Doors and Statistics and Monty Haul - Is MitD helping?

[Son of A Lich!]

I'm not so Hot on the Numbers...

So, I'm pretty bad with Statistics.

I get dice and what can be expected and how to figure them out, but... This is something I'm not too sure if I understand the math behind MitD's Gambit with the Monster Hollow correctly.

The Monty Haul problem is set up with N doors. 1 door has a Car behind it, N-1 doors have Goats behind them. When the contestant picks a door at random, Monty reveals X of the doors (Where X=N-<2) that all have goats behind them, and gives the contestant a chance to change his answer.

When you make the numbers huge, it becomes obvious that you should always change your answer;

I had a 1 in 100 (N) chance of picking the right door, he revealed 98 doors that are bad answers, if I change my answer to the only other door, I am much more likely to get the door with a Car.

Makes sense right?

For sake of argument, lets say the Monster Hollow has 100 doors. One has the Gate (The Car), and 99 have monsters (the Goat).

Xykon picks a door and it is wrong. MitD picks 2 random doors and eliminates them. For the next trial, is Xykon more or less likely to pick the right door because of MitD?

Follow up, if the Order figures out that MitD was doing this, is it more beneficial to them to check the crossed off doors first and ignore the ones that haven't been checked?

I'm not even 100% sure on how to approach this question from a mathematics side and was hoping that someone with better math skills could help me understand.


Populations - P1 (Doors in Monster Hollow), P2 (Doors that have Monsters behind them), P3 (Doors that have been crossed off), P4 (Doors that have been checked).

Each trial is randomly selecting [P1-P3], Then if no gate is found randomly selecting 2 more [P1-P3] and adding 2 to the sum of P3. The chances of selecting a door with a gate is (P2-P1)/P1 or 1%. The chances of selecting a door that does not have a gate is P2/P1, or 99%.

Then, I would have to figure out what the probability of the second half of the trial is to having found a gate and making the rest moot right?[/spoilers]

[Spoilers= OotS and the Doors]
If I can find the probability of MitD crossing off the door with the gate behind it, then I think it's simply a matter of which ever is the highest population; Unchecked doors or checked doors. If there are 100 doors total and 60 are unmarked, the order should check them first before double checking the marked doors. If there are 60 marked doors, they should go for the Marked doors first. This is because the probability that MitD checked the right door off is still lower then the available population of unchecked doors... Right?

[Ruck]

The difference between MitD's situation and the Monty Hall Problem is that Monty Hall knows which door the prize is behind, so he's never opening that door. MitD is marking doors at random.

That said, an unmarked door is one that Team Evil has certainly not checked, whereas a marked door is either one that Team Evil has checked or one that MitD marked randomly. Thus, the chances of an unmarked door being a previously explored door are 0%, and therefore Team Evil should still explore all unmarked doors first.

I do not think MitD functionally affects the probability of Team Evil finding Kraagor's Gate in any one door they explore**; however, it increases the chances they don't find the gate before going through all the unmarked doors, meaning they have to go through the marked doors and thus waste time exploring some of them twice.

** EDIT: I wasn't sure about this so I thought it through some more, but I think it's correct:

If X = total number of doors
Y1 = number of doors explored
Y2 = number of doors marked
Z = number of doors unexplored (X - Y1)

Then the chance of finding the gate on any individual expedition is 1/Z. They're not going to explore a door twice in a row, because it's been marked. The difference is, when Y1 does not equal Y2, then when Y2 = 0, they'll have to start over at the beginning.

[BeerMug Paladin]

In a statistical sense, MitD is doing something to interfere with their search method. As noted by Ruck, if MitD happens to mark off the correct door, then it means they will keep searching until all the doors are eventually marked. Forcing them to ditch information they thought they had which would have been useful to increase their odds of finding the door.

This doesn't guarantee that some doors will be searched a second time (they could potentially choose doors at random which were unexplored) but given as numbers grow larger, it's pretty much as good as a guarantee.

Like with the Monty Hall problem, imagine the extreme case. There are 3 choices. You pick one. The chance of that choice being correct is 1/3. But when you mark off the choice to indicate you have already selected it, someone comes along and makes the marking on the other two choices. Now you don't have that information you tried to create for yourself and have to start over with the same 1/3 chance you had the first time.

In the case that the marking is placed on only one of the other two left, that leaves you with only one option, which is 1/2 odds to be what you want. Normally, there would be a 1/2 chance on day 2 to get the correct door. But if it's not found then, you have to begin searching all three again, which reduces your odds to 1/3 once again on the same day that you would have been guaranteed to have found the correct option if your markings were not interfered with.

These sorts of interference methods with information can create a case where it takes any arbitrary number steps of picking doors in order to find the correct one. (Although I think it's safe to assume that Xykon and Redcloak will not be restarting with the same conditions to their search method when they run out of doors to find.)

[Knaight]

Average unmodified searches necessary for N doors is N/2. On multiple searches this works out to kN+N/2, where k is a non negative integer. Consider the MitD as marking a fraction of the doors, x (between 0 and 1).

We can relate k to x, treating it as a probabalistic arithmetic sum, where the average k is the infinite series expansion. That gets us k=x/(1-x). For instance if they mark 3/4 of the doors gets k=3, which works out to 4 searches total (as k is repeat searches), matching intuitive expectations. The number of searches per cycle decreases with x however, proportional to 1/(1-x). That gets the overall equation

S=x/((1-x)^2)N+N/2(1-x)

N can be treated as a constant and thus ignored for the question of increase, so some quick simplification and algebra yields:

S=(2x(1-x)+1)/(1-x)=(-2x^2+2x+1)/(1-x)

Wolfram Alpha shows that this is increasing in the 0-1 domain. Outside it it gets weird, but outside it you're inputting negative probabilities or those above 1. 1 itself is a discontinuity, which is to be expected. Marking every single door every time is itself slightly outside the bounds of both the equation and its scenario.

[lopa12]

Average unmodified searches necessary for N doors is N/2. On multiple searches this works out to kN+N/2, where k is a non negative integer. Consider the MitD as marking a fraction of the doors, x (between 0 and 1).

[Mike Havran]

My take is as follows:

Original idea was:
1 door with the Gate
N = number of doors
MC = number of checked and marked doors
So probability of a success of finding the Gate during k iteration is PX(k)=1/(N-MC(k-1)).
They can make only one iteration per day.
On average, it would take N/2 days to reach the gate, and at most N days if they are extremely unlucky (and the gate is indeed behind one of the doors)

But then MitD started to sabotage their mission and we have to add

o = number of unchecked doors marked by MitD after every iteration (can be anything random from 0 to N, but assume low numbers such as random from 0 to 3)
MO = number of all doors marked by MitD that are otherwise unchecked
MA = number of all marked doors, MA = MC + MO

MitD marks unchecked doors after every iteration and probability of marking door that leads to Gate is PY(k) = o(k)/(N-MA(k))

So, now every iteration Team Evil makes has a chance of success that is marginally increased by lower number of unmarked doors, but largely diminshed by increasing odds that MitD managed to mark the correct door. Something like P(X(k+1)) = 1/(N-MA(k1)) * (1 - ∑(i=1;i=k)(oi/(N-MA(i-1))) if ∑[SUB](i=1;i=k)oi <= ∑[SUB](i=1;i=k)(N-MA(i-1))

I apologize if it's hard to understand, but I don't know how to write equations nicer here.

The bottom line is that Team Evil is quickly approaching the point where their effort with random selection and uniform marking is a complete waste of time. MitD should only take care not to raise suspicion. It demonstrates how unwise Xykon is: if they had chosen predetermined search pattern and meticulous evidence of the checked doors, they would have already discovered the sabotage.

[Chronos]

Of course, it's also possible that, once they see that all doors are marked but they still haven't found the Gate, that they come to a wrong conclusion. Sure, if they conclude that someone or something is marking extra doors, then they'll start re-checking marked doors, and get there eventually. But maybe they instead conclude that the doors are a red herring, and start searching the statue's butt crack, or something.

[Sir_Norbert]

Exactly.

Which is why the legion of forum members proclaiming "I'm so clever! I've worked it out! The Gate is really in the statue!" are certainly wrong.

[Kish]

Before the creature in the darkness marks anything: 100 unsearched doors.

After he does: 100 unsearched doors. Four of which Redcloak and Xykon will never search.

He's making only one change to the odds: He's making it possible for Redcloak and Xykon to believe that they've searched all the doors, and not found what they're looking for, even if what they're looking for is in fact behind one of the doors.

What he's doing would only be helpful to Redcloak and Xykon if an unmarked door was somehow inherently more likely to have the Gate than a marked, unsearched door. Going from Door 1 to Door 6 is exactly the same, in terms of "likelihood the Gate is behind this door," as going from Door 1 to Door 2; there's no actual help in "helping them get to Door 6 faster."

[BaronOfHell]

Imagine you have a sack of marbles, and one of these marbles are special.

Now you take one out at a time, the amount in the sack diminishes, and eventually you find the right marble.

Imagine in stead that every time you take one out, you put it back in again so the amount in the sack is never reduced, that is kind of what the MitD is doing by making team Evil not know which dungeons they have already crawled. Not entirely though, because if the MitD marked all doors, team Evil could begin a new approach and be more careful that no one interferes with their system.

Alternatively, if you have to pick between several white cards among which only one is correct, ask yourself if you somehow magically have a higher chance of succeeding if you randomly start to color some of these cards red.

So what the MitD is doing does make it more difficult to find the correct door in my opinion, assuming there exist a correct door.

[Harbinger]

Like with the Monty Hall problem, imagine the extreme case. There are 3 choices. You pick one. The chance of that choice being correct is 1/3. But when you mark off the choice to indicate you have already selected it, someone comes along and makes the marking on the other two choices. Now you don't have that information you tried to create for yourself and have to start over with the same 1/3 chance you had the first time.

In the case that the marking is placed on only one of the other two left, that leaves you with only one option, which is 1/2 odds to be what you want. Normally, there would be a 1/2 chance on day 2 to get the correct door. But if it's not found then, you have to begin searching all three again, which reduces your odds to 1/3 once again on the same day that you would have been guaranteed to have found the correct option if your markings were not interfered with.

This is actually not correct, with regards to Monty Hall. If you decide to switch doors after Monty Hall opens the initial door, your chance of getting the car goes from 1/3 to 2/3, not from 1/3 to 1/2. This is because Hall will always open a door with a goat, so unless you pick right initially you will always get the car.

[Ruck]

This is actually not correct, with regards to Monty Hall. If you decide to switch doors after Monty Hall opens the initial door, your chance of getting the car goes from 1/3 to 2/3, not from 1/3 to 1/2. This is because Hall will always open a door with a goat, so unless you pick right initially you will always get the car.

This is true, but in this case I believe that BeerMug Paladin is describing the situation we have in the story, where MitD is marking doors at random, reduced to three doors for simplicity's sake.

[Manga Shoggoth]

Before the creature in the darkness marks anything: 100 unsearched doors.

After he does: 100 unsearched doors. Four of which Redcloak and Xykon will never search.

He's making only one change to the odds: He's making it possible for Redcloak and Xykon to believe that they've searched all the doors, and not found what they're looking for, even if what they're looking for is in fact behind one of the doors.

What he's doing would only be helpful to Redcloak and Xykon if an unmarked door was somehow inherently more likely to have the Gate than a marked, unsearched door. Going from Door 1 to Door 6 is exactly the same, in terms of "likelihood the Gate is behind this door," as going from Door 1 to Door 2; there's no actual help in "helping them get to Door 6 faster."

Indeed. We have only seen him do this once, and I suspect that he has done this several times, and is getting away with it because there are enough doors and they are being opened randomly. If they followed a pattern per Redcloak's suggestion then they would have caught on very quickly. All he has to do is mark three extra doors on three visits (or do several doublets) and he can easily make it a significant number of doors missed.

The flaw in the plan is that if he keeps going too long and gets caught then it is trivial to check the doors that have already been done, and if they then start on a clear search pattern they won't really lose any time.

Also, if Redcloak is keeping count he might notice that they have run out of doors early.

[Knaight]

The flaw in the plan is that if he keeps going too long and gets caught then it is trivial to check the doors that have already been done, and if they then start on a clear search pattern they won't really lose any time.

The doors repopulate, so it's not trivial to check them - at minimum this adds rechecking every door, which adds time.

[Ruck]

Indeed. We have only seen him do this once, and I suspect that he has done this several times, and is getting away with it because there are enough doors and they are being opened randomly. If they followed a pattern per Redcloak's suggestion then they would have caught on very quickly. All he has to do is mark three extra doors on three visits (or do several doublets) and he can easily make it a significant number of doors missed.

The flaw in the plan is that if he keeps going too long and gets caught then it is trivial to check the doors that have already been done, and if they then start on a clear search pattern they won't really lose any time.

Also, if Redcloak is keeping count he might notice that they have run out of doors early.

I don't think it'll be trivial to re-check the doors, unless Redcloak catches on pretty quickly that MitD is marking other doors.

Fortunately for MitD, his supposed stupidity makes for a great excuse as to why he "screwed up."

[hroþila]

The doors repopulate, so it's not trivial to check them - at minimum this adds rechecking every door, which adds time.
While I personally believe that the monsters respawn all video game-like, what Oona said allows for the possibility that the bugbears are simply careful not to hunt too many monsters so that the population increases naturally. If that's the case, that shouldn't be a problem in sections of the dungeon that have been completely cleared, or at the very least it shouldn't be a problem in a very long time.

[Knaight]

While I personally believe that the monsters respawn all video game-like, what Oona said allows for the possibility that the bugbears are simply careful not to hunt too many monsters so that the population increases naturally. If that's the case, that shouldn't be a problem in sections of the dungeon that have been completely cleared, or at the very least it shouldn't be a problem in a very long time.

Maybe. There's also the real possibility that the segments aren't quite as segmented as they seem, and that the entire dungeon would need to be depopulated heavily to really make a mark on any individual segment.

[RatElemental]

I have to wonder if the mitd could in fact pull off not marking an explored door if he could also pick a door far away from it the next day (See: Xykon asking if anyone was feeling especially lucky. It doesn't matter who picks so he might let the mitd pick), thus wasting even more time once the team finally gets back around to it.

[BeerMug Paladin]

This is actually not correct, with regards to Monty Hall. If you decide to switch doors after Monty Hall opens the initial door, your chance of getting the car goes from 1/3 to 2/3, not from 1/3 to 1/2. This is because Hall will always open a door with a goat, so unless you pick right initially you will always get the car.
As Ruck already noted, I was speaking there about this particular case. I was only referencing Monty Hall in regards to the fact that the extreme case makes the overall understanding of what the scenario does easier to intuit. In Monty Hall, the extreme case is a million doors. In this door searching scenario, the extreme case is 3 doors with the particular impact that each individual sub-scenario will leave you with.

[Son of A Lich!]

Average unmodified searches necessary for N doors is N/2. On multiple searches this works out to kN+N/2, where k is a non negative integer. Consider the MitD as marking a fraction of the doors, x (between 0 and 1).

We can relate k to x, treating it as a probabalistic arithmetic sum, where the average k is the infinite series expansion. That gets us k=x/(1-x). For instance if they mark 3/4 of the doors gets k=3, which works out to 4 searches total (as k is repeat searches), matching intuitive expectations. The number of searches per cycle decreases with x however, proportional to 1/(1-x). That gets the overall equation

S=x/((1-x)^2)N+N/2(1-x)

N can be treated as a constant and thus ignored for the question of increase, so some quick simplification and algebra yields:

S=(2x(1-x)+1)/(1-x)=(-2x^2+2x+1)/(1-x)

Wolfram Alpha shows that this is increasing in the 0-1 domain. Outside it it gets weird, but outside it you're inputting negative probabilities or those above 1. 1 itself is a discontinuity, which is to be expected. Marking every single door every time is itself slightly outside the bounds of both the equation and its scenario.


My take is as follows:

Original idea was:
1 door with the Gate
N = number of doors
MC = number of checked and marked doors
So probability of a success of finding the Gate during k iteration is PX(k)=1/(N-MC(k-1)).
They can make only one iteration per day.
On average, it would take N/2 days to reach the gate, and at most N days if they are extremely unlucky (and the gate is indeed behind one of the doors)

But then MitD started to sabotage their mission and we have to add

o = number of unchecked doors marked by MitD after every iteration (can be anything random from 0 to N, but assume low numbers such as random from 0 to 3)
MO = number of all doors marked by MitD that are otherwise unchecked
MA = number of all marked doors, MA = MC + MO

MitD marks unchecked doors after every iteration and probability of marking door that leads to Gate is PY(k) = o(k)/(N-MA(k))

So, now every iteration Team Evil makes has a chance of success that is marginally increased by lower number of unmarked doors, but largely diminshed by increasing odds that MitD managed to mark the correct door. Something like P(X(k+1)) = 1/(N-MA(k1)) * (1 - ∑(i=1;i=k)(oi/(N-MA(i-1))) if ∑[SUB](i=1;i=k)oi <= ∑[SUB](i=1;i=k)(N-MA(i-1))

I apologize if it's hard to understand, but I don't know how to write equations nicer here.

The bottom line is that Team Evil is quickly approaching the point where their effort with random selection and uniform marking is a complete waste of time. MitD should only take care not to raise suspicion. It demonstrates how unwise Xykon is: if they had chosen predetermined search pattern and meticulous evidence of the checked doors, they would have already discovered the sabotage.

Thank you for helping, I'm following along as best as I can, but I'm still a little bit lost with the specifics here. I'm going to have to pull out my Stat's text book and see if I can get back up to speed with what I'm looking at here.

It's not so much the process as it is the notation, I believe both of you two have the general formula for what I'm looking for solved, but I don't know if I could plug numbers in and see the results as of yet.


In a statistical sense, MitD is doing something to interfere with their search method. As noted by Ruck, if MitD happens to mark off the correct door, then it means they will keep searching until all the doors are eventually marked. Forcing them to ditch information they thought they had which would have been useful to increase their odds of finding the door.

This doesn't guarantee that some doors will be searched a second time (they could potentially choose doors at random which were unexplored) but given as numbers grow larger, it's pretty much as good as a guarantee.

Like with the Monty Hall problem, imagine the extreme case. There are 3 choices. You pick one. The chance of that choice being correct is 1/3. But when you mark off the choice to indicate you have already selected it, someone comes along and makes the marking on the other two choices. Now you don't have that information you tried to create for yourself and have to start over with the same 1/3 chance you had the first time.

In the case that the marking is placed on only one of the other two left, that leaves you with only one option, which is 1/2 odds to be what you want. Normally, there would be a 1/2 chance on day 2 to get the correct door. But if it's not found then, you have to begin searching all three again, which reduces your odds to 1/3 once again on the same day that you would have been guaranteed to have found the correct option if your markings were not interfered with.

These sorts of interference methods with information can create a case where it takes any arbitrary number steps of picking doors in order to find the correct one. (Although I think it's safe to assume that Xykon and Redcloak will not be restarting with the same conditions to their search method when they run out of doors to find.)

This on the other hand is Perfect. I would have never thought to take it the other extreme in a million years.

Run with me on this though - Lets take it to four doors. Xykon picks one and (Assuming there is no gate behind it) MitD marks off the second and third, then Xykon picks the last door. This would improve his odds, if I'm following this right. (It improves MitD's odds of running interference, from 0 to 33%, but I was wondering if this helps Xykon in the long run).


So, Xykon's first pick is 1/4 for .25

MitD picks two doors and marks them. Since order does matter, we solve for the chances of picking two 'Goats' in a row and subtract it from 1. (1-((2/3)*(1/2)))= .33

Now, since the fourth door is guaranteed to be determined by this point, Xykon's odds are .66 in finding the right door(If MitD did not find the Car, the Car must be behind the last door). If MitD did not run interferance, he would have (1-((3/4)*(2/3)))= .50


So, a .16 improvement on his odds, right?

I referred to Monty Hall due to the way that it changes the numbers from what you would expect. The fact that MitD could pick the right door is not so much a problem as the fact that he is unlikely to pick the right one.

If we up this to 10 doors but the same set up (Xykon picks one, MitD picks two, Xykon picks one again) it results like this -


Xykon picks 1 door with a 1/10 chance of getting the gate for .1

MitD picks 2 at random. (1-((8/9)*(7/8)))= .22

Xykon picks a new door (1-(6/7))+ .1 = .24

The chances of the door with the car is undiscovered is (1-(.22+.24))= .54

The chances of Xykon finding the door with just two random pulls is (1-((9/10)*(8/9)))= .20


So, Xykon is still getting a benefit from the random pulls that aren't specifically to help him...

Right? Or am I fluffing the math up hard here?

The difference between MitD's situation and the Monty Hall Problem is that Monty Hall knows which door the prize is behind, so he's never opening that door. MitD is marking doors at random.


Well, this is a bit different then Monty Hall in a few different ways; First, the first door is revealed regardless of goat/car. Secondly, the extra Marked doors are applied at random with no knowledge of the Goats/Cars. Then you HAVE To pick a new door afterward.

Also - Its a fictional dilemma in a universe where the probability of something working is tied to the Drama. I doubt that 'Analyzing Statistical Probabilities' is going to be the climax of Kraagor's gate.

Our friendly local bard would tell us that the fact that MitD is hiding what he did and O-Chul witnessed it proves that it's helpful to the Good Guys and bad for Team Evil; but that since the plan has been announced it will only be a matter of time before it fails, spectacularly.

I was just curious as to whether or not the math helps or hinders Xykon over all, in the event that this ever comes up elsewhere in life. *Shrug*

Thank you for re-splitting up the populations, that helped out a lot as well, Ruck.

[BeerMug Paladin]

This on the other hand is Perfect. I would have never thought to take it the other extreme in a million years.

Run with me on this though - Lets take it to four doors. Xykon picks one and (Assuming there is no gate behind it) MitD marks off the second and third, then Xykon picks the last door. This would improve his odds, if I'm following this right. (It improves MitD's odds of running interference, from 0 to 33%, but I was wondering if this helps Xykon in the long run).


So, Xykon's first pick is 1/4 for .25

MitD picks two doors and marks them. Since order does matter, we solve for the chances of picking two 'Goats' in a row and subtract it from 1. (1-((2/3)*(1/2)))= .33

Now, since the fourth door is guaranteed to be determined by this point, Xykon's odds are .66 in finding the right door(If MitD did not find the Car, the Car must be behind the last door). If MitD did not run interferance, he would have (1-((3/4)*(2/3)))= .50


So, a .16 improvement on his odds, right?

Since the markings don't actually give information to further cut down on the possible correct choice, the markings have no impact on the second door's likelihood of being the correct one when opened. The second door being opened has a 1/3 chance of being correct regardless of markings placed on the doors. All the markings do is attempt to register which doors were already searched beforehand.

It only changes the odds after all the doors are marked (and without success), making you know you've lost the information you had thought was accurate. At that point, the odds just simply reset to its beginning state.


I referred to Monty Hall due to the way that it changes the numbers from what you would expect. The fact that MitD could pick the right door is not so much a problem as the fact that he is unlikely to pick the right one.

If we up this to 10 doors but the same set up (Xykon picks one, MitD picks two, Xykon picks one again) it results like this -


Xykon picks 1 door with a 1/10 chance of getting the gate for .1

MitD picks 2 at random. (1-((8/9)*(7/8)))= .22

Xykon picks a new door (1-(6/7))+ .1 = .24

The chances of the door with the car is undiscovered is (1-(.22+.24))= .54

The chances of Xykon finding the door with just two random pulls is (1-((9/10)*(8/9)))= .20


So, Xykon is still getting a benefit from the random pulls that aren't specifically to help him...

Right? Or am I fluffing the math up hard here?
You're overcomplicating things here.

The chance of Xykon's second door being the correct choice is 1/9, because there are 9 unsearched doors. Only Xykon's pick of doors will remove that door as a possibility, MitD has no impact on the second choice, even if his markings make it seem like he's narrowed down the possibilities. Only on a completed cycle of unsuccessful choices do the markings make a difference, which results in a reset back to the starting state of the scenario.

The odds of the MitD marking off the correct door are as follows. After the first choice keeps the scenario going, since he picks two the odds he picked a correct door are 2/9. Since there are two selections made out of nine unsearched doors at that point.

Xykon looks at the three markings, and so believes his odds are 1/7. But since two of those aren't accurate markings, his odds are really 1/9, it could be behind any of the unmarked doors, or one of the two falsely marked doors.

The key here is information. MitD is a random actor with no information, so his actions don't have any real impact on the odds of Xykon's success. It only has an impact on the actual statistics if his random activity ends up forcing a restart based on the fact that his actions are destroying useful information. Up until that resetting point, his actions make no difference on step 2, 3 or whatever until the restart.

At step two, with one searched door and three marked doors, choosing a marked door has a (1/9)*(2/3) chance of success (based on the fact that two of those are possible right answers, but choosing one among that group you have a 2/3 chance of getting one of the MitD marked doors). So, that set is actually worse than going with one of the seven others where your odds are simply 1/9. If MITD were not doing this, then choosing a marked door would have a 0% chance of success.

This is because the information that one of those markings contains is still in the scenario. It's just that MitD is destroying information by ensuring the odds of a marked door being a "wrong answer" is not 100%. As Xykon adds doors, the chances of those marks indicating a wrong answer go up but as MitD adds marks, the chances go down.

If the scenario gets to a step where all the doors are marked, then the information of which doors are incorrect aren't indicated in any way by a marking. This is what I mean by the information being destroyed.

With 10 doors, the odds of Xykon succeeding in a correct choice within two trials is (1/10) + (9/10)*(1/9) = (1/10) + (1/10) = 2/10 = .2 (Keep in mind, the 9/10 factor is the likelihood that the scenario continues after one door searching.)

It's deceptively simple. Again, MitD's actions only impact the odds when they force a restart.

[Chronos]

Suppose that they did go through every door in order. But after searching each door, the Monster marks that door, and then also marks another door at the end of the sequence. In this case, the Monster could never help the search, but if the true door happens to be in the last half, he will hinder it (exactly how much he hinders it depends on how Xykon and Redcloak react once all doors have been marked without the Gate being found).

And I would maintain that this situation is exactly equivalent to what they're doing.

[MartianInvader]

Should we factor in a prior probability of the gate actually being under Kraagor's statue?

[Reboot]

Should we factor in a prior probability of the gate actually being under Kraagor's statue?
I think that's unlikely. The "Haley's Shell Game" option is possible, but more in a "there is no path which leads to the Gate, it's just walled off" way (why give them any chance, eh rogues?).

The statue has a *higher * chance of being searched, being out in the open, than the apparent "Gate at the hard-to-reach end of one random path from hundreds" scenario.

[Knaight]

Should we factor in a prior probability of the gate actually being under Kraagor's statue?

We could do that, it's an easy enough modification to the equations Mike Havran and I provided (assume they check once filled, set up a weighted average of sort). That said, they seem pretty spectacularly low. Kraagor seems like a straightforward person, and while that sort of thing would absolutely be how Girard operates Serini trying to build an homage to Kraagor would operate differently.

[Sir_Norbert]

Since you mention Girard... he pulled the whole "Your Gate is in another Pyramid" as one final line of defence, after the gate-seeker had already got through all the other defences he set up. He did not leave the Gate out in the open, so that it would be completely undefended if the seeker saw through his ruse.

None of the Scribblers would be that stupid.

[RatElemental]

None of the Scribblers would be that stupid.

I for one think you underestimate just how stupid otherwise intelligent people can be sometimes. I don't think the gate is there either, though. That said, Serini definitely could have put a mechanism in the statue that opens a path in one of the random doors that leads to the actual gate which is otherwise completely blocked off. Or a trap, in case someone decided to go poking at it.

[Chronos]

Sure, even Soon wouldn't actually put the Gate in a statue's butt-crack, but you still have to search there anyway if you're going to be thorough, because "he wouldn't put it there, that's too stupid" is the same sort of mistake as "it's so stupid it's perfect, because nobody will look there".

In other words, you have to search the butt-crack for the same reason that it isn't the butt-crack.

[locksmith of lo]

he isn't helping if the answer is none of the doors. and they do not catch his antics...

[Aquillion]

The flaw in the plan is that if he keeps going too long and gets caught then it is trivial to check the doors that have already been done, and if they then start on a clear search pattern they won't really lose any time.It's not trivial, is it? It takes the same amount of time as any other search. And more importantly, they can't distinguish between doors they actually searched and doors MitD falsely marked, which means that if they have to go back and check every marked door, they end up wasting time checking some doors twice.

Basically:

1. Since he's marking doors at random, up until they run out of door or discover what he's doing, his actions actually have no effect (any door he falsely marks could have just been at the end of the list of doors they check instead, so it doesn't, statistically, change anything as long as there are still unchecked doors.)

2. If they run out of doors without finding anything (because he marked the correct door), they waste a bunch of time because they essentially have to start over. So in the long run, statistically, he is costing them time (but of course it's statistical, so he might or might not in practice.)

3. But there are two other considerations: Is the actual gate behind any of the doors? And what will Team Evil conclude when they run out of unmarked doors?

4. One possibility is that they will conclude (incorrectly) that the gate is not behind any doors. In that case, MitD's plan could not only delay them but prevent them from ever finding the Gate.

5. Of course, another possibility is that they will conclude correctly that the gate is not behind any doors. In that case MitD's trickery did in fact help them! But this is a low-probability outcome.

6. Or they'll start over, and the gate is not behind any doors. This could happen if they discover MitD's trickery near the end. Obviously this wastes even more of their time.

7. In fact, if MitD only marked a small number of doors, another trick he could pull is claim to have marked a large number of doors - this makes it seem much more likely that the gate is behind one of the marked doors (and they only missed it because of his trickery), rather than not being there at all, so they'll waste more time searching doors. Conversely, if MitD marked a large number of doors and they reach the end, he could remain silent (or, if caught, say that he marked only a small number of them) to convince them that the gate isn't there.

[Manga Shoggoth]

It's not trivial, is it?

Of course it's trivial. The minute they find one of the corpses from a monster they killed they can say "We've done this one".

[Grey_Wolf_c]

Of course it's trivial. The minute they find one of the corpses from a monster they killed they can say "We've done this one".

It wouldn't be much of an ultimate protection for the rift if it was that easy. New monster may spawn, or be kept in reserve, or by any other means the author need not concern itself with, take over the old haunts, consume any corpses left behind and be effectively back at fulls strength by the next morning.

Grey Wolf

[Ruck]

Of course it's trivial. The minute they find one of the corpses from a monster they killed they can say "We've done this one".


It wouldn't be much of an ultimate protection for the rift if it was that easy. New monster may spawn, or be kept in reserve, or by any other means the author need not concern itself with, take over the old haunts, consume any corpses left behind and be effectively back at fulls strength by the next morning.

Grey Wolf

Indeed, it seems like Oona was pretty clear that the monsters "come back" (http://www.giantitp.com/comics/oots1039.html) after some time.

[Manga Shoggoth]

Indeed, it seems like Oona was pretty clear that the monsters "come back" (http://www.giantitp.com/comics/oots1039.html) after some time.

Yes, but no timescale is given, nor was any mechanism given. We have no reason to suppose that there is a stock of creatures. We don't even know (yet) if someone is still maintaining the system.

Until we have any further evidence my position is quite reasonable, and as valid as any other.

Edit: I personally wouldn't put it past Serini - a rogue, if I recall correctly - to be pulling some sort of trick here, just not using magic (so it will actually last longer than Giriad's setup after death).

[Grey_Wolf_c]

Yes, but no timescale is given, nor was any mechanism given. We have no reason to suppose that there is a stock of creatures. We don't even know (yet) if someone is still maintaining the system.

Until we have any further evidence my position is quite reasonable, and as valid as any other.

Sure, but not reasonable enough to attack other people's theories that are too built on equally reasonable positions. To whit: you cannot counter "they would not be able to tell if they've visited a cave or not" with "the dead bodies will still be there", because reasonable or not, it is not canon.

Grey Wolf

[Manga Shoggoth]

Sure, but not reasonable enough to attack other people's theories that are too built on equally reasonable positions. To whit: you cannot counter "they would not be able to tell if they've visited a cave or not" with "the dead bodies will still be there", because reasonable or not, it is not canon.

Grey Wolf

Where did I attack a theory? I have three earlier posts here:

The first post I am agreeing with Kish.
The second post I am responding to a single point of Aquillion's.
The third post I am defending my position when it is being questioned (Note: questioned, not attacked. Disagreeing is not attacking).

I have not said other points are not valid, I have only said - in response to people who are dismissing my points - that my points are equally valid at the moment.

And yes, I can respond with "the dead bodies will still be there" because on the current timescales (~one cave per day, IIRC) unless someone is literally clearing out the cave afterwards they will still be there (less any used for undead, of course).

[Grey_Wolf_c]

Where did I attack a theory? I have three earlier posts here:

You attacked a theory the moment you categorically asserted that re-testing a cave is trivial. You can assume that, but you cannot prove that.

And Thor, do I hate "oh let's play semantics with words" posts. "Attack" is a perfectly valid verb to be used here. You do not get to tell me what I can and cannot use when expressing myself.


3. begin to deal with (a problem or task) in a determined and vigorous way.
"a plan of action to attack unemployment"
synonyms: attend to, address, see to, deal with, grapple with, confront, direct one's attention to, focus on, concentrate on, apply oneself to

Grey Wolf

[Manga Shoggoth]

You attacked a theory the moment you categorically asserted that re-testing a cave is trivial. You can assume that, but you cannot prove that.

Neither can you prove the opposite. The only evidence we have (to the best of my recall) is the comment from Oona which is less than useful in timescales, and the the test case suggested is quite trivial. You don't have to search the entire cave (including fighting monsters) if there are bodies there that you have already killed.



And Thor, do I hate "oh let's play semantics with words" posts. "Attack" is a perfectly valid verb to be used here. You do not get to tell me what I can and cannot use when expressing myself.

In which case nor should you get to complain if you use an aggressive word like "attack", and are then misunderstood. I can't tell what you are thinking and what weight you are assigning to the words you use.

[Grey_Wolf_c]

In which case nor should you get to complain if you use an aggressive word like "attack", and are then misunderstood. I can't tell what you are thinking and what weight you are assigning to the words you use.

Yes, yes I can get to complain when you decide to take issue with how your language works, not of "being misunderstood". You are the one that decided to police my vocabulary use, not me with you. I merely complained that of all the ways you could address my post, you decided to make this a semantics fight. Seemingly while ignorant of the definitions of the word in your own language.

Grey Wolf

[Aquillion]

If it were trivial to determine if a cave had already been explored, they wouldn't need to mark them off, would they?

[Ruck]

Yes, but no timescale is given, nor was any mechanism given. We have no reason to suppose that there is a stock of creatures. We don't even know (yet) if someone is still maintaining the system.

Until we have any further evidence my position is quite reasonable, and as valid as any other.

Edit: I personally wouldn't put it past Serini - a rogue, if I recall correctly - to be pulling some sort of trick here, just not using magic (so it will actually last longer than Giriad's setup after death).

From a Doylist perspective, it seems rather evident to me that the only reason that the monster restocking was mentioned was to make explicit that they won't be able to go back into previously explored doors and have them be empty.


If it were trivial to determine if a cave had already been explored, they wouldn't need to mark them off, would they?

Also, yeah, that.

[Chronos]

I think it unlikely that they would find any corpses. Dangerous monsters tend to be carnivorous, and few carnivores would pass up fresh carrion, free for the taking.

[BeerMug Paladin]

Neither can you prove the opposite. The only evidence we have (to the best of my recall) is the comment from Oona which is less than useful in timescales, and the the test case suggested is quite trivial. You don't have to search the entire cave (including fighting monsters) if there are bodies there that you have already killed.
Initially, I almost posted the same sentiment regarding the corpses. Then I recalled Oona's statement in the mentioned comic and thought better of it.

If it were trivial to determine if a cave had already been explored, they wouldn't need to mark them off, would they?
This also implies that they don't have an easy way to identify past caves. Redcloak obviously isn't high enough level to cast Commune, either.

The comic kind of implies the markings are useful. So the scenario presented is pretty easy to accept at face value.

Although if we want to go into the realm of stating that we don't know certain things about the search, we can always say things like they aren't thoroughly exploring the chambers accessible through the doors. We have no evidence they aren't, but hey, we don't have evidence that they are either, so either one is just as likely, isn't it?

[Manga Shoggoth]

Initially, I almost posted the same sentiment regarding the corpses. Then I recalled Oona's statement in the mentioned comic and thought better of it.

The way I read the full statement was that Oona (and tribe) treats it as a hunting ground - every so often they kill something and over time it naturally restocks over time - they may not even be clearing entire caves out, so the remaining creatures can still breed. In which case, that is why Oona in the next frame is concerned that all the doors caves are being accessed and cleared out in quick succession - Xykon and Redcloak are over-hunting, and also presumably leaving the bodies where they lay.

Of course, the other way of reading it is possible at the moment - we have no other evidence that I am aware of. But my theory does seem simpler.

When I re-read Oona's statement, I briefly thought it could be interpreted to mean that Oona and tribe are maintaining the hollow themselves (having found it abandoned), but I don't think that fits.

[Chronos]

Oona wouldn't be cooperating with Xykon and Redcloak if they were leaving the dungeons completely incapable of restocking. This means one of three things: Either they are leaving a few monsters in each dungeon alive to breed, or monsters can get into the dungeons from the other ones to repopulate them, or the monsters magically respawn. In any one of those three cases, there are likely to be at least a few monsters in any dungeon, shortly after they leave it, and which could eat the corpses they've left behind.

[Sir_Norbert]

If it were trivial to determine if a cave had already been explored, they wouldn't need to mark them off, would they?

That's nonsense. Supposing it's trivial to determine if a cave has already been explored, they still have to go inside to make that determination, and that adds up to a lot of time for each attempt, given the number of caves. Marking them creates a sign that's visible from the outside.

[Fish]

With 10 doors, the odds of Xykon succeeding in a correct choice within two trials is (1/10) + (9/10)*(1/9) = (1/10) + (1/10) = 2/10 = .2 (Keep in mind, the 9/10 factor is the likelihood that the scenario continues after one door searching.)

It's deceptively simple. Again, MitD's actions only impact the odds when they force a restart.
This suggests a simpler solution. If the odds of each door after the first are multiplied by the ratio of remaining doors (eg, [1/n]*[[n/10] + [1/n-1]*[n-1/10]...) then cancelation the total odds for Xykon are easy to calculate.

If Xykon tries 1 door and the MITD marks 3 untried doors, then Xykon’s total odds become (1/4) for any n doors. For 100 doors, Xykon only tries 25, so his odds are 25/100 or 1/4. The odds that the MITD marked the right door are 75/100 or 3/4.

The only thing that changes is the rate at which Xykon reaches his conclusion (either finding it or not) 4 times faster.

[AutomatedTeller]

Of course it's trivial. The minute they find one of the corpses from a monster they killed they can say "We've done this one".


Honestly, it's not clear to me that any Team Evil pays enough attention to detail to notice.

[Aquillion]

Oona probably does.

Also, Oona probably doesn't care.

[rbetieh]

How many straight tries did team evil get before starting to mark extras? I would expect that a high Wis character would have remembered that 1 != 4.... At what point did they lose track of the number of days they had tried the dungeon out?

[Mike Havran]

How many straight tries did team evil get before starting to mark extras? I would expect that a high Wis character would have remembered that 1 != 4.... At what point did they lose track of the number of days they had tried the dungeon out?I would bet a small sum on that the on-screen marking was the first instance. MitD probably needed more time to figure out that he can mess up with their search that way.

[Mordaedil]

I would bet a small sum on that the on-screen marking was the first instance. MitD probably needed more time to figure out that he can mess up with their search that way.

I agree, and if not the first instance, the first one where he marked off that many at once. You need at least a dozen caves down before it starts to blend together, albeit if they were more aware, they'd notice as long as it was less than 100 doors.

Regardless, it does seem like a matter of time before the deception is discovered and I reckon it'll be before the last few doors.

[RatElemental]

It would probably be prudent to realize that the Mechane isn't the only thing traveling at the speed of plot. Team evil will/won't find the gate exactly when they're meant to by Rich, and whatever happens will probably involve them being defeated in an unexpected manner by the OotS, possibly involving the planet in the rifts.

The significance of MitD marking the doors is that it shows he's not just going along with it anymore, as O-Chul already pointed out. He even pointed out it's dubious whether or not it actually delayed team evil at all. It may also serve as an extra source of drama if he is found out/almost found out doing it. Statistics and chance is going to play precisely 0 part in this.

[Manga Shoggoth]

Honestly, it's not clear to me that any Team Evil pays enough attention to detail to notice.


Oona probably does.

Also, Oona probably doesn't care.

True on both counts. The only problem is that Redcloak does pay attention to detail, and Xykon does when he feels like it (and often unnoticed). We'll see when the die is cast.

I do wonder what game Oona is playing here.

[Fish]

I just ran some simulated trials using 100 doors, using the following assumptions:

1. For every 1 door that Xykon picks, the MITD marks off 3 unexplored doors.
2. When Xykon picks the door with the Gate, keep track of how many times it takes to reach that result, and begin the next trial.
3. When Xykon does not pick the door with the Gate, count all of those trials plus the number assigned to the correct door. Since the number is from 1 to 100, this averages out to approximately (tot_doors + 1) / 2.
4. Xykon is presumed not to notice the manipulation with the door markings until he gets through all of his doors.

Using one particular sample of 100,000 trials:

Xykon succeeds in finding the Gate 25,320 times out of 100,000 attempts (roughly 25%, what we would expect if MITD is hiding 75% of the doors). He fails in 74,680 attempts.

When Xykon succeeds, it takes an average of 12.99 guesses to find the right door. This is what we would expect to see if Xykon is only choosing 25% of the doors. In other words, he is choosing only 1/4 of the doors, so he goes through his choices 4 times faster than would be expected without interference.

When Xykon fails, it takes an average of 76.086 guesses to find the right door. This is also what we'd expect: 25 guesses to realize that he's been duped, then roughly (1 + tot_doors)/2 guesses to start again from scratch without using the old markings.

Average number of guesses to get the correct door: 60.11
Expected number of guesses to get the correct door: 50.5

Under those conditions, the MITD is helping.

However, if you change condition 3 to the following:

3. When Xykon does not pick the door with the Gate, count all of those trials plus the average number of times it takes to search all the remaining doors (presuming there is some way to discern when a door has not been previously searched). This averages out to approximately (0.75 * tot_doors + 1) / 2.

Then Xykon takes an average of 50.59 trials to find the correct door, even with the interference. This is almost dead-on what we'd expect for the number of attempts without interference (50.5).

[Aquillion]

Something else I noticed: They're letting MitD choose the doors they open. Assuming they don't notice if it's a previously-cleared cave, that would allow him to be much more effective at wasting time (since he could just send them to caves he knows are empty.) It's much more risky, though.

Although it's also worth considering that we don't know that he's doing this in any systematic way. It's possible painting those doors was a spur-of-the-moment choice.

(Another possibility is that he could sabotage them somehow within the hollows, eg. by offering to search branches for them and then not searching them - though they know he can't see the gates, so this could be difficult.)

But the real answer is of course what O'Chul said. This isn't really going to matter - the important thing is what it says about MitD, and in that respect it's his intent that counts.

[Sir_Norbert]

Well, the Monster can only direct them to a previously explored cave by removing marking from a marked cave, and we haven't observed him doing that yet.

[RatElemental]

Well, the Monster can only direct them to a previously explored cave by removing marking from a marked cave, and we haven't observed him doing that yet.

Or by not marking it in the first place.

[lio45]

The only thing that changes is the rate at which Xykon reaches his conclusion (either finding it or not) 4 times faster.

Sure, but let's not forget that Xykon reaching "the end of unmarked doors" sooner does not make it one iota more likely he'll have found the correct door by that point (his odds of having found it at that point are unchanged). The only thing that this "speeding the end by randomly marking unexplored doors" changes is that there is now a chance - getting higher with each extra marked-but-unexplored door - that Xykon will reach the end of unmarked doors before reaching the end of all doors AND will not have found the gate yet, which would be a severe blow to his plans (causing significant delay; back to square one if you believe monsters respawn and cadavers vanish, or back many squares even if you don't believe that).

[RatElemental]

Here's an interesting thought: Say team evil makes it to the end of the doors with no gate in sight. Do they even suspect sabotage at all or assume Serini worked some trick into the gate? I think it likely they'd think some trick was involved, meaning they might waste even more time trying to figure out what it was before starting over with all the doors.

[Aquillion]

I mean, let's be real, as someone said above, Serini was a rogue. There's a 0% chance that the gate is actually behind any of those doors.

(Come to think of it, why are the gates even accessible? Why not encase it entirely in that special everything-proof stone she has?)

...come to think of it, the location of the gates are from her diary. Is there any reason she'd even have to write down the actual location of her gate? One of the themes of the tomb is that Kraagor's sacrifice was forgotten, and the fact that nobody even remembers where he died would fit in with that. I mean, narratively "haha, oops, Serini's gate isn't here! You have no idea where it is!" doesn't really work storywise, but it's a thought.

[Grey_Wolf_c]

I mean, let's be real, as someone said above, Serini was a rogue. There's a 0% chance that the gate is actually behind any of those doors.

Serini wanted a monument to physical strength in remembrance of Kraagor. It would defeat the purpose of the tomb if it wasn't approachable as a massive slug fest.


(Come to think of it, why are the gates even accessible? Why not encase it entirely in that special everything-proof stone she has?)
Repairs.


One of the themes of the tomb is that Kraagor's sacrifice was forgotten
No, it is not. Its theme is that his sacrifice should NOT be forgotten.

Grey Wolf

[woweedd]

Serini wanted a monument to physical strength in remembrance of Kraagor. It would defeat the purpose of the tomb if it wasn't approachable as a massive slug fest.


Repairs.


No, it is not. Its theme is that his sacrifice should NOT be forgotten.

Grey Wolf
I eman, logically speaking, it would make the most sense to have it so that, say, touching a certain set of doors in a certain order opens the path. That way, you can access it, and your enemies can't. But, then, people don't act perfectly logically, and, if any of the Scribblers had, things wouldn't have gotten this bad in the first place.

[Grey_Wolf_c]

I eman, logically speaking, it would make the most sense to have it so that, say, touching a certain set of doors in a certain order opens the path. That way, you can access it, and your enemies can't. But, then, people don't act perfectly logically, and, if any of the Scribblers had, things wouldn't have gotten this bad in the first place.

A complex enough pattern that can't be found by coincidence would also need to be written down. That's just asking for her diary to fall into the wrong hands and give the information away. The doors are secured, in a way that Kraagor would approve of, which is also perfectly logical.

Grey Wolf

[RatElemental]

A complex enough pattern that can't be found by coincidence would also need to be written down. That's just asking for her diary to fall into the wrong hands and give the information away. The doors are secured, in a way that Kraagor would approve of, which is also perfectly logical.

Grey Wolf

It could be something as simple as a lever at the end of all doors of a certain shape, all of which need to be flipped to open the path at the end of a door that's a different color than the others. An intruder might not make the connection and even if they do they still have to fight their way through many doors anyway, and it's easy to keep track of if you need to visit the gate for some reason.

[Grey_Wolf_c]

It could be something as simple as a lever at the end of all doors of a certain shape, all of which need to be flipped to open the path at the end of a door that's a different color than the others. An intruder might not make the connection and even if they do they still have to fight their way through many doors anyway, and it's easy to keep track of if you need to visit the gate for some reason.

I don't dispute that Serini might have put in a roguish shortcut to access the gate, BTW, just dispute that it's the only way to get to the gate. I think that Serini would not disrespect Kraagor's memory by not making the gate reachable via sheer force of arms.

Grey Wolf

[Kish]

One of the members of the Order of the Scribble was obsessed with deception. He wasn't a rogue and his Gate has already been destroyed. There's no indication that Serini valued deception at all; to what extent she had a theme in the Order of the Scribble, it currently appears to have been--what the others needed and ultimately didn't have--cooperation. "A rogue must be super-deceptive" is the same kind of shallow stereotype that dictates that Roy should be acting more like Thog; Serini could easily have been a Lawful Good rogue with 0 ranks in Bluff.

(Though Haley and Vaarsuvius both cleave to their class stereotypes.)

[Chronos]

In fact, we know that she was Lawful Good, because she considered multiclassing to paladin.

Though, just for the sake of completeness... She does appear not to have any ranks in Bluff, but then, that's just what it would look like, if she were really, really good at bluffing.

[Aquillion]

No, it is not. Its theme is that his sacrifice should NOT be forgotten.Check out the inscription on his statue (http://www.giantitp.com/comics/oots1039.html). It doesn't say that his sacrifice shouldn't be forgotten, it says that it was forgotten.

I think it's very likely that the version of the story passed down by the Paladins was not 100% accurate, or that she didn't fully express her feelings at the time.

(I also like this theory because that would mean the statue was a clue and that Team Evil completely missed the incongruously odd nature of the inscription when they were vandalizing it.)


In fact, we know that she was Lawful Good, because she considered multiclassing to paladin.Her party members told her it was a bad idea, which might mean that she didn't actually meet the requirements. And it's also possible (even likely) that she wasn't serious.

[Kish]

Kraagor said "I don't know if you're the paladin type, Serini." That should not be expanded to "her party members, plural, told her it was a bad idea."

[hroþila]

Serini was the one to suggest the five of them should retire, pick a gate and never contact the other Scribblers, but she kept adventuring and she stayed in touch with Girard. We know very little of her, but what we do know doesn't point towards Lawful.

[Ruck]

Serini was the one to suggest the five of them should retire, pick a gate and never contact the other Scribblers, but she kept adventuring and she stayed in touch with Girard. We know very little of her, but what we do know doesn't point towards Lawful.

Which makes it pretty interesting that, for all Girard's suspicions, Soon was the only one who didn't violate the oath to never make contact.

[Knaight]

Check out the inscription on his statue (http://www.giantitp.com/comics/oots1039.html). It doesn't say that his sacrifice shouldn't be forgotten, it says that it was forgotten.

That's what the inscription says, yes. Between the emotional context of the scene in which that inscription was made and how pointing out that someone was forgotten by literally engraving it in stone has a certain level of inherent contradiction to it which suggests not to take it at face value. Which in this case works out to "he was a hero who's sacrifice shouldn't be forgotten, but the nature of his heroism meant it had to be", from the embittered position of other people working in that same secrecy.

[Fish]

Serini wanted a monument to physical strength in remembrance of Kraagor. It would defeat the purpose of the tomb if it wasn't approachable as a massive slug fest.
I should point out that in Shojo's story (http://www.giantitp.com/comics/oots0277.html), it says:

She decided that she would build a tomb for Kraagor, and fill it with the nastiest monsters in the world, to reflect his belief in the power of physical might.
It does not say she built the tomb at, over, around, or even near the Gate. The likelihood, yes, is that the tomb is intended to guard the Gate. The art suggests that both would be found in a cold climate. It is perfectly sensible to go to the trouble and expense of such a complex construction in order to protect the Gate.

Nevertheless, the comic doesn't say so explicitly. If one is looking for a loophole, this might be it.

[woweedd]

I should point out that in Shojo's story (http://www.giantitp.com/comics/oots0277.html), it says:

It does not say she built the tomb at, over, around, or even near the Gate. The likelihood, yes, is that the tomb is intended to guard the Gate. The art suggests that both would be found in a cold climate. It is perfectly sensible to go to the trouble and expense of such a complex construction in order to protect the Gate.

Nevertheless, the comic doesn't say so explicitly. If one is looking for a loophole, this might be it.
I would say that's the dumbest theory Grey_Wolf's heard, but, let's be honest: It's Grey_Wolf. He's probably heard dumber theories today. Point is, neat idea, but it's a real stretch.

[Grey_Wolf_c]

I would say that's the dumbest theory Grey_Wolf's heard, but, let's be honest: It's Grey_Wolf. He's probably heard dumber theories today. Point is, neat idea, but it's a real stretch.

Thor damnit. I may have (with your permission) to add this to my sig somewhere.

Grey Wolf

[woweedd]

Thor damnit. I may have (with your permission) to add this to my sig somewhere.

Grey Wolf
Permission granted. Now please excuse me to Squee.

[RatElemental]

I don't dispute that Serini might have put in a roguish shortcut to access the gate, BTW, just dispute that it's the only way to get to the gate. I think that Serini would not disrespect Kraagor's memory by not making the gate reachable via sheer force of arms.

Grey Wolf

Serini was also the voice of reason to the scribblers. I find it plausible that she'd understand the gravity of what she was defending, and would not be above putting a trick on things. Especially since she was defending something that Kraagor died in order to fix in the first place.

[woweedd]

Serini was also the voice of reason to the scribblers. I find it plausible that she'd understand the gravity of what she was defending, and would not be above putting a trick on things. Especially since she was defending something that Kraagor died in order to fix in the first place.
Notably, as others have pointed out, she seems to have done what they should have been doing all along: pool their strengths. The multiple doors, only one, or none, of which have the answer channels Girard's fondness for illusions and deception. The memorial aspect matches Soon's obsession with honor. The gate has magic on it to keep Xykon from being able to bypass it, like Dorukan's. The tomb is in the middle of harsh wilderness, and filled with deadly monsters harnessing the most powerful parts of nature, showing a respect for nature, as Liriam would have, and Kraggor..Well, that's obvious. She took all their "themes," illusions, honor, arcane magic, nature, and physical strength, and pooled them to create an uber dungeon. If only they could have done that with all of them, things might not have gotten this bad. Almost as if the comic's trying to send a message about the importance of cooperation and not letting petty grudges get in the way of the greater good.

[MartianInvader]

Though, just for the sake of completeness... She does appear not to have any ranks in Bluff, but then, that's just what it would look like, if she were really, really good at bluffing.
Now you've filled my head with a whole bunch of unlikely plot twists (Serini was working for the Dark One all along! Serini was Loki in disguise all along! Serini was from the Snarl's world all along!)

[georgie_leech]

Now you've filled my head with a whole bunch of unlikely plot twists (Serini was working for the Dark One all along! Serini was Loki in disguise all along! Serini was from the Snarl's world all along!)

Serini was Kraggor all along :smallamused:

[Manga Shoggoth]

Serini was Kraggor all along :smallamused:

I was going to suggest that Oona is Serini, but your suggestion takes a lot more skill...

[Fish]

I would say that's the dumbest theory Grey_Wolf's heard, but, let's be honest: It's Grey_Wolf. He's probably heard dumber theories today. Point is, neat idea, but it's a real stretch.
As I said, I think it unlikely myself. It’s not the kind of wordplay gotcha game that’s likely to come up. There’s no good reason to think ithe theory is true. However, it’s as well to remind ourselves where the text ends and where our assumptions — even very confident ones — begin.

[Aquillion]

Also, the massive slugfest can still serve to murder anyone who comes looking for the gate, adding to its defenses while putting the actual gate elsewhere. Or the actual hidden door could have even more monsters behind it.

But either way, I totally don't think the gate is behind any of those doors. Why would she do that? Kraagor's belief in physical strength doesn't require being stupid. And putting the gate in an obvious place is stupid. If you're going to set up a big "THE GATE IS HERE" display, you put the gate somewhere else.

Also, again - her statements about how she's defending her gate come from her. If we're assuming she's willing to be deceptive - and, again, regardless of her alignment, she was a rogue, so she absolutely was - there's no reason she'd give anyone the full details, especially knowing that one of them is a Paladin who won't lie to others, making him a security risk.

If she's going to add a twist at the end to add another layer of security, she's obviously not going to tell anyone. That would defeat the whole purpose of it.

(There's also a narrative issue to consider here. Why have the defenses on Kraagor's gate be built by Serini if they're going to be exactly the defenses Kraagor himself would build? What's the narrative purpose of having an extra member of the Order of the Scribble, and to have Serini, a rogue, building the gate for Kraagor, the fighter? I mean, there's a lot of possible answers to that, and "it's just part of the story for other reasons" is valid, but it's something to consider.)

[Kish]

and, again, regardless of her alignment, she was a rogue, so she absolutely was
Again:

One of the members of the Order of the Scribble was obsessed with deception. He wasn't a rogue and his Gate has already been destroyed. There's no indication that Serini valued deception at all; to what extent she had a theme in the Order of the Scribble, it currently appears to have been--what the others needed and ultimately didn't have--cooperation. "A rogue must be super-deceptive" is the same kind of shallow stereotype that dictates that Roy should be acting more like Thog; Serini could easily have been a Lawful Good rogue with 0 ranks in Bluff.

(Though Haley and Vaarsuvius both cleave to their class stereotypes.)

Bolding the parts you're ignoring to act like my post was just "she might have been Lawful Good."

[Morty]

Last I checked, Bluff was a class skill for rogues, not a class feature. Serini could have easily gone through 20+ levels without ever putting ranks in it.