Kraagor's tomb for math/CS types

[pendell]

Problem: We have 121 doors arranged in sequence. Let's assume that the gate really is behind the doors but we don't know which one it is.

So: Serious question. What is the optimal search strategy we could pursue that would allow us the greatest probability of finding the door in the least amount of tries? Put differently, how can we maximize our coverage of the search space?

The naive approach, of course, is to start with door 0, then door 1, 2, in sequence. The problem with this approach is that if the door is door 80 or above, you're going to have a minimum of 79 tries before you get to the right door.

So intuitively speaking it might give us a better chance to make sure that we cover the entire search space in the first ten tries -- one pick from the first twenty doors, one pick from the last twenty, one pick from the range 40-60, and so on. The idea is to cover more of the search space in fewer tries, in the hopes of stumbling on the right one more quickly. Continue this form of jumping around until all doors are covered or until we find the correct door.


But that's intuitive , and may not actually yield a better probable outcome than simply going through the doors in order.

So that's why I'm asking: What can we say about this problem, theoretically? Is going through the doors in sequence really the best possible search strategy? If not, what would be a better one, and why?

ETA: Not to mention, this tomb is designed by an intelligent mind, not a random one, so it's likely the true door is NOT located in the first or the last few doors, to deliberately cause trouble for people attempting a sequential search.


Respectfully,

Brian P.

[Crusher]

Problem: We have 121 doors arranged in sequence. Let's assume that the gate really is behind the doors but we don't know which one it is.

So: Serious question. What is the optimal search strategy we could pursue that would allow us the greatest probability of finding the door in the least amount of tries? Put differently, how can we maximize our coverage of the search space?

The naive approach, of course, is to start with door 0, then door 1, 2, in sequence. The problem with this approach is that if the door is door 80 or above, you're going to have a minimum of 79 tries before you get to the right door.

So intuitively speaking it might give us a better chance to make sure that we cover the entire search space in the first ten tries -- one pick from the first twenty doors, one pick from the last twenty, one pick from the range 40-60, and so on. The idea is to cover more of the search space in fewer tries, in the hopes of stumbling on the right one more quickly. Continue this form of jumping around until all doors are covered or until we find the correct door.


But that's intuitive , and may not actually yield a better probable outcome than simply going through the doors in order.

So that's why I'm asking: What can we say about this problem, theoretically? Is going through the doors in sequence really the best possible search strategy? If not, what would be a better one, and why?

ETA: Not to mention, this tomb is designed by an intelligent mind, not a random one, so it's likely the true door is NOT located in the first or the last few doors, to deliberately cause trouble for people attempting a sequential search.


Respectfully,

Brian P.

I think your final point is the key one. With that as your starting point, its worth remembering the biggest assumption in your question: That ANY of the 121 doors is the correct one.

Of course, that's fine for your exercise, but for Team Evil is probably a crucial point.

[Ranzear]

Divide and conquer doesn't work if each access provides no info besides 'not here'.

The only thing to go on is your last line: People are bad random number generators. I'd suggest rule of thirds.

[pendell]

Divide and conquer doesn't work if each access provides no info besides 'not here'.

The only thing to go on is your last line: People are bad random number generators. I'd suggest rule of thirds.

I'm not familiar with that particular concept as applied to searching. I don't see it on wikipedia (https://en.wikipedia.org/wiki/Rule_of_three) either. Can you explain what you mean?

Respectfully,

Brian P.

[Spartakus]

The key point is indeed of the Bulder not having a random mind. However, if Serini was aware of this (and I think she was), the optimal hiding strategy would be to use some random process (like rolling a fair die) for choosing a door to be the correct one. This would ensure, that there is no strategy better then any other when trying to find the correct one.

However, team evils strategy gets really bad, when a third party like the OOTS becomes involved. If you paint an X on each searched door, not only do you show to the other team which doors are already determined to be false, they could theoretically paint every door that has not yet been chosen and so completely eliminate your progress thus far.
If they had chosen a deterministic approach, but one that becomes obvious (like starting top left and ending bottom right) it would be better, since you can't just erase progress, but you can predict which doors are next and search ahead, thus ensuring Xycon could only choose doors already proven to be false.

Best strategy is using a random approach each day, but taking secret notes, which doors are false.

[Roland Itiative]

Assuming the correct door is truly random, there is no best approach. None of the strategies used in computer programming to optimise searches apply in this situation. The only thing that can improve the search is being able to check more doors in the same time frame, by having two or more groups going into different doors (and coming back alive).

And even considering the human mind is not really good at true randomness, it's still pretty hard to guess which door would be correct, unless Serini's diary is detailed enough that some quirks of her methods of choosing random things becomes apparent.

On a somewhat unrelated note, I don't think there are only 121 doors. The impression I got from that panel is that the Monster Hollow is a pit, and the entire inner surface of the pit's walls are covered in doors, from all angles. And even if that's not the case, and it's really just a wall, we don't see the end of it from either side in that panel, so all we can say is that there are at least 121 doors.

[Quartz]

The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.

[Shining Wrath]

If you don't gain any information from any door other than "yes/no", then the best way to narrow it down is not via Computer Science methods, but psychological ones; where would Serini have put the gate?

However, we may well be going up against a Sicilian Halfling Rogue with death the safety of the world on the line, in which case Serini would anticipate that someone would try to guess where she would have put the Gate, and therefore not make the choice herself - either randomly, or by allowing someone with completely different thought processes (like, say, a Bugbear from a nearby village) to choose.

[Sir_Norbert]

The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.

Far from "of course", this is extremely unlikely to be the solution, for reasons that have already been gone over thoroughly in the main #1039 discussion topic:


I predict that the Gate won't be beneath Kraagor's statue, for the same reason it wasn't beneath Girard's statue: you do not put your double bluff outside your defenses.

Because if you do - and someone guesses the con: it's defenseless.

It is perfectly possible, of course, that none of the doors lead to the Gate. Why make the gate accessible at all? Clearly some pretty major landscaping has been involved, so why not hide the Gate inside solid rock, wrapped around with dungeons full of monsters, so you can't even dig down to it without setting them all off.

Or, you know, maybe Kraagor's statue is the doorway into the real boss dungeon, complete with stuff that dumps the contents of all the other doors down into it the moment the doorway is opened.

[pendell]

The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.


Most likely. But for the sake of learning about search theory, I'm going to hypothesize for the sake of argument that it really IS one of the doors.



Assuming the correct door is truly random, there is no best approach. None of the strategies used in computer programming to optimise searches apply in this situation. The only thing that can improve the search is being able to check more doors in the same time frame, by having two or more groups going into different doors (and coming back alive).

And even considering the human mind is not really good at true randomness, it's still pretty hard to guess which door would be correct, unless Serini's diary is detailed enough that some quirks of her methods of choosing random things becomes apparent.

On a somewhat unrelated note, I don't think there are only 121 doors. The impression I got from that panel is that the Monster Hollow is a pit, and the entire inner surface of the pit's walls are covered in doors, from all angles. And even if that's not the case, and it's really just a wall, we don't see the end of it from either side in that panel, so all we can say is that there are at least 121 doors.



Hmm... I know where I've seen this before. Skip to 27:00 in this video for Starship Titanic's Bomb Puzzle (https://www.youtube.com/watch?v=sONIZUf3UbY). And a shorter version (https://www.youtube.com/watch?v=_sB24zOcV_Q).

As John Cleese tells us, There are 23 barrels with 15 combinations each, which gives us several million possible combinations. The puzzle then hints... "Look at it this way. You can either try all the possible combinations which will take us neatly up to the heat death of the universe OR you can try being clever." It then proceeds to give us the exact hint we need.

So I guess this is a problem less for computer science than it is for social engineering (https://en.wikipedia.org/wiki/Social_engineering_(security)) -- to research Serini thoroughly to figure out what the key to the puzzle is. Surely there must be one -- she would need a way to access the gate herself in time of trouble, so she must have had SOME way to bypass the decoys and go directly to the true entrance.

Simply tackling the doors one at a time is a sucker's game. But if it keeps Team Evil occupied, it serves its purpose. Only ... only I thought Redcloak was smarter than that.

Respectfully,

Brian P.

[King of Nowhere]

Assuming the correct door is truly random, there is no best approach. None of the strategies used in computer programming to optimise searches apply in this situation. The only thing that can improve the search is being able to check more doors in the same time frame, by having two or more groups going into different doors (and coming back alive).



Exactly that. It is mathematically demonstrated that in a game of pure random chance like this one, there is no better strategy than simply picking at random.
You could try some pshycological things, though.
For example, serini is unlikely to have hid the gate under the first or last door; when we pick a random number, we are more likely to avoid good "round" numbers like 10 or 50, and we prefer to instead pick numbers with 3 or 7. Also, the instintive first approach would be to go through them in sequence, or maybe in reverse sequence, so the instinctive place to hide it would be around two thirds of the sequence.
For those reasons, if I were facing such a trial, I'd start with doors 78, 87, 83, 73, followed by all remaining doors between 70 and 90.
Of course, the best way to hide the gate would be to throw a dice to decide where to place it (or rather, how to build the dungeon around it). Nothing forces serini to fall for the typical failings of the human mind when coming to chance - though I expect rogues to not be mathematically-minded people.


The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.
Well, maybe not inside the statue, as that would be a fairly obvious place to look for if you suspect a misdirection - and really, when facing a situation like that, a misdirection is one of the first things you should suspect. But yes, there is a significant chance the gate is somewhere else. Or that, when you open a door, it won't be apparent if the gate is there or not.

[Bulhakov]

If the correct door is in a purely random location, then there is no optimal search strategy. A simpler example would be to throw a dice and have someone guess the result. He can ask 1?,2?,3?... or 3?,1?,5? etc. - no order of asking will have an advantage over the rest.

The intelligent design (instead of pure randomness) of the tomb might of course change this, but we have no info on that yet, but there are some good real-life examples such as PIN numbers - theoretically they should be evenly distributed, but in practice some numbers are much more popular than others: http://www.datagenetics.com/blog/september32012/g2.png

[Roland Itiative]

Most likely. But for the sake of learning about search theory, I'm going to hypothesize for the sake of argument that it really IS one of the doors.




Hmm... I know where I've seen this before. Skip to 27:00 in this video for https://www.youtube.com/watch?v=sONIZUf3UbY] ( [url) Starship Titanic's Bomb Puzzle [/url]. And a shorter version (https://www.youtube.com/watch?v=_sB24zOcV_Q).

As John Cleese tells us, There are 23 barrels with 15 combinations each, which gives us several million possible combinations. The puzzle then hints... "Look at it this way. You can either try all the possible combinations which will take us neatly up to the heat death of the universe OR you can try being clever." It then proceeds to give us the exact hint we need.

So I guess this is a problem less for computer science than it is for social engineering (https://en.wikipedia.org/wiki/Social_engineering_(security)) -- to research Serini thoroughly to figure out what the key to the puzzle is. Surely there must be one -- she would need a way to access the gate herself in time of trouble, so she must have had SOME way to bypass the decoys and go directly to the true entrance.

Simply tackling the doors one at a time is a sucker's game. But if it keeps Team Evil occupied, it serves its purpose. Only ... only I thought Redcloak was smarter than that.

Respectfully,

Brian P.

If Serini wanted a way to bypass the decoys, all she had to do is memorise which is the correct door. She designed the thing, she would know. That, by itself, doesn't give any information to an invader, the correct door could still be completely random.

Then there is the possibility that she really made it in a way that even she doesn't know the answer. Presumably, if she knew, it might have been easier for Team Evil to track her (or her corpse) and extract the information directly at the source. She'd have to protect herself as well as she protects the Gate.

[Mx56]

We know that random number generation exists and is used to hide things in this world, and has been used by at least one member of the Order of the Scribble (http://www.giantitp.com/comics/oots0694.html). If you had some means of knowing which method was used to randomly generate the numbers (and hence the probability distribution over the doors), that may or may not give you a more useful strategy than sequentially opening doors - if the method was to take 11 rows and 11 columns and chose row and column by 2d6-1, you'd be best looking at the middle rows and columns first, while if Serini got a d121 from somewhere, you're screwed. The smart strategy if you're Serini is to assign numbers to doors not based on their sequential order, sample numbers from a uniform distribution to pick the door, and then burn any notes that you made on how you did it before feeding the ashes to your pet sphere of annihilation.

(all of this assumes 121 doors, but still applies regardless)

ETA: My silly comment about a d121 aside, you could sample from a uniform distribution with OoTS-world technology using a well shuffled set of numbered cards pretty straightforwardly.

[pendell]

Can we trust random number technology, though? Might the same anti-divination measures which prevent Redcloak from divining the tomb's location also bollix an RNG to prevent it returning a true answer?

*Thinks*

Actually, if it DOES prevent a true number from being generated, that makes the problem easier, not harder. Use magic to summon a scribe or create a golem to write down the answers, then roll a couple thousand times. Write down all the numbers that show up, taking special note of any numbers which don't. The number that doesn't turn up is the place you search :smallamused:.

Respectfully,

Brian P.

[Jasdoif]

The naive approach, of course, is to start with door 0, then door 1, 2, in sequence. The problem with this approach is that if the door is door 80 or above, you're going to have a minimum of 79 tries before you get to the right door.

So intuitively speaking it might give us a better chance to make sure that we cover the entire search space in the first ten tries -- one pick from the first twenty doors, one pick from the last twenty, one pick from the range 40-60, and so on. The idea is to cover more of the search space in fewer tries, in the hopes of stumbling on the right one more quickly. Continue this form of jumping around until all doors are covered or until we find the correct door.


But that's intuitive , and may not actually yield a better probable outcome than simply going through the doors in order.

So that's why I'm asking: What can we say about this problem, theoretically? Is going through the doors in sequence really the best possible search strategy? If not, what would be a better one, and why?Here's the thing: any sequence of opening doors maps to opening "door 0", then "door 1", etc., in order. It's no different than an integer index on an array, "door 0" is simply the first door you check and so on. If there's really no information available other than "one and only one of these doors has what we're looking for", and the extent of information from a wrong check is "false", checking any full sequence of doors individually is effectively the same strategy as checking any other full sequence of doors individually.

[pendell]

Here's the thing: any sequence of opening doors maps to opening "door 0", then "door 1", etc., in order. It's no different than an integer index on an array, "door 0" is simply the first door you check and so on. If there's really no information available other than "one and only one of these doors has what we're looking for", and the extent of information from a wrong check is "false", checking any full sequence of doors individually is effectively the same strategy as checking any other full sequence of doors individually.

I'm not quite sure that's fair. While you can map the sequence that way, it definitely has an impact on the outcome. If the correct door is #37 (say), starting a sequential search at 1 will get the correct answer in 37 tries, starting from 30 will give it to you in 7 tries, and starting on 1 and searching only primes gives you 12 tries. So it seems a false equivalence, to me. While it is possible to describe all sequences thus, the sequences are not all identical and do not all have the same outcome.

Respectfully,

Brian P.

[Mx56]

Can we even assume that Serini actually got to pick a door, by the way? It could be, for instance, that the rift was in a naturally occurring network of caves and the dungeon was built into it, doors being added to all possible entrances.

[Shining Wrath]

I'm not quite sure that's fair. While you can map the sequence that way, it definitely has an impact on the outcome. If the correct door is #37 (say), starting a sequential search at 1 will get the correct answer in 37 tries, starting from 30 will give it to you in 7 tries, and starting on 1 and searching only primes gives you 12 tries. So it seems a false equivalence, to me. While it is possible to describe all sequences thus, the sequences are not all identical and do not all have the same outcome.

Respectfully,

Brian P.

I believe you misunderstand Jasdoif. Suppose you pick the door by counting the number of knots visible in the wood and opening them in order of knotiness. And it just happens that the Gate is behind the first door you open. Yeah! You win! But that doesn't mean that knotiness was actually a better method than rolling dice, going from left to right and bottom to top, or asking the MitD to choose.

Once you know where the Gate is you'll know what method works best. However, until you know you don't have any reason to believe that starting at 1 and working up is better than starting at 121 and working down, or that either is better than rolling dice.

Or that any of the three are better than counting knots in the wood of the doors.

To summarize, there are many ways you could conduct this search. Depending on where the Gate is, some of them will work better than others. However, you don't know which ones work better until you find the Gate at which point you don't have to search any longer.

[Jasdoif]

I'm not quite sure that's fair. While you can map the sequence that way, it definitely has an impact on the outcome. If the correct door is #37 (say), starting a sequential search at 1 will get the correct answer in 37 tries, starting from 30 will give it to you in 7 tries, and starting on 1 and searching only primes gives you 12 tries. So it seems a false equivalence, to me. While it is possible to describe all sequences thus, the sequences are not all identical and do not all have the same outcome.It impacts the outcome, yes. But not the strategy.

What you've described is "door 36" being correct in one sequence (assuming you're still going with zero-indexing), "door 6" in another sequence, and "door 11" in yet another sequence. But they're all the same door. And we don't know which door is correct until we find it, and until we find it there's no basis for saying one sequence is "better" than another, because there's no way to know what position the correct door has in any sequence without knowing which door is the correct door.

[Roland Itiative]

Can we trust random number technology, though? Might the same anti-divination measures which prevent Redcloak from divining the tomb's location also bollix an RNG to prevent it returning a true answer?

*Thinks*

Actually, if it DOES prevent a true number from being generated, that makes the problem easier, not harder. Use magic to summon a scribe or create a golem to write down the answers, then roll a couple thousand times. Write down all the numbers that show up, taking special note of any numbers which don't. The number that doesn't turn up is the place you search :smallamused:.

Respectfully,

Brian P.
Well made dice (with someone rolling them "right", without trying to cheat in any way) are one of the most reliable sources of random numbers that exist, and are pretty primitive technology. Sure, a medieval world wouldn't be able to make dice as reliable as the ones used in modern casinos (or maybe they can, considering magic is common in that universe), but they still should be able to make pretty decent ones.

And I don't get the second part of your post. Anti-divination measures shouldn't affect the outcome of a dice roll at all, it only affects divination magic.

[Mad Humanist]

What if some of the monster are earth elementals. They could rearrange the tunnels every day so the question is even meaningless.

[Roland Itiative]

What if some of the monster are earth elementals. They could rearrange the tunnels every day so the question is even meaningless.

If that's the case, there is no point in marking the doors that were already explored. It's definitely possible, and would indeed be a pretty decent way to defend the Gate. In that case, choosing the right door would always be a 1/121 chance, rather than a cumulative (x+1)/121, where x is the number of doors previously explored.

[pendell]

Hmm .. I'll have to think on Jasdoif's comments because I'm not fully understanding. Still, a bit of cogitation should set that to rights!



And I don't get the second part of your post. Anti-divination measures shouldn't affect the outcome of a dice roll at all, it only affects divination magic.


Divination is closely related to random number generation -- tao sticks, tea leaf patterns, flipping a coin, magic 8-ball, Bibliomancy (https://en.wikipedia.org/wiki/Bibliomancy), you name it, all revolve around the assumption that some supernatural force will influence a random event to give an answer. I surmise it's possible that anti-divination magic would trigger on an attempt to divine the correct answer through random number generation.

Of course, when I say this I have Erfworld in mind. In that comic, Parson asked what the probability was that he could cast a specific spell. His calculator artifact at first gave a very high answer, then almost immediately revised the answer down to 0%. This is because Fate would intervene and guarantee a result that normal probability would not.

Likewise, I think it possible sufficiently well-crafted anti-divination magic would interfere with normal random rolls being used in their vicinity, given the close relationship between random numbers and divination.

Respectfully,

Brian P.

[Kish]

allowing someone with completely different thought processes (like, say, a Bugbear from a nearby village) to choose.
Then clearly it's the last door, because that means an intruder has to go through all the doors to get to it.

Is there a problem with that line of reasoning?

Not that I'd expect Oona to notice.

Hmm .. I'll have to think on Jasdoif's comments because I'm not fully understanding.
If you have no reason to expect a pattern, every random search method is equally good or bad. If it happens to be the upper left door, opening them in order starting with that one looks ideal and opening them in order starting with the lower right one looks awful--but it's still just a coincidence. Bubble sort and similar methods presume that you get more information than "not this one."

That is, if every door included the message, "The correct tomb is [up/down] and [left/right] of this one," then the logical method would be to start at the midpoint, cross off half the doors, follow it up with the midpoint of the new range, and continue until either finding the door, or getting a message which narrows it down to only one door. But since they only give the information "not this one (unless you missed a secret door, heh heh heh)," there are two basic alternatives. One is to try to figure out Serini's mind, and hope she didn't anticipate your ideas. The other is to accept that complete randomness is the best you can do here. Using prime numbers or some equivalent is only better than taking the doors in order if you have some reason to think Serini used primes (i.e., it's option 1); pointing out that hypothetically it might lead to the door quickly doesn't make it statistically better, since it's equally likely, hypothetically, that it will lead to the door last.

[KillingAScarab]

*sigh* This would be the perfect time for Donald Knuth to be swinging a sword (http://www.explainxkcd.com/wiki/index.php/342).

[DaggerPen]

Then clearly it's the last door, because that means an intruder has to go through all the doors to get to it.

Is there a problem with that line of reasoning?

Not that I'd expect Oona to notice.

... you and I have very different readings of Oona.

Regarding the doors - it's very true that, if all doors are equivalent except in that one has the Gate at the end, assuming no bias in choosing on Serini's part, you can't do better than random chance.

However. I personally would be inclined to start with the upper doors. Why? Because, barring extra dimensional spaces, you can get more information from them.

Consider: the hollow can only sustain so many tunnels of that size before the tunnels start running into each other. The ones at the bottom have room to go down, but what of the ones at the top? If they don't go all the way down, a sampling of several of them and marking the inside would give you a good sense that they have less room in them or that run together, meaning that either you eliminate multiple doors with one trip or there's less space for the Gate to be in there. If the former, search more, because you can hit multiple doors more efficiently. If the latter, move your search to the lower ones, because there's more room for the Gate down there. And if the tunnels go down all the way, keep at the ones on the top, because then you're covering more ground behind one door.

It's not that much more efficient, but there's probably an edge over random chance.

[dancrilis]

So: Serious question. What is the optimal search strategy we could pursue that would allow us the greatest probability of finding the door in the least amount of tries? Put differently, how can we maximize our coverage of the search space?


Phylactery search technique I would say, supported by summoned minions would seem best - if the caves are no too dangerous for Bugbears than Hobgoblins should be fine - if you care about that kind of thing (especially with extra planer support), and gating them in should not be a problem.

Ignoring that than using magic to Ghost Form (perhaps invisibly) through the tunnels and not bother to engage with the monsters would seem quickest for Xykon (but this to an extent depends on what protections the gate has which we may not know yet).

Those would be my suggestions on greatest probability of finding the door in the least amount of time at any rate.

[Jasdoif]

Hmm .. I'll have to think on Jasdoif's comments because I'm not fully understanding. Still, a bit of cogitation should set that to rights!Well, let me try it from the other direction.

To demonstrate this more easily, let's suppose a similar scenario with only 3 doors. Let's call them A, B and C.

That means there's 6 full sequences (permutations) of doors:

ABC
ACB
BAC
BCA
CAB
CBA

So in ABC, A is the first door you check, B is the second door you check, and C is the last door you check.


Now, let's suppose A is the correct door. That means there's 2 sequences where you check the correct door first (ABC & ACB), 2 where you check the correct door second (BAC & CAB), and 2 where you check the correct door last (BCA & CBA). So a randomly chosen sequence has the same chance of having the correct door first, and it does of having the correct door second, and as it does of having the correct door last.

Now, let's suppose B is the correct door. That means there's 2 sequences where you check the correct door first (BAC & BCA), 2 where you check the correct door second (ABC & CBA), and 2 where you check the correct door last (ACB & CAB). So a randomly chosen sequence has the same chance of having the correct door first, and it does of having the correct door second, and as it does of having the correct door last.

Finally, let's suppose C is the correct door. That means there's 2 sequences where you check the correct door first (CAB & CBA), 2 where you check the correct door second (ACB & BCA), and 2 where you check the correct door last (ABC & BAC). So a randomly chosen sequence has the same chance of having the correct door first, and it does of having the correct door second, and as it does of having the correct door last.

....So no matter which door is correct, randomly choosing a sequence has equal odds of having the correct door in any position. And each sequence has a different "ranking" for each door (ABC has you open one door if A is correct, two doors if B is correct, and all three if C is correct), so any chosen sequence also has equal odds of having the correct door first/second/last if the correct door is randomly chosen.

And since we don't know which door is correct, nor any information in advance to narrow down which door is correct, or weigh which is more likely to be correct...from our perspective, each door is equally likely, giving it the same probability as if it were randomly chosen.


So ultimately, unless/until we find the correct door to know where it is, every sequence has equal odds (from our perspective) of having the correct door at any given position. With just the information we have, there's no basis on which to believe any particular sequence is more or less likely to have the correct door sooner than any other particular sequence.

[pendell]

Then clearly it's the last door, because that means an intruder has to go through all the doors to get to it.

Is there a problem with that line of reasoning?

Not that I'd expect Oona to notice.

If you have no reason to expect a pattern, every random search method is equally good or bad. If it happens to be the upper left door, opening them in order starting with that one looks ideal and opening them in order starting with the lower right one looks awful--but it's still just a coincidence. Bubble sort and similar methods presume that you get more information than "not this one."

That is, if every door included the message, "The correct tomb is [up/down] and [left/right] of this one," then the logical method would be to start at the midpoint, cross off half the doors, follow it up with the midpoint of the new range, and continue until either finding the door, or getting a message which narrows it down to only one door. But since they only give the information "not this one (unless you missed a secret door, heh heh heh)," there are two basic alternatives. One is to try to figure out Serini's mind, and hope she didn't anticipate your ideas. The other is to accept that complete randomness is the best you can do here. Using prime numbers or some equivalent is only better than taking the doors in order if you have some reason to think Serini used primes (i.e., it's option 1); pointing out that hypothetically it might lead to the door quickly doesn't make it statistically better, since it's equally likely, hypothetically, that it will lead to the door last.

*Nods* That makes sense -- so what you, Jasdoif, and others are saying is that more clever search strategies need more information than simply "not there" in order to do dividing-and-conquering or some other method of gaining information. If all we get is a simple "false", then all search algorithms are equally valid.

Of course, if we can gain even a hint of what was going through Serini's mind, or any kind of clue as to how the dungeon is constructed, then alternate search strategies become more feasible.

But at the moment, we're reduced to either picking doors at random, or researching Serini to gain clues. Me thinks the second is probably a better choice, unless we care enough about leveling up minions like Oona to wish to fight these difficult battles any way.

Respectfully,

Brian P.

[chy03001]

If the correct door is in a purely random location, then there is no optimal search strategy. A simpler example would be to throw a dice and have someone guess the result. He can ask 1?,2?,3?... or 3?,1?,5? etc. - no order of asking will have an advantage over the rest.

The intelligent design (instead of pure randomness) of the tomb might of course change this, but we have no info on that yet, but there are some good real-life examples such as PIN numbers - theoretically they should be evenly distributed, but in practice some numbers are much more popular than others: http://www.datagenetics.com/blog/september32012/g2.png

I love this diagram.

[Shining Wrath]

How many high-CR monsters have burrow?

Phaetheon (http://www.d20srd.org/srd/epic/monsters/abomination.htm)
Primal Earth Elemental (http://www.d20srd.org/srd/epic/monsters/elementalPrimal.htm)
Genius Loci (http://www.d20srd.org/srd/epic/monsters/geniusLoci.htm)
Dragons: blue, white, brass (http://www.d20srd.org/srd/monsters/dragonTrue.htm)

I'm sure there's more, but if any of those creatures are in the dungeon then the tunnels are continually being changed.

[dancrilis]

How many high-CR monsters have burrow?

Phaetheon (http://www.d20srd.org/srd/epic/monsters/abomination.htm)
Primal Earth Elemental (http://www.d20srd.org/srd/epic/monsters/elementalPrimal.htm)
Genius Loci (http://www.d20srd.org/srd/epic/monsters/geniusLoci.htm)
Dragons: blue, white, brass (http://www.d20srd.org/srd/monsters/dragonTrue.htm)

I'm sure there's more, but if any of those creatures are in the dungeon then the tunnels are continually being changed.

Burrow doesn't change the tunnels there is no tunnel left by the burrower (normally).


Burrow (http://www.d20srd.org/srd/specialAbilities.htm#burrow)
A creature with a burrow speed can tunnel through dirt, but not through rock unless the descriptive text says otherwise. Creatures cannot charge or run while burrowing. Most burrowing creatures do not leave behind tunnels other creatures can use (either because the material they tunnel through fills in behind them or because they do not actually dislocate any material when burrowing); see the individual creature descriptions for details.

I also doubt that the gate is guarded by high intelligence evil beings - would seem a like it might backfire somehow.

[Hopeless]

Firstly you should remember this tomb was constructed by a Halfling... specifically a Chaotic aligned Halfling whose hero was Girard.

So taking that into regard my solution would be to blow up all of the doors as the real gate would undoubtedly be hidden behind them.

Secondly if I wanted to keep this tomb intact and populated with the strongest monsters I would set as a guard a force that is dependent on that tomb remaining intact since as long as it remains intact they have a ready access of food and resources they would be disinclined to lose.

So its not because its guarded by evil monsters but its guarded by evil monsters who have a pretty d**n reason to want nothing to happen to it unlike Girard's reliance on his family that through demonic and devilish mischance were all wiped out by a wizard possessed by a spirit who knows that particular epic level spell...

Anyway how long do you think it'll be before Xykon loses what little patience he ever had and just blows up the place?:smallwink:

[Kish]

Firstly you should remember this tomb was constructed by a Halfling... specifically a Chaotic aligned Halfling whose hero was Girard.
Not sure where you get any of that but "halfling."

... you and I have very different readings of Oona.
We sure do! You seem to be in the majority (I am glad to say that neither of us has shown any desire to shove her into a romantic pairing), but I expect we'll find out who's closer to right pretty soon. (Counting only strips Oona is in, that is. We may well cut away from Oona for another 500 strips now.)

[SirKazum]

Maybe Team Evil doesn't have access to that intel (or maybe it did, from the Dungeon of Dorukan), but there's a small piece of information that we've got about Kraagor's Gate: its height. Yes, I'd like to call your attention to the device in panel 2 (http://www.giantitp.com/comics/oots0844.html). Some of those orbs look higher than others, and at first I thought it meant they were bobbing up and down in mid-air (which would be a nice magical effect), but no, it seems consistent from panel to panel.

Now, if you look at how high each orb is, and compare it to what we know about the Gates, they seem to map pretty well to their physical location. Lirian's Gate is at (nearly) ground level, and the orb is touching the board - same for Girard's Gate, which is at the bottom of a pyramid, and thus roughly at ground level (look at the rift here (http://www.giantitp.com/comics/oots0900.html)). The orb corresponding to Dorukan's rift is a bit above the board, and while the gate is at the bottom of a dungeon, the dungeon itself is on a castle in mountainous terrain (http://www.giantitp.com/comics/oots0013.html), so it might still be a ways above "ground level" (and the bare rift in the Scribbles of Time (http://www.giantitp.com/comics/oots0276.html) story seems to support that, although I place less stock in that since it's second-hand information). And the Azure City Gate is rather high up in the air, being on top of the throne that's at one of the higher floors of a huge castle, and the resulting rift in Gobbotopia supports that as well, corresponding to a rather high orb.

So, the orb corresponding to Kraagor's Gate is about the same height as Soon's, maybe slightly lower, depending on the angle of the board. The Scribbles of Time drawing seems to show it at just above ground level - then again, not only it's a second-hand account, but the landscape was also changed by digging an enormous pit, so maybe the height "counts" down to the bottom of that pit. At a rough comparison between this (http://www.giantitp.com/comics/oots1039.html) and this (http://www.giantitp.com/comics/oots0510.html) (bottom panel), I'd expect the gate to be toward the top - if not at the topmost row, then not very far from it.

With that in mind, my strategy would be to start searching the topmost row, at whatever order, and work my way down.

[DaggerPen]

Not sure where you get any of that but "halfling."

We sure do! You seem to be in the majority (I am glad to say that neither of us has shown any desire to shove her into a romantic pairing), but I expect we'll find out who's closer to right pretty soon. (Counting only strips Oona is in, that is. We may well cut away from Oona for another 500 strips now.)

That we shall!

(Don't get me wrong, regardless of how much of Oona's and Greyview's relationship is just "mutual respect involving a warg who is just relentlessly fatalistic for unrelated reasons" vs. how much is "Oona doesn't really respect Greyview and Greyview has just given up", her insistence that all "beasts" need a "master" is the opposite of good or wise. But she does strike me as rather a shrewd sort with solid sense and reasoning who just speaks Common as a second language, given her analysis of the bugbear/goblin relationship and pan goblin narrative and her approach to Kraagor's Tomb. She may not be exceptionally wise, but I'd peg her as having far more sense than "Put it behind the last door!" That's pre-character-development Elan level of lack of sense.]


Maybe Team Evil doesn't have access to that intel (or maybe it did, from the Dungeon of Dorukan), but there's a small piece of information that we've got about Kraagor's Gate: its height. Yes, I'd like to call your attention to the device in panel 2 (http://www.giantitp.com/comics/oots0844.html). Some of those orbs look higher than others, and at first I thought it meant they were bobbing up and down in mid-air (which would be a nice magical effect), but no, it seems consistent from panel to panel.

Now, if you look at how high each orb is, and compare it to what we know about the Gates, they seem to map pretty well to their physical location. Lirian's Gate is at (nearly) ground level, and the orb is touching the board - same for Girard's Gate, which is at the bottom of a pyramid, and thus roughly at ground level (look at the rift here (http://www.giantitp.com/comics/oots0900.html)). The orb corresponding to Dorukan's rift is a bit above the board, and while the gate is at the bottom of a dungeon, the dungeon itself is on a castle in mountainous terrain (http://www.giantitp.com/comics/oots0013.html), so it might still be a ways above "ground level" (and the bare rift in the Scribbles of Time (http://www.giantitp.com/comics/oots0276.html) story seems to support that, although I place less stock in that since it's second-hand information). And the Azure City Gate is rather high up in the air, being on top of the throne that's at one of the higher floors of a huge castle, and the resulting rift in Gobbotopia supports that as well, corresponding to a rather high orb.

So, the orb corresponding to Kraagor's Gate is about the same height as Soon's, maybe slightly lower, depending on the angle of the board. The Scribbles of Time drawing seems to show it at just above ground level - then again, not only it's a second-hand account, but the landscape was also changed by digging an enormous pit, so maybe the height "counts" down to the bottom of that pit. At a rough comparison between this (http://www.giantitp.com/comics/oots1039.html) and this (http://www.giantitp.com/comics/oots0510.html) (bottom panel), I'd expect the gate to be toward the top - if not at the topmost row, then not very far from it.

With that in mind, my strategy would be to start searching the topmost row, at whatever order, and work my way down.

... this is a SUPER good catch. Well spotted.

[Roland Itiative]

Hmm .. I'll have to think on Jasdoif's comments because I'm not fully understanding. Still, a bit of cogitation should set that to rights!



Divination is closely related to random number generation -- tao sticks, tea leaf patterns, flipping a coin, magic 8-ball, Bibliomancy (https://en.wikipedia.org/wiki/Bibliomancy), you name it, all revolve around the assumption that some supernatural force will influence a random event to give an answer. I surmise it's possible that anti-divination magic would trigger on an attempt to divine the correct answer through random number generation.

Of course, when I say this I have Erfworld in mind. In that comic, Parson asked what the probability was that he could cast a specific spell. His calculator artifact at first gave a very high answer, then almost immediately revised the answer down to 0%. This is because Fate would intervene and guarantee a result that normal probability would not.

Likewise, I think it possible sufficiently well-crafted anti-divination magic would interfere with normal random rolls being used in their vicinity, given the close relationship between random numbers and divination.

Respectfully,

Brian P.
Oh, I see what you mean now. But, in this case, I'd say that an effective anti-divination spell wouldn't just keep the right answer from appearing on the dice rolls, but keep the magic from affecting the randomness at all, since deliberately blocking a result also gives important information. It's better to give your enemies a chance to find the right result by chance, than trying to block it altogether and leaving a loophole that allows you to be outsmarted :smalltongue:


Maybe Team Evil doesn't have access to that intel (or maybe it did, from the Dungeon of Dorukan), but there's a small piece of information that we've got about Kraagor's Gate: its height. Yes, I'd like to call your attention to the device in panel 2 (http://www.giantitp.com/comics/oots0844.html). Some of those orbs look higher than others, and at first I thought it meant they were bobbing up and down in mid-air (which would be a nice magical effect), but no, it seems consistent from panel to panel.

Now, if you look at how high each orb is, and compare it to what we know about the Gates, they seem to map pretty well to their physical location. Lirian's Gate is at (nearly) ground level, and the orb is touching the board - same for Girard's Gate, which is at the bottom of a pyramid, and thus roughly at ground level (look at the rift here (http://www.giantitp.com/comics/oots0900.html)). The orb corresponding to Dorukan's rift is a bit above the board, and while the gate is at the bottom of a dungeon, the dungeon itself is on a castle in mountainous terrain (http://www.giantitp.com/comics/oots0013.html), so it might still be a ways above "ground level" (and the bare rift in the Scribbles of Time (http://www.giantitp.com/comics/oots0276.html) story seems to support that, although I place less stock in that since it's second-hand information). And the Azure City Gate is rather high up in the air, being on top of the throne that's at one of the higher floors of a huge castle, and the resulting rift in Gobbotopia supports that as well, corresponding to a rather high orb.

So, the orb corresponding to Kraagor's Gate is about the same height as Soon's, maybe slightly lower, depending on the angle of the board. The Scribbles of Time drawing seems to show it at just above ground level - then again, not only it's a second-hand account, but the landscape was also changed by digging an enormous pit, so maybe the height "counts" down to the bottom of that pit. At a rough comparison between this (http://www.giantitp.com/comics/oots1039.html) and this (http://www.giantitp.com/comics/oots0510.html) (bottom panel), I'd expect the gate to be toward the top - if not at the topmost row, then not very far from it.

With that in mind, my strategy would be to start searching the topmost row, at whatever order, and work my way down.

They are all floating above the board, it's just the perspective that makes it seem otherwise.

[pendell]

Not sure where you get any of that but "halfling."


First of all, Serini's a rogue, so she can't be lawful, correct?

Second of all, what we've seen of her diary shows a bunch of hearts around Girard and a "meanie" caption for Soon -- which strongly implies A) a chaotic alignment B) a liking for Girard, if not a schoolgirl crush.

If I were seriously trying to break this puzzle myself I would definitely explore the hypothesis that Serini admired Girard and would incorporate some of his ideas into her dungeon.

Respectfully,

Brian P.

[Frankablu]

Problem: We have 121 doors arranged in sequence. Let's assume that the gate really is behind the doors but we don't know which one it is.

So: Serious question. What is the optimal search strategy we could pursue that would allow us the greatest probability of finding the door in the least amount of tries? Put differently, how can we maximize our coverage of the search space?

The naive approach, of course, is to start with door 0, then door 1, 2, in sequence. The problem with this approach is that if the door is door 80 or above, you're going to have a minimum of 79 tries before you get to the right door.

So intuitively speaking it might give us a better chance to make sure that we cover the entire search space in the first ten tries -- one pick from the first twenty doors, one pick from the last twenty, one pick from the range 40-60, and so on. The idea is to cover more of the search space in fewer tries, in the hopes of stumbling on the right one more quickly. Continue this form of jumping around until all doors are covered or until we find the correct door.


But that's intuitive , and may not actually yield a better probable outcome than simply going through the doors in order.

So that's why I'm asking: What can we say about this problem, theoretically? Is going through the doors in sequence really the best possible search strategy? If not, what would be a better one, and why?

ETA: Not to mention, this tomb is designed by an intelligent mind, not a random one, so it's likely the true door is NOT located in the first or the last few doors, to deliberately cause trouble for people attempting a sequential search.


Respectfully,

Brian P.

The relevant Computer Science theory is called the No free lunch theorem.
Go to Wikipedia - No_free_lunch_in_search_and_optimization article

[Jasdoif]

First of all, Serini's a rogue, so she can't be lawful, correct?The rogue class doesn't have an alignment restriction. You might be thinking of barbarians or bards.

[goodpeople25]

First of all, Serini's a rogue, so she can't be lawful, correct?

Second of all, what we've seen of her diary shows a bunch of hearts around Girard and a "meanie" caption for Soon -- which strongly implies A) a chaotic alignment B) a liking for Girard, if not a schoolgirl crush.

If I were seriously trying to break this puzzle myself I would definitely explore the hypothesis that Serini admired Girard and would incorporate some of his ideas into her dungeon.

Respectfully,

Brian P.
As said before me, rouges have no alignment restrictions.
Also which strip/book is this meanie caption in? There's a mean! Grr! caption for Kraggor though, (does that mean she's lawful?) Is that what you're thinking of?
Also if it is there, you don't need to be chaotic for a paladin (or just the person) to rub you the wrong way and being a rouge likely wouldn't have helped. (Well in general for paladins and rouges, we don't really know that much about them)

[pendell]

As said before me, rouges have no alignment restrictions.
Also which strip/book is this meanie caption in? There's a mean! Grr! caption for Kraggor though, (does that mean she's lawful?) Is that what you're thinking of?
Also if it is there, you don't need to be chaotic for a paladin (or just the person) to rub you the wrong way and being a rouge likely wouldn't have helped. (Well in general for paladins and rouges, we don't really know that much about them)

I may have mis-remembered. I thought it was soon but I suppose it could have been Kraagor.

I suppose rogues don't have an alignment restriction in 3.5E -- I just have a hard time imagining a lawful one.

Respectfully,

Brian P.

[Hopeless]

What if that isn't 121 doors but actually one immense door with a multitude of smaller doors leading into extradimensional space that restocks with fresh monsters (every time it shuts which would be every time they close whether at the end of the day or once shut before then)?

So they actually need to open that door to access the true dungeon leading to the Gate which I figure has the dwarven city at the very back, but back to the thread.

[Kish]

It was Kraagor. There is no indication that Serini ever showed hostility toward Soon in any way. "Chaotic rogue who admired Girard" is apparently backed entirely by forty percent stereotypes and sixty percent assumptions.

Which may not make it inaccurate--Rich has thus far portrayed every rogue he's shown as disappointingly-to-me stereotypical--but let's not talk as though it was actually established.

[Jasdoif]

It was Kraagor.Indeed. (http://www.giantitp.com/comics/oots0196.html)


There is no indication that Serini ever showed hostility toward Soon in any way.In fact, I'd say Serini seems too playful towards Soon to suggest hostility. (http://www.giantitp.com/comics/oots0276.html)

[Aeson]

"Chaotic rogue who admired Girard" is apparently backed entirely by forty percent stereotypes and sixty percent assumptions.
The sketch of the Order of the Scribble seen in Serini's diary in 196 (http://www.giantitp.com/comics/oots0196.html) arguably constitutes evidence that Serini admired Girard (although it'd probably be more accurate to take it as evidence that Serini felt that Girard was attractive in some way), as Girard, alone of the members of the Order of the Scribble, has hearts drawn around him.

[Ruck]

So I guess this is a problem less for computer science than it is for social engineering (https://en.wikipedia.org/wiki/Social_engineering_(security)) -- to research Serini thoroughly to figure out what the key to the puzzle is. Surely there must be one -- she would need a way to access the gate herself in time of trouble, so she must have had SOME way to bypass the decoys and go directly to the true entrance.

Unfortunately, the team best equipped to do this is the one with her diary - Team Evil. And even so, they haven't figured out any clues from it that would hasten their search. We could be here a while.

[Roland Itiative]

Unfortunately, the team best equipped to do this is the one with her diary - Team Evil. And even so, they haven't figured out any clues from it that would hasten their search. We could be here a while.

Does Team Evil even know Serini designed this dungeon? The explanation Xykon gives in the page where the diary is feature makes it seem like Kraagor actually guards this Gate. Sure, we know that's not the case, but maybe she kept the details hidden even from her diary.

[Quild]

Problem: We have 121 doors arranged in sequence.

I'd bet there are more doors than the ones we see.
And I'm not sure of what you call a sequence. There's a chance that there is no way to consider a door as the first one. Maybe there are doors on the two faces of the cliff.

[Ruck]

In truth, I have no idea one way or the other. Maybe there's a clue in the No Cure for the Paladin Blues strips (and bonus strips) where Xykon and Redcloak return to his old lair to retrieve the diary.

EDIT: It seems you may be correct. I had the PDF on my phone, so I checked. In #196, Xykon says:

:xykon: There were five other adventurers in the party. Each built a secret stronghold atop one of the five gates, while the sixth—the halfling—recorded the locations in code in her diary.

So it sounds like he's under the impression Kraagor himself built the gate. Not sure if that's a trivial detail or one that will play into Team Evil's failure to find it so far.

[Shining Wrath]

Is it possible that Serini would have considered the possibility of her diary falling into the wrong hands and deliberately included misleading information? For example, we know she liked Kragor but has him noted as a "meanie grrrr" in the picture.

[Fish]

There are a few issues with search optimization that could also come up. I've no idea if Serini would have taken these into consideration, but they're interesting to contemplate. First, there's monster CR; second, there's enemy psychology; and third, there's enemy frustration.

First, Serini would have doubtless been aware of XP in general, but she might not have considered it vis-a-vis dungeon preparation: how difficult the monsters were inside. The monsters' CR and XP value would certainly influence any attack strategy against the doors. There are, as I see it, a few ways it could be done:

All the monsters are the same CR. This would suggest that all monsters are equally difficult, more or less. There are only two outcomes: either the invaders are strong enough to defeat everything, or they're not strong enough to defeat anything. (Let's set aside variability for the moment; obviously not all creatures are necessarily equally difficult for all invaders; some invaders, such as Xykon, might find undead a piece of cake, but have difficulty with other creatures. So read "everything" as "almost everything.") However, it would also suggest that if all monsters are equally difficult, the enemy will soon reach a point where he doesn't get any XP from defeating anything, and he'll cease leveling up. Depending on how powerful the enemies here are, Xykon may be at or near this stage; he doesn't necessarily get more powerful by defeating things in the Hollow.

The monsters scale in CR. It's possible that the "easy" and "hard" monsters are randomly distributed throughout the Hollow, or that they are lined up neatly; either way, it represents a range of CRs. This would suggest two possibilities: the invaders can defeat something in the Hollow; or the invaders can defeat nothing. In the former case there is an optimal strategy: start with the things you can defeat and gradually level up until you can defeat the other things. This seems like a fairly foolish strategy for Serini to have used; she would be providing the means for defeating the dungeon in the dungeon itself.

The monsters get more difficult somehow. I don't know how Serini would manage to arrange this.

The size of the door is relative to the size of the danger inside. Big powerful enemies would likely choose doors proportionate to their own size; a halfling might hide the gate behind a door that's too small for most powerful invaders even to use, unless they magically diminished themselves. It doesn't seem like a particularly good strategy; anything with sufficient magic to shrink itself and fit through a tiny door probably also has other magic.

The second question is, "Against whom was Serini planning to defend?" If she were defending the Gate against random people who happened to stumble across the Hollow, fine; they'd probably not realize that a systematic approach was even required. They'd attack a few things, and they might even last a while, but they wouldn't necessarily be aware that there was something being defended. The other kind of invader is the kind that knows about the Gates, and further knows that a Gate is there. If you were really trying to keep a guy like that out, you'd look for a way to re-randomize the doors after they've been used. Then it wouldn't matter if you'd checked that door previously; it would still have an equal chance of being correct this time.

And the last question is this: "What do you do when you have an enemy more powerful than your monsters?" Such an enemy would, of course, be impossible to keep out by powerful monsters alone. If the invader wasn't getting any XP, he might grow bored and give up, or grow angry and attack randomly (maybe take out his frustrations on the statue), or try to hack the dungeon itself (bashing down the separating walls between monster lairs, using passwall to get around, whatever). I don't know if Serini would have planned against a powerful, persistent enemy like Xykon, one who knows the Gate exists, knows it's here, and has no natural lifespan to stop him from getting in.

So what's my opinion?

I think there is a way to solve this. After all, we expect the Order to come by to save the day; it makes a certain sense that the Order would have to have a smart way to choose correctly. Therefore, I think Serini's placement of powerful monsters is a misdirection. She built the dungeon to honor a fallen warrior, who prized strength; but she didn't say that it would require strength to solve. It may be a fool's game to bash through doors one by one. She herself wouldn't solve the dungeon that way.

[Ruck]

Is it possible that Serini would have considered the possibility of her diary falling into the wrong hands and deliberately included misleading information? For example, we know she liked Kragor but has him noted as a "meanie grrrr" in the picture.

Possibly, but I wouldn't rely on that too much given that we already know the locations of the five gates were recorded in code, and Xykon and Redcloak accurately cracked it. Were the code some kind of double bluff such that cracking it would reveal false information, I imagine that would have come up already. (Of course, it's easy to argue that since she knew where she built Kraagor's Gate, she could deliberately include misleading information there, but in that case, why not send anyone looking for the gates somewhere else entirely, a la Girard, instead of "right on top of the actual gate," which this has to be, given the statue of Kraagor?)

[Princess Tracy]

The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.

The Shell Game is absolutely the first thing that comes to mind, but it seems like... too much of a bluff. You only use three cups in the shell game, you don't throw out literally dozens. The Shell Game relies on the assumption that the game, as presented, is still being played and played fairly with no trickery on the part of the shuffler and that there's a decent chance of you picking out the right cup, or door in this case. Having a huge monster pit like this means that you really have to start applying logic to see where you're being gamed, to get into Serini's head, try to find something outside of the logic of simply rote checking of every door. The human (well, halfling) element is the only angle they can realistically analyse, and that necessarily means considering the possibility that the "rules of the game" (i.e. that the gate is behind ONE of the doors) is not being followed.

In short, there is an incredible risk that, if the gate or the true dungeon entrance were out in the open or connected to the statue (the only external feature of the dungeon not guarded by monsters), someone would almost certainly check for such a trick before resorting to checking all the dungeons one by one. Hell, just look on the forum, pretty much everyone's saying "hey, its the statue, check the statue" just on instinct even without necessarily remember it being foreshadowed at Azure City.

So either Serini honoured her intention of having a pure monster dungeon reflecting kragor's philosophy, or... well, if it is under the statue as everyone seems certain of already, I suspect it won't just be in the statue but like a mile down through the most dangerous epic level dungeon of all time that puts every other pit of beasts there to shame.

Oh wait, it was mentioned that the location of the gates is coded in Serini's diary somehow so maybe she has some love for cryptography. There could be a hint there but we need more information and some kind of cipher on the dungeon layout... *scribble scribble*

[dtilque]

figure out what the key to the puzzle is. Surely there must be one -- she would need a way to access the gate herself in time of trouble, so she must have had SOME way to bypass the decoys and go directly to the true entrance.

Perhaps it's a failure of imagination on my part, but I can't see any reason that Serini needs a backdoor access to the gate. I just can't think of any form of trouble where getting to the gate would help. After all, the only thing inside the gate is a rift in the fabric of space-time which leads to a world-destroying monster. What kind of trouble could that be a solution to?


My solution:

1. Choose a door but don't open it.
2. Convince Monty Hall the DM to open all the doors except two, the one you chose and the one with the car gate, thus revealing N-2 goats monsters.
3. Decide whether to keep your original door or switch (hint: switch)

Drawbacks: step 2.1 Defeating N-2 goats monsters.

[Roland Itiative]

Perhaps it's a failure of imagination on my part, but I can't see any reason that Serini needs a backdoor access to the gate. I just can't think of any form of trouble where getting to the gate would help. After all, the only thing inside the gate is a rift in the fabric of space-time which leads to a world-destroying monster. What kind of trouble could that be a solution to?

The Order of the Scribble fought a lot of people who wanted to use the Rifts for something or another, apparently. Also, they deliberately set up defences for each of the Gates, implying they expected them to be attacked by someone. Also also, they set up a separate device that shows the current state of every Gate to every member of the party, showing that they wanted to be forewarned to a future threat, if possible.

Four out of five of the members of the Order decided to actually start living very close to their Gates, so they wouldn't need such a backdoor, but Serini seems to have resumed travelling around the world afterwards, so a backdoor would be quite useful in case she needed to come back. But it would also be a big security risk, so maybe it really doesn't exist.

[Porthos]

In short, there is an incredible risk that, if the gate or the true dungeon entrance were out in the open or connected to the statue (the only external feature of the dungeon not guarded by monsters), someone would almost certainly check for such a trick before resorting to checking all the dungeons one by one. Hell, just look on the forum, pretty much everyone's saying "hey, its the statue, check the statue" just on instinct even without necessarily remember it being foreshadowed at Azure City.

This and entirely this.

I can check 1 thing or I can check 120+ things. Gee, I wonder which I am going to do first?

Haley and Roy put it best when they said: (http://www.giantitp.com/comics/oots0697.html)

:roy:: Plus the desert is so unbelievably large that anyone with more than two brain cells would know better than to try to search the whole thing.
:haley:: - which would lead them to concentrate their search right here again, just in case it WAS a double-bluff. The message would actually increase the chances of someone finding the Gate.

Substitute just a wee little bit and one could see how that line of thinking is going to make people check out the statue JUST IN CASE. If, if the doors were a gigantic shell game, in that the Gate is in none of them, having 120 of them is counter-productive since many people would lose paitence with the whole thing and look for a cheat code. Say, that conspicuous statue over there.

Heck, if I was leading a group, I'd check out the statue realtively quickly. If only to keep the folks in my party who were yammering about bluffs and shell games in my ear happy. :smalltongue:

[Roland Itiative]

This and entirely this.

I can check 1 thing or I can check 120+ things. Gee, I wonder which I am going to do first?

Haley and Roy put it best when they said: (http://www.giantitp.com/comics/oots0697.html)

:roy:: Plus the desert is so unbelievably large that anyone with more than two brain cells would know better than to try to search the whole thing.
:haley:: - which would lead them to concentrate their search right here again, just in case it WAS a double-bluff. The message would actually increase the chances of someone finding the Gate.

Substitute just a wee little bit and one could see how that line of thinking is going to make people check out the statue JUST IN CASE. If, if the doors were a gigantic shell game, in that the Gate is in none of them, having 120 of them is counter-productive since many people would lose paitence with the whole thing and look for a cheat code. Say, that conspicuous statue over there.

Heck, if I was leading a group, I'd check out the statue realtively quickly. If only to keep the folks in my party who were yammering about bluffs and shell games in my ear happy. :smalltongue:
That only works for debunking the assumption that the true path being related to the statue. It can still be inside a completely inconspicuous hidden passage
anywhere else on the mountain range, not an easily searchable place. So the right answer would not be wasting time on the 120 doors, nor with the statue, but actually combing the entire mountain range for secret rock door or something.

In that case, you get both the people who failed to see the misdirection at all and the people who saw the misdirection but thought the statue was part of the ruse suckered, both will waste their time with no chance of success.

[Porthos]

That only works for debunking the assumption that the true path being related to the statue. It can still be inside a completely inconspicuous hidden passage
anywhere else on the mountain range, not an easily searchable place. So the right answer would not be wasting time on the 120 doors, nor with the statue, but actually combing the entire mountain range for secret rock door or something.

In that case, you get both the people who failed to see the misdirection at all and the people who saw the misdirection but thought the statue was part of the ruse suckered, both will waste their time with no chance of success.

Think that is a bit much. Even Girard had a very conspicuous Enter HERE!!! for his dungeon once True Seeing was turned on.

I would also point out that while it might make sense "in reality" for somone to make it nearly impossible to find the Gate That Could Destroy the World, Rich is probably not going to make it as difficult. Be kinda anti-climatic if no one could get to the Doomsday Device, no? :smallwink:

So throw up roadblocks to delay things? Preferably in an entertaining manner? Absolutely. Make it so this is a gigantic waste of time? Call me slightly doubtful, though I am willing to admit I might be wrong.

[Tarqiup Inua]

I think this is closely related to the problem of choosing the door where to hide the prize.

One cheap way to get good result would be getting a large group of people, asking each of them to secretly choose a door they think noone else will. Those that choose the least-chosen door will be promised some sort of a prize. (but won't really get it, that would make your choice weak to a torture-based cryptography)

You could try something similar to get an educated guess where the best choice might be.

Of course, some might say that the least likely door will be the one where you shouldn't hide your prize at all, but that's the nice thing about this approach - it's only approximation and people that choose the door might already take this threat into account when they are participating in the survey.


Think that is a bit much. Even Girard had a very conspicuous Enter HERE!!! for his dungeon once True Seeing was turned on.

I am not sure I am looking at the strip correctly, but there seems to be unique combination of colour, structure, hinge, doorknob and their respective positions and amounts to each door.

(Maybe if there is a door that is somehow unique even among these combinations, that could be it?)

[dtilque]

Four out of five of the members of the Order decided to actually start living very close to their Gates, so they wouldn't need such a backdoor, but Serini seems to have resumed travelling around the world afterwards, so a backdoor would be quite useful in case she needed to come back. But it would also be a big security risk, so maybe it really doesn't exist.

But she doesn't actually have to get to the gate itself to defend it. Say she has some device/spell to notify her when someone opens the correct door. She (and any assistants) could just teleport to the statue area and then enter the same door. Following the intruders would be simple since the intruders have already defeated any monsters between themselves and the door. She just follows the path of dead monsters until she reaches the intruders. So there's no need to get to the gate via a secret passage or whatever. And, as you say, such a backdoor is a big security risk, so why have it?

[DaggerPen]

In some fairness, there is one fairly compelling reason to have some way to access the Gate outside of active siege - what happens if something goes wrong with the spell? This was epic experimental magic sealing a hole through which a god-killing abomination can reach - it's something that's getting quite a lot of stress testing without quite a lot of experience with that sort of magic. I know that if I were making it, I'd want a way to patch it super quick if it turns out there was a flaw in the binding.

[dtilque]

I know that if I were making it, I'd want a way to patch it super quick if it turns out there was a flaw in the binding.

She didn't make the gate or seal the rift and has no magic that would help fix those things. Still don't see any need for special access to the gate.

Remember, it's a big security hole to have such a backdoor. If someone discovers it, then all the other stuff you've done (monsters, multiple doors, whatever) has been a waste of time. You need to have an overriding reason to have one, and I just don't see one.

[theasl]

Perhaps it's a failure of imagination on my part, but I can't see any reason that Serini needs a backdoor access to the gate. I just can't think of any form of trouble where getting to the gate would help. After all, the only thing inside the gate is a rift in the fabric of space-time which leads to a world-destroying monster. What kind of trouble could that be a solution to?


My solution:

1. Choose a door but don't open it.
2. Convince Monty Hall the DM to open all the doors except two, the one you chose and the one with the car gate, thus revealing N-2 goats monsters.
3. Decide whether to keep your original door or switch (hint: switch)

Drawbacks: step 2.1 Defeating N-2 goats monsters.

Dang, now I want to see a dungeon full of goats.

[Shining Wrath]

Serini presumably has magic notifying her the gates are being messed with.

I wonder if she's still alive?

[Mx56]

Serini presumably has magic notifying her the gates are being messed with.

I wonder if she's still alive?
Assuming Rich uses the lifespans given in core 3.5, it's entirely possible. Per page 109 of the PHB, Halflings are venerable at 100 years old and can live up to 200 (100+5d20 years, mean would be 152.5, I think), it's apparently been 66 years since Soon first discovered a rift (http://www.giantitp.com/comics/oots0276.html), so if Serini was in her 30s then (reasonable for a high-level halfling rogue), she might just about be pushing venerable now.

[Quild]

The answer, of course, is that it isn't behind any of the doors. It's inside Kraagor's statue. Serini's a rogue, and, as foreshadowed in the Azure City battle, it's a shell game.
Far from "of course", this is extremely unlikely to be the solution, for reasons that have already been gone over thoroughly in the main #1039 discussion topic:



I predict that the Gate won't be beneath Kraagor's statue, for the same reason it wasn't beneath Girard's statue: you do not put your double bluff outside your defenses.

Because if you do - and someone guesses the con: it's defenseless.

It is perfectly possible, of course, that none of the doors lead to the Gate. Why make the gate accessible at all? Clearly some pretty major landscaping has been involved, so why not hide the Gate inside solid rock, wrapped around with dungeons full of monsters, so you can't even dig down to it without setting them all off.

Or, you know, maybe Kraagor's statue is the doorway into the real boss dungeon, complete with stuff that dumps the contents of all the other doors down into it the moment the doorway is opened.


The problem here is that if Serini did indeed hide the Gate in the end of one unique passageway, then it makes a big part of the defenses useless.

Is Serini sure that no one will have a clue of which door is the right one?
Or have a lucky shot?
Or does not care crawling the caves one after one losing a few days? (Xykon waited months in Dorukan's dungeon and in Gobbotopia and spent years decyphering the diary)

The point of Girard's phantasm was not to stop anyone, but to delay them so Draketooth family can act.
Thinking of it, Dorukan's glyph door probably also was made to delay an evil party.
The alive paladins were not the last line of defense in Azure City.
And if Lirian's virus sure was a big help to her defenses, that was not all...
And the Order of the Scribble had ways to know if another Gate had fell, so they could do something about it.

But the number of doors can not be more time consuming than having a very big dungeon which each and every of the monsters that are split in several dungeons. Nor more dangerous.
I don't get it.
Also I wonder if you can go from the end of a dungeon to another dungeon with "Meld into Stone"

[Shining Wrath]

The problem here is that if Serini did indeed hide the Gate in the end of one unique passageway, then it makes a big part of the defenses useless.

Is Serini sure that no one will have a clue of which door is the right one?
Or have a lucky shot?
Or does not care crawling the caves one after one losing a few days? (Xykon waited months in Dorukan's dungeon and in Gobbotopia and spent years decyphering the diary)

The point of Girard's phantasm was not to stop anyone, but to delay them so Draketooth family can act.
Thinking of it, Dorukan's glyph door probably also was made to delay an evil party.
The alive paladins were not the last line of defense in Azure City.
And if Lirian's virus sure was a big help to her defenses, that was not all...
And the Order of the Scribble had ways to know if another Gate had fell, so they could do something about it.

But the number of doors can not be more time consuming than having a very big dungeon which each and every of the monsters that are split in several dungeons. Nor more dangerous.
I don't get it.
Also I wonder if you can go from the end of a dungeon to another dungeon with "Meld into Stone"

Meld Into Stone doesn't allow for travel. It's only 3rd level.

[Jasdoif]

Dang, now I want to see a dungeon full of goats.Take the blink dog's (http://www.d20srd.org/srd/monsters/blinkDog.htm) entry, change the bite attack into a gore attack, and adjust and refluff the rest into a blink goat.

On top of the obvious potential of a dungeon full of teleporting goats, now it's entirely justifiable to have telegoats appear on the other side of any door as it's opened.

[DaggerPen]

The golden conversation about goats aside, I do think that there's a fairly practical limit to how many monsters you can fit into a linear dungeon. Some may squabble or hunt the others for food, only so many can attack at a time anyway, etc. Throwing the doors into the mix allows for more monsters to be fought ultimately unless a party is very lucky, and that's barring the potential for stuff like "pull a switch behind this door before going to the other one," which is a twist I am certainly not ruling out.

I do, however, definitely not think it's under the statue, for the reasons others have given above. Having that many doors to search makes trying all the clever/desperate/guide dammit solutions super tempting.

[Roland Itiative]

The golden conversation about goats aside, I do think that there's a fairly practical limit to how many monsters you can fit into a linear dungeon. Some may squabble or hunt the others for food, only so many can attack at a time anyway, etc. Throwing the doors into the mix allows for more monsters to be fought ultimately unless a party is very lucky, and that's barring the potential for stuff like "pull a switch behind this door before going to the other one," which is a twist I am certainly not ruling out.

The thing is, why put the doors besides each other, when the next door could be right at the end of the previous door's dungeon, effectively making a single giant corridor? One way leaves a luck factor that allows for a possibility of 99% of the dungeon to be bypassed, the other guarantees an invader will have to face every single encounter, with nothing left to chance.

Monsters migrating from one area to the other doesn't really work as an explanation, as they could just as simply leave their own door and go hunting inside another.

[Kish]

From Oona's words, Serini seems to have managed some way for the monsters to keep respawning indefinitely, like it was a video game--not leave, not run out of food and eat each other and die out.

[Jigsaw Forte]

If this were my game? I wouldn't put the gate behind a door... or at least, no single door.

I'd make the gate switch doors at random. That way, I wouldn't need to remember a certain door, but just the algorithm that I used. It'd be pointless to rubber-hose it out of me, unless you could somehow get me to tell you the algorithm. Even if you HAD my algorithm, you'd still need to be smart enough to use it.

Since 121 is just 11 squared, let's run with that idea and make it 11 cubed. If so, there's actually 1331 rooms in total behind those doors, and only ONE of them contains the gate. The rest contain a random spattering of monsters, traps, and other lovely dead ends. This also has the bonus of boosting the random odds from 1:121 to 1:1331. Instead of being "unlikely", we're now in Vegas territory.

It also means we can refer to each room as [X][Y][Z]. Rooms range from [0][0][0] to [10][10][10]. (Yes, we ARE starting our count at 0! We're couching this as a CS problem, right?)

Here, Xykon's "paint logic" starts to make sense. Mark each room as you traverse it, and if you see THAT room again, you know it's pointless to explore it further. Sooner or later, Xykon will run out of unique rooms, and thus find the gate.

Of course, Xykon's logic requires the rooms to have static coordinates. If the rooms change their coordinates, he's going to have a bad time.

... You're asking if I've seen a certain series of Canadian horror films? Why do you ask? ;)

[pendell]

This raises an interesting question for the DMs reading this : Monster ecology. How do you set up a self-sustaining dungeon if you don't want to handwave away the various logistics requirements? Has anyone written any articles on the subject?

You're going to need a primary food source (fungae? some underground crop?), something to eat that primary source, and secondary hunting creatures which will prey on the primary and, as a bonus, kill any intruders who happen to blunder in.

Taking a leaf from Salvatore, the most likely way to set up such a dungeon would be to have some intelligent caretakers (Drow? Mind flayers?) who can set up the various other species and manage them as a kind of "farm". You would need some kind of precious resource for this civilization to exploit (mithril ore? some medicinal plant that only grows here?) in order for them not to move and to keep sustaining the dungeon.


Serini can't use this, I think, because as others have said, an intelligent caretaker or group of same could be tempted to seize the gate themselves. Push that line of thought far enough and you need caretakers not unlike Soon's Sapphire Guard or Girard's family -- a band of sapients sworn to the defense of the gate.

However, we've already seen that mechanism at Girard's gate and Soon's Gate, so presumably Serini will have to try something different. How do you automatically attract high-level monsters and keep them in the dungeon, neither escaping nor killing each other nor overpopulating?

Of course you can handwave these issues with "a wizard did it"... but where's the fun in that?

Respectfully,

Brian P.

[dancrilis]

However, we've already seen that mechanism at Girard's gate and Soon's Gate, so presumably Serini will have to try something different. How do you automatically attract high-level monsters and keep them in the dungeon, neither escaping nor killing each other nor overpopulating?


Another issue that Serini has is that she presumedly does not want to attract adventurers (even when there are magic items within).

As for ecology I think this can work farily well by allowing ecology to play out naturally - when you have fungus and fungus eaters in a forest you tend not to run out of fungus as ecological balance will be achieved same with wolves and deer, etc - there is no reason that ecological balance cannot be achieved within the dungeons.

[Sniffnoy]

121 doors to check, huh? Well, obviously, the answer for our heroes to put themselves and the dungeon in a coherent quantum superposition, and then they can (with probability over 99.9%) solve it opening only 8 doors. :P

[Hopeless]

Do you get xp for summoned monsters?

Easiest way to restock her dungeon is that every time they shut the door or say the sun rises or sets its automatically replenished so no need to feed them since it requires the door to be opened to summon them.

Might prove a problem as if they do kill the visitors and leave but once killed summoned monsters disappear don't they?

[Mx56]

Do you get xp for summoned monsters?
If they're summoned by another opponent, no, the original opponent's CR includes their ability to summon. If they're summoned by a trap, on the other hand, then I think you get experience as it counts as beating a trap (and traps have their own CR and hence experience for beating them), though I haven't DMed in a while so I forget exactly how that works. The latter case more relevant to what you suggest there.

[Kish]

I doubt very much that Serini's defenses center around endlessly respawning summon traps. Superficial appearances aside, that would mean she'd guarded her gate in a way thematically appropriate for a conjurer, not a barbarian or a rogue.

[Roland Itiative]

We have also seen the process of the dungeon being made, and it included the monsters being lowered into the dungeon, and some of them going out of control. No magic seems to have been involved in that.

[Hopeless]

Kraagor's Tomb or the Epic level ruse Xykon & co are facing outside?

What if those doors are all red herrings designed to summon monsters once they're opened so no matter if they explore all of them they can't find the true entrance to Kraagor's Tomb as its all designed to keep their attention away from the true entrance.

I think that cliff face isn't a multitude of doors its one enormous door intended as a test of exemplary strength to reach the true entrance to Kraagor's Tomb.

I'm actually hoping this is the case so it would be the MITD who solves this as being an innocent and naïve it would see through this ruse and when asked will open the true door revealing the truth behind Serini's opening defence.

There's also the possibility Xykon will lose his patience and just blast the cliff face revealing the ruse for what it is, but I'm hoping whatever the solution is it'll be as usual very funny!

[dtilque]

This raises an interesting question for the DMs reading this : Monster ecology. How do you set up a self-sustaining dungeon if you don't want to handwave away the various logistics requirements? Has anyone written any articles on the subject?

You're going to need a primary food source (fungae? some underground crop?),

What you're really going to need is a primary energy source. Aboveground, it's the light from the sun, but below? Either the Earth's heat (not really a good source) or something magical like mana seem to be the only options.

There's also the problem that as you go up a level in an ecology, everything decreases by at least an order of magnitude. That is, there's at least an order of magnitude more plants than plant-eaters, and an order of magnitude more plant-eaters than carnivoires. But that's not the way monsters in the typical dungeon are allocated. So doing this is going to result in a very atypical dungeon.

[dancrilis]

Thinking about ecology.
If Serini has recruited some 'cast's as cleric' creatures to operate as management she could solve the food issue.
Lammasu, Androsphinx, Djinni, Ghaele, Trumpet Archon - can all cast Create Food and Water - so having some of them on staff would possible solve some of the food issues.
All are largely trustworthy.

[Roland Itiative]

I think the ecology talk is reading a bit too much into things :P Fantasy worlds are notable for not giving any hard thought to that, and so far the OotS has been no exception, there's no reason to think this one dungeon will.

[DaggerPen]

I think the ecology talk is reading a bit too much into things :P Fantasy worlds are notable for not giving any hard thought to that, and so far the OotS has been no exception, there's no reason to think this one dungeon will.

Well, the thread title did specifically invite math/CS types. Asking us not to overthink this sort of thing is like asking a fish to climb a tree :smalltongue:

[dancrilis]

Well, the thread title did specifically invite math/CS types. Asking us not to overthink this sort of thing is like asking a fish to climb a tree :smalltongue:

Ah yes the mangrove killifish capable of suriving for over a month outside water and climbing trees (or so it is said - not personally tested, happy to accept marine biologist input).

[DaveMcW]

This raises an interesting question for the DMs reading this : Monster ecology. How do you set up a self-sustaining dungeon if you don't want to handwave away the various logistics requirements? Has anyone written any articles on the subject?

Check out the Dungeonbred template. (Dungeonscape. Jason Bulman and Rich Burlew. 2007. page 112)

[Cracklord]

This raises an interesting question for the DMs reading this : Monster ecology. How do you set up a self-sustaining dungeon if you don't want to handwave away the various logistics requirements? Has anyone written any articles on the subject?

You're going to need a primary food source (fungae? some underground crop?), something to eat that primary source, and secondary hunting creatures which will prey on the primary and, as a bonus, kill any intruders who happen to blunder in.


MSTK rule aside? Sure. But it's self-defeating.

Look, any proper DnD player loves dungeons. I do, you do, presumably the Giant does as well (and good for him). But a few seconds consideration will leave you pretty sure there is no good reason for the silly things. For starters, the average D&D game world is quite frankly incapable of the technology or manpower needed to build vast underground complexes, short of powerful magic. For comparison consider real world history; aside from a single underground city in Turkey and a couple of pyramids and tombs, the ancient world took a pass on underground life. Just to accept their existence you need the age old excuse 'A wizard did it', much less to get on to the actual problem of why anyone would bother.

Let's accept the presumption that 'Wizards can magic it up and they do it because its defensible' - accepting the existence of magic as a common method for problem solving becomes a bit lame considering that we are now talking about a world with teleport and burrowing and ethereal travel; being underground is actually a liability since its harder to escape and people can drop the roof onto you, not to mention the incredible costs involved in doing it even if magic is available. And what do you get in return? Not much.

Don't even get me started on the Underdark.

Lets say you have completed your herculean engineering task, or the cave structures are naturally occurring and you've just put in a few doors and some furniture despite these problems. So where do you go from there? Once you populate it (assuming they all eat one another), you have to keep it going, which is harder then it looks. Entropy - the simple fact that it requires energy inputs to maintain an open system, and a closed system only degrades over time. If there is to be Life, there has to be some way of getting energy into the system. For us here on the surface, that’s not even a problem: the Sun shines energy down on the surface all day. Obviously, that's not the case in dungeons. And you have to assume that new energy is being brought in to this closed system some way or another.

But for those who live in the dark realms: be it sewers, the ocean floor, or the classic dungeon complex; there has got to be a source of energy in the D&D world that just doesn’t exist in ours. You don’t see lush forests in marine trenches because the energy inputs just aren’t there. So what? More magic? By now the rules are so arbitrary that you're wishing you never asked yourself the question.


What you're really going to need is a primary energy source. Aboveground, it's the light from the sun, but below? Either the Earth's heat (not really a good source) or something magical like mana seem to be the only options.

There's also the problem that as you go up a level in an ecology, everything decreases by at least an order of magnitude. That is, there's at least an order of magnitude more plants than plant-eaters, and an order of magnitude more plant-eaters than carnivoires. But that's not the way monsters in the typical dungeon are allocated. So doing this is going to result in a very atypical dungeon.

To compound on this, particularly in a dungeon, resources are limited, and thus only a finite number of creatures can be supported on any particular diet within any area. Now DnD usually has every single species that Earth has (Horses and elephants, bears and walruses, whales and snakes, mosquitoes and finches, not to mention animals that are technically extinct like sabertooth tigers and tyrannosaurus rex's), and then it has thousands of additional monsters, many of which are not only top predators, but apex predators. There is a reason the area around the Lonely Mountain was called 'The Desolation of Smaug', and it's because feeding such a big creature has dire effects on the local environment, or anywhere within a days flying distance.

That’s hard to manage. Remember that to support a single top predator requires a huge amount of energy inputs.

Fungus is the traditional answer, thanks in part to R.A Salvatore. Lets go with that, although it doesn't really make much sense.

You want a bigger problem? Bad things happen to a species if there aren’t many instances of them in the world. Cheetahs can all accept skin grafts from each other and are all weak to the same diseases. Seriously, all cheetahs could go extinct next year, their existence is that fragile. Much as George R. R Martin is prone to ignoring the effects of incest, that's not common in fantasy. And yet there are numerous species that are less well represented in the world than cheetahs, so why don’t they have the same problems?

Now I'm sure some of them do. Many times a dungeon will contain a 'unique monster' that the player characters will go and kill. That’s an extinction event right there. But some of them do not, and the reason is that DnD has a wealth of completely magical ways to keep a lineage from drying up. Even if a creature can’t find a mate of its own kind anywhere in the world (a likely distressingly common state of affairs) there’s always the planes and the realms of sorcery. As a matter of last resort, humans, dragons and Demons (as well as celestials, technically) can successfully reproduce with anything (there are no half-dwarf half elves, but humans can have viable offspring with either, in an old campaign module there was a half dragon shambling mound (plant base!), and the half-fiend and half-celestial templates speak for themselves).

Your best bets, therefore, are monsters that don't require a stable population (can successfully reproduce with whatsoever they encounter), and get the vast majority of their sustenance through magical means and only need small influxes of real food (from their human food). The classic example, of course, is Vampires. They consume much less in food energy than they use up in maintaining their undead existence. Most of their energy is actually siphoned off the negative energy plane, and the drinking of human blood is presumably just a symbolic evil act that they need to perform in order to keep that juice flowing, and to reproduce. Zombies are actually perpetual motion machines, when they eat your brain it's in fact (like vampires) their reproductive process, creating more of their kind (some people say it's enough for them to bite you, but that's getting them confused with vampires. Wendigoes, it's the other way around (or maybe you have to bite another person, but hopefully that will never be relevant)).

In the same way, Mindflayers are not sustained not by the nutritive value of brains (unlike those robots in the Matrix), but by their own psychic powers. They eat the brains of intelligent creatures to keep their psychic powers sharp, and they keep their psychic powers sharp in order to survive day to day without eating enough calories to keep themselves alive in the normal way. So they're all good.

Other species don't care one way or another. Plenty of creatures are made rather than born and appear in the world fully formed. Golems, elementals, and most outsiders simply don’t need a stable pool of Golems, outsiders or elementals in order to maintain genetic diversity, and they don't really eat either. They don’t even have genetic diversity. So they have the distinction of being the most common encounters in my dungeons. However, that gets old really fast.

Should I go on? Because I can go on.

[Deliverance]

121 doors to check, huh? Well, obviously, the answer for our heroes to put themselves and the dungeon in a coherent quantum superposition, and then they can (with probability over 99.9%) solve it opening only 8 doors. :P
While that is the problem as outlined by the OP, it must be noted that we don't know how many doors the hollow has.

In #1039 (http://www.giantitp.com/comics/oots1039.html), the one picture where we've seen doors, we only see part of the hollow. it clearly continues both to the right and left of what is shown in the picture, and there might be more doors there, so the ones we see only provide us with the minimal number of entrance doors in the hollow.

There might be a lot more.

[Sniffnoy]

While that is the problem as outlined by the OP, it must be noted that we don't know how many doors the hollow has.

In #1039 (http://www.giantitp.com/comics/oots1039.html), the one picture where we've seen doors, we only see part of the hollow. it clearly continues both to the right and left of what is shown in the picture, and there might be more doors there, so the ones we see only provide us with the minimal number of entrance doors in the hollow.

There might be a lot more.

Yeah, I know, but the joke worked better if I assumed a specific number of doors.

[JessmanCA]

I'm not quite sure that's fair. While you can map the sequence that way, it definitely has an impact on the outcome. If the correct door is #37 (say), starting a sequential search at 1 will get the correct answer in 37 tries, starting from 30 will give it to you in 7 tries, and starting on 1 and searching only primes gives you 12 tries. So it seems a false equivalence, to me.

Surely this is true, once you know the solution. But how can you know to start from 30? What if you started from 40? Then it'd take 119 tries to get to 37 (sequentially you'd try all of the 121 doors except 38 and 39). Not knowing the solution or any clues, there is no better way to search than random number generation. Perhaps if they could split up and conquer more than 1 door at once.

I would suggest however searching every other, or every third door, and using divination or brute force to try and peer into the nearby tunnels.

[pendell]

Here's an out-of-the-box idea:

A wizard can have a familiar, right? Make a strain of bacteria your familiar. Disperse in the tunnels and allow them to multiply until they've covered the entire area; use their senses to find the gate.

If bacteria won't do it, how about the demonic cockroaches? Or some kind of self-replicating golem a la a Von Neumann probe (https://en.wikipedia.org/wiki/Self-replicating_spacecraft#Von_Neumann_probes)? As you can see, this device was originally intended for exploring and mapping interstellar space. If it can map THAT environment, a single planetary valley should present a trivial challenge.

Respectfully,

Brian P.

[JBiddles]

Assuming Serini has some way of sorting out food and water etc., could she have gone for the true Kraagor approach as her final line of defence? Gather enough Epic monsters in one place, and not even Xykon is getting through.

Serini was "never one to settle down". She could have spent the past few decades trawling the earth for the biggest, nastiest monsters around.

Of course, that leaves the problem of how the story would proceed.

[Shining Wrath]

Assuming Serini has some way of sorting out food and water etc., could she have gone for the true Kraagor approach as her final line of defence? Gather enough Epic monsters in one place, and not even Xykon is getting through.

Serini was "never one to settle down". She could have spent the past few decades trawling the earth for the biggest, nastiest monsters around.

Of course, that leaves the problem of how the story would proceed.

The horde of monsters kill Xykon and Redcloak - and then talky man talks his way past them. That would be anti-climactic in a few ways, but it represents Roy thinking his way past a problem.

[DaggerPen]

Here's an out-of-the-box idea:

A wizard can have a familiar, right? Make a strain of bacteria your familiar. Disperse in the tunnels and allow them to multiply until they've covered the entire area; use their senses to find the gate.

If bacteria won't do it, how about the demonic cockroaches? Or some kind of self-replicating golem a la a Von Neumann probe (https://en.wikipedia.org/wiki/Self-replicating_spacecraft#Von_Neumann_probes)? As you can see, this device was originally intended for exploring and mapping interstellar space. If it can map THAT environment, a single planetary valley should present a trivial challenge.

Respectfully,

Brian P.

I can't tell if you just completely ignored the premise of this thread, or exemplified it.

[pendell]

I can't tell if you just completely ignored the premise of this thread, or exemplified it.

I didn't ignore the premise -- I'm the OP, remember? The conclusion several pages ago is that, because we don't get enough information from exploring any one door, there's no use to a particularly clever search strategy depending on which doors we choose. So the next plan , I think, would be to increase the number of searchers. A geometrically increasing number of searchers, such as a virus or other rapidly-multiplying organism, seems like a reasonable choice. It's not that dissimilar from modern crowd-sourcing projects such as Planet Hunters (https://www.planethunters.org/), which relies heavily on volunteer labor to find planets.

Respectfully,

Brian P.

[DaggerPen]

I didn't ignore the premise -- I'm the OP, remember? The conclusion several pages ago is that, because we don't get enough information from exploring any one door, there's no use to a particularly clever search strategy depending on which doors we choose. So the next plan , I think, would be to increase the number of searchers. A geometrically increasing number of searchers, such as a virus or other rapidly-multiplying organism, seems like a reasonable choice. It's not that dissimilar from modern crowd-sourcing projects such as Planet Hunters (https://www.planethunters.org/), which relies heavily on volunteer labor to find planets.

Respectfully,

Brian P.

... whoops. >_>;;

Well in that case, I tip my hat to you

[pendell]

Latest strip, 1040 (http://www.giantitp.com/comics/oots1040.html).

Looks like Redcloak and I are on the same wavelength :smallamused:. But as we've discussed, Xykon is actually right about this -- there IS no more efficient way than random guessing, knowing what we know now.

Respectfully,

Brian P.

[Sonata Arctica]

For such ellaborated problem, there always is Haley's Conjecture: "Why bother hiding the gate in one of the doors when hiding it in none of them is a win-win?"

R: The gate is inside the statue. At least the first thing I'd do is destroy that.

[dancrilis]

For such ellaborated problem, there always is Haley's Conjecture: "Why bother hiding the gate in one of the doors when hiding it in none of them is a win-win?"

My suspicion?
The gate is in the dungeon and the they know this - the coordinantes show them that the gate is in the center of the complex, which is why they have to explore it as they cannot simply phase through it/teleport etc.

Playing such a shell game would be one of the first things Redcloak (or Xykon for that matter) thought of I imagine.

[Fish]

My suspicion?
The gate is in the dungeon and the they know this - the coordinantes show them that the gate is in the center of the complex...
This isn't a geocaching expedition; they don't have GPS. I doubt they are using anything as accurate as "coordinates." First, there would be no reason for Serini to provide anything so useful, even in her diary. If she left any clues to the gate's location, it would probably be in a more general sense: "beyond the polar sea, past the craggy mountain, west of the canyon, and two days from where we stopped at that inn that was serving the smoked halibut."

[Sniffnoy]

Latest strip, 1040 (http://www.giantitp.com/comics/oots1040.html).

Looks like Redcloak and I are on the same wavelength :smallamused:. But as we've discussed, Xykon is actually right about this -- there IS no more efficient way than random guessing, knowing what we know now.

I'm telling you, quantum superposition is the way, get it done in 8! :P

[dancrilis]

This isn't a geocaching expedition; they don't have GPS. I doubt they are using anything as accurate as "coordinates." First, there would be no reason for Serini to provide anything so useful, even in her diary. If she left any clues to the gate's location, it would probably be in a more general sense: "beyond the polar sea, past the craggy mountain, west of the canyon, and two days from where we stopped at that inn that was serving the smoked halibut."

They appeared pretty close to the Girard's gate, do you assume that they location they got was 'in the desert north of some country that probably doesn't exist anymore'.

Her diary was not encoded - the coordinates were. We know that Girard gave Soon an area of approximatly 1 square mile to search, and that the trap was located right in the middle of that (he even used the word coordinates for them).

As such I am assume that Xykon and Redcloak have roughly similiar location information for the Kraagor's gate - and given the amount of doors and how small a 1square mile area is I think it likely that the gate is in the complex.
Leaving it outside means that anyone that looks for the gate outside the complex (to rule it out) may find it easilly - setting defences for your house and leaving your TV in your front garden is not a safe way to prevent people stealing your TV.

[Porthos]

For such ellaborated problem, there always is Haley's Conjecture: "Why bother hiding the gate in one of the doors when hiding it in none of them is a win-win?"

R: The gate is inside the statue. At least the first thing I'd do is destroy that.

As said by too many people, including myself, having a statue out in the open makes it MORE probable to be searched than not searched. Haley even said as much when talking about Girard's Gate in the desert.

And let's not forget, for all of the people who quote the shell game in Azure City..... The real Xykon was invisible. Thus the statue completely fails the test in all regards. It's just as much a shell in the shell game as all of the doors.

(and, yes, after about Door #5, I'd be checking it too, just in case)

[littlebum2002]

R: The gate is inside the statue. At least the first thing I'd do is destroy that.

And that's why it's almost certainly not in there

[Fish]

They appeared pretty close to the Girard's gate, do you assume that they location they got was 'in the desert north of some country that probably doesn't exist anymore'.
No, I assume that the location that Team Evil got was whatever location that Serini recorded. Serini may have recorded or encoded them as coordinates, provided that Girard had given them all to her, but Xykon never calls them by that name (at least that I can tell). Xykon twice says they are (http://www.giantitp.com/comics/oots0196.html) "locations" (http://www.giantitp.com/comics/oots0197.html) rather than "coordinates." Serini could have said "the big plain ringed by mountains and west of the winding canyon," or similar.

Based on what we know, I'd be surprised if Girard's methods were so precise as to say "the exact, precise location of the Gate is right here, plus or minus a meter" or, as you previously said, "the gate is in the center of the complex." Girard was the one keeping track of all this stuff, since he had two levels of ranger (http://www.giantitp.com/comics/oots0694.html). Girard also characterized his role as the mapmaker, which suggests that his information is a skill, not a spell. Even so, any "coordinates" he supplies are likely to be within the range of accuracy of his map (however created) and his tools. Let's give Girard the benefit of the doubt and assume he had been continuously his taking his locations with an accurate modern sextant; let's assume he took measurements from the map with accurate modern standardized rulers (and Xykon had rulers in the same scale and to the same tolerance). The very best accuracy I would expect to see from this method, based on the accuracy of a professional sailor using a modern sextant, would be precise to within about 200 meters. For Girard, who had only 2 levels of ranger, maybe 500 meters. Pretty accurate, but not GPS-quality accurate to the point where you could stick a pin in the center of the Hollow and say "right here."

If Xykon had had precise coordinates, he wouldn't have needed to invade Azure City at all; he could have teleported to the throne room. Instead, he flew on an invisible dragon. Azure City wasn't protected against scrying or teleportation like Dorukan's gate had been; Xykon could have popped right in. This doesn't make it impossible for there to be super-accurate GPS-quality coordinates, but Xykon's failure to teleport doesn't make the coordinate theory look good.

[Roland Itiative]

Since the comic discussion thread is bound to become hectic soon, and this thread is still active, I thought it would be better to bring this discussion here.

I was writing a more elaborate post here, but really, it was too complicated and didn't make the point in a better way than this.


Let's look at a small sample, shall we? Ten doors, numbered 0 to 9. Team Evil picks a door, opens it. They have 1/10 of chance of succeeding. If they fail, MitD picks 3 doors. They have a 3/9 of chance of marking the right door (they pick 3 out of a total of 9, since Team Evil already eliminated one door).

Next night, Team Evil picks one out of 6 doors. But their chance to succeed is not 1/6 this time. They only have a 1/6 chance assuming MitD failed last turn, a completely independent chance of 6/9. When you have two independent events, and you need them both to happen at once, you multiply the probabilities. (1/6)*(6/9) is 1/9, exactly the same probability Team Evil would have to find the right door on this night had the MitD not interfered before.

Why is that? It doesn't matter that the MitD is reducing the choice pool, but that they aren't increasing the amount of information available. Team Evil will not make a more informed choice the next day because of the MitD, so their odds of success should not, and are not, any greater.

Interestingly enough, for each night the MitD keeps doing that, they are getting increased chances because of Team Evil. Team Evil's actions give the MitD extra information about a door, and increases the MitD's odds (similar to the Monty Hall problem), but the reciprocate isn't true. Both groups are playing the game by different rules, and have different effects on each other.

[dtilque]

We actually know what form the coordinates for Girard's Gate that Roy received from Soon were. First of all, Roy was using a sextant, which pretty much means that they were lat-long. Secondly, when they got close to it, he said it was somewhere in a square mile area. If the lat-long is only to minute-precision, that is, degrees-minutes but no seconds, that gives an area of about 1 square mile. It wouldn't be a perfect square -- one nautical mile north-south and somewhat less than that east-west. But it's close enough to call it a square mile.

Those coordinates were not what Serini recorded in her diary, but it's probably best to assume all gate coordinates are given that way until proven otherwise.



If Xykon had had precise coordinates, he wouldn't have needed to invade Azure City at all; he could have teleported to the throne room. Instead, he flew on an invisible dragon. Azure City wasn't protected against scrying or teleportation like Dorukan's gate had been; Xykon could have popped right in. This doesn't make it impossible for there to be super-accurate GPS-quality coordinates, but Xykon's failure to teleport doesn't make the coordinate theory look good.

The throne room was protected from scrying and likely from teleportation too. But if they were given coordinates as above, the gate could have been almost anywhere in the city. Hence the subterfuge they pulled with Miko to find its exact location.

Now Kraagor's Gate coordinates would actually describe a somewhat smaller area. Still a nautical mile north-south, but much less than a mile east-west. That's because it's in the arctic where lines of longitude get closer together when you get to high latitudes. But I think it may be that, for that gate, she included other information, such as it being near Kraagor's statue, that would narrow things down quickly. Why would she give more info for this gate? It's the one she set up the defenses for, so it's probably in her diary somewhere, if not right next to the coordinates.

[Fish]

We actually know what form the coordinates for Girard's Gate that Roy received from Soon were. First of all, Roy was using a sextant, which pretty much means that they were lat-long. Secondly, when they got close to it, he said it was somewhere in a square mile area.
Technically, that only tells us how Roy found his way there, not how Girard recorded them (or would have recorded them, had he done so; Roy was looking for a location that had been rolled randomly). That is a reasonable range of error for such a device. (Of course, we would have to account for Girard's error and Roy's as well; I won't go any further because perhaps Roy already accounted for this.)

But if they were given coordinates as above, the gate could have been almost anywhere in the city.
You misunderstand: I agree with you.

The point I was making is that it's unlikely that the coordinates are so precise, so narrow and so accurate that we can say the point lies in the direct center of the hollow (which was the claim). As you say, knowing the coordinates are somewhere within a square mile is insufficient for that. And so I said: if Xykon did have enough precision to say that, he would have had enough precision to teleport directly to the throne room — and he's an epic sorcerer, so he'd stand a good chance of beating whatever anti-teleportation defense was in place.

[Fuzzypickles]

Serini is a rogue, but she built this gate to honor Kraagor. I think having the whole thing be misdirection would be undermining the point of this gate and what Kraagor believed in. There are no tricks. It is a pure test of stamina and combat. Also from a narrative perspective, I think Girard's Gate had enough bluffing and illusions where we don't need to see another gate defended by a big bluff.

I'd say there's no better strategy than brute force, which is quite thematic.

[DaggerPen]

If nothing else, even if he couldn't beat the anti-teleportation wards, I imagine he could teleport 100 feet up outside the wards and barrel down nicely.

[dancrilis]

The point I was making is that it's unlikely that the coordinates are so precise, so narrow and so accurate that we can say the point lies in the direct center of the hollow (which was the claim).
No it wasn't.

The stated suspicion was that the gate was in the center of the complex, no measure of precision was given for this center other than within the complex*.

This was elaborated in the follow up post:

As such I am assume that Xykon and Redcloak have roughly similiar location information for the Kraagor's gate - and given the amount of doors and how small a 1square mile area is I think it likely that the gate is in the complex.

*my belief is that the entire complex is greater than one square mile.

[dtilque]

Technically, that only tells us how Roy found his way there, not how Girard recorded them (or would have recorded them, had he done so; Roy was looking for a location that had been rolled randomly).
No it tells us the form of the coordinates.

Roy was using a sextant. A sextant is useless for anything but lat-long. Ergo, he had lat-long coordinates for Girard's gate. He got those coordinates from Soon, Soon got those coordinates from Girard. They were false coordinates, but Girard also had correct ones, because he gave them to Serini. Chances are both were in lat-long. In fact, the coordinates for ALL the gates were probably in lat-long.


That is a reasonable range of error for such a device. (Of course, we would have to account for Girard's error and Roy's as well; I won't go any further because perhaps Roy already accounted for this.)
Girard rolled dice to determine a false location, so he had no error. He just teleported to that spot and installed his booby trap. Roy would have had errors in his measurements, but those would have made the area to search even larger than a square mile. Chances are, he ignored that and assumed there was no error in his measurements.




The point I was making is that it's unlikely that the coordinates are so precise, so narrow and so accurate that we can say the point lies in the direct center of the hollow (which was the claim).

OK, I agree with that. As I said in my previous post, there was probably other info, elsewhere in the diary, that allowed them to narrow the location down quickly.