(Poll) Which makes more sense? Success on a 12 or Success on a 1?

[Souju]

Poll: https://docs.google.com/forms/d/e/1FAIpQLSfwkaeZ2biIQfYltbolPmmXjM_IUTg5TZafy6rK_Xzs z7qE3A/viewform?usp=sf_link

Explanation: So recently I've been involved in a bit of development.

However, an argument has ensued between me and my co-developer on his system (we're both making different systems using the same lore background)

He wants his system to have 1s as successes and 12s as failure (d12 is being used as the dice-of-choice)

Now, as for his reasoning I can't really get him to provide a satisfactory response other than d12s being underrepresented, and it doesn't bother me too much which dice is chosen. There are pros and cons to everything yadda yadda yadda.

My personal sticking point is that he wants low numbers to be a success and high numbers to be a failure on these dice rolls, especially "to hit", all the while maintaining additive damage rolls for things like damage (So you'd crit succeed on a 1 and crit fail on a 12).

He gives the following as an example of how this works (keep in mind d12s are being used, not d6s. I made that mistake as well)


Ejariak the level 6 thief is missing two wounds and attempts to pick a lock, a class skill for him, he rolls: 1,2,3,4,5,6. He gets 2 successes, in the 1 and 2, and fails to pick the lock. He gets angry and attempts to attack his friend, defense 2. Ejariak’s melee skill is 7, and because of his wounds and his friends defense he needs a 3 or less to hit. Ejariak rolls a 1 critting his friend, negating his ability to dodge. Ejariak finishes his betrayal by rolling damage, normally 2d6+4; he rolls 2 and 6, doing a total of 9 damage that ignores armor due to critical strike.

So my argument to him is that players honestly like bigger number=success when rolling dice. But he isn't really hearing it, and doesn't seem to want to actually look at other games to see why no one does this.

Thus, the poll. I'm asking what other people think about the idea of success on a 12 sided die. Simple question, 4 possible responses.

Thank you for your time.

[MoiMagnus]

"Roll under" is actually pretty common. Percentile systems are an example, and if my memory is correct The Dark Eye uses "roll under" for d20 and is one if the most played games in Germany.

But I agree that it is not intuitive at all for players who mostly played D&D or d20system games, which is most modern players.

One of the underused advantages of "roll under" is that you can vary the dice. For example, you can say that a character that I'd blessed rolls a d8 instead of a d12, and the rules will just work.

[Vahnavoi]

Rolling under for one thing while rolling over for another makes just as much sense as always rolling over or always rolling under.

Switching between them only fails to make sense if you sometimes roll under and sometimes roll over for the same thing, and even that's debatable. If given enough time, I could likely come up with a dice game with that trait that would be both interesting and suitable for use as a roleplaying game mechanic.

Actual reasons to use roll-under:

1) Bounded accuracy. If you're supposed to stay within a certain band of probability anyway, it makes sense your choice of die and numbers would reflect this. Nearly all percentile-based system exemplify this.

2) Working with small numbers is faster than working with large number. This is related to the above in the sense that it's a reason to use bounded accuracy to begin with. With roll under, values approach zero, rather than balloon upwards without an end.

Actual reasons to use d12, specifically:

1) 12 is neatly divisible by 6, 4, 3 and 2, something that doesn't apply to, say, 10 or 20 or 100. All of these numbers are also easy to deal with in one's head. Used smartly, this allows for use of division and multiplication in ways that many games sleep on.

2) Again, working with small numbers.

3) It's a Platonic solid, hence easy to make into a fair dice (shares this quality with d4, d6, d8 and d20)

4) Has double the spread of d6 while retaining all useful properties of d6.

[King of Nowhere]

higher numbers are better.

it is good to maintain consistency. having to roll low for some things and high for others is confusing.
so, for damage you already want to roll high. you may as well want to roll high for everything.

i remember playing 2e d&d that uses high numbers for something and low numbers for something else. it was confusing. especially when you have to factor in bonuses. sometimes you add, sometimes you subtract? no thanks. d&d 3e streamlined it into every roll being dice + modifiers, and it was a great improvement.

[Morgaln]

One reason some games (like The Dark Eye) use roll under for certain things is that they use the stats of the character as target numbers, most commonly attributes. It means that GMs don't have to make up difficulties. In order to succeed at a strength check, roll lower than your strength. In order to perform a feat of agility, roll under your agility, and so on and so forth. It is a very simple system, that is intuitive and immediately allows players to understand how a higher stat is better. It also makes GM bias less of a problem.

It's certainly not a perfect system and not necessarily better than others, but it has its advantages.

[stoutstien]

I have a soft spot for rull under systems but it's hard for them to gain traction in the general player pools.

[Vahnavoi]

One reason some games (like The Dark Eye) use roll under for certain things is that they use the stats of the character as target numbers, most commonly attributes. It means that GMs don't have to make up difficulties. In order to succeed at a strength check, roll lower than your strength. In order to perform a feat of agility, roll under your agility, and so on and so forth. It is a very simple system, that is intuitive and immediately allows players to understand how a higher stat is better. It also makes GM bias less of a problem.

It's certainly not a perfect system and not necessarily better than others, but it has its advantages.

None of those positive traits follow from rolling under. They follow from having target number of your rolls right there on yout character sheet. The same can be achieved in a roll over system, but it requires some thinking of how to display information. For example, saving throws in lot of OSR games and old D&D are roll over, but this means the number on your character sheet goes down as the character improves. That's less intuitive than rolling under a number that goes up as a character improves.

If target numbers or modifiers to them are hidden, a roll under system has all the same potential for game master chicanery as a roll over system. In general, changing minor things of how dice are rolled has minimal effect on any biases. Roll under and roll over are mathematically equivalent, choice between them is largely about what is easier to process, not some tangentially related thing.

[Kurald Galain]

it is good to maintain consistency. having to roll low for some things and high for others is confusing.
This. It doesn't matter which one you pick, but being inconsistent will confuse (some) players.

[KorvinStarmast]

Rolling under for one thing while rolling over for another makes just as much sense as always rolling over or always rolling under.

Switching between them only fails to make sense if you sometimes roll under and sometimes roll over for the same thing, and even that's debatable. If given enough time, I could likely come up with a dice game with that trait that would be both interesting and suitable for use as a roleplaying game mechanic. We used "roll under" for ability check things (ad hoc, mostly) before it went mainstream in Old D&D and in AD&D 1e. Am currently in a game where Roll Under is the standard approach (it's called Mothership). I agree with your point about consistency.
(Rest of your post appreciated), but this bit is why I've used D12's A Lot for years:

1) 12 is neatly divisible by 6, 4, 3 and 2, something that doesn't apply to, say, 10 or 20 or 100. All of these numbers are also easy to deal with in one's head. Used smartly, this allows for use of division and multiplication in ways that My favorite d4's have 12 sides on them, for example.

It doesn't matter which one you pick, but being inconsistent will confuse (some) players. True.

For the OP: I'd suggest the roll under if your overall system uses roll under as its approach.

[GloatingSwine]

Ultimately whether success happens on big numbers or small numbers is aesthetic.

One thing though is that it looks like the game is using some number of dice, the example seems to have rolled 6D12. The more dice you're rolling at once the more valuable it is to use D6s because it's less mental load to look for the dot patterns you want rather than to care about the actual numbers.

Also they're cheaper and easier to get in bulk.

[Souju]

I appreciate a lot of the feedback so far. These have been the best replies I've gotten thus far and I'm glad I keep coming here for esoteric questions like this.

Most of the feedback I've gotten from other places is that they prefer higher numbers to equate to success (and the poll shows me that 60% of people think this is the case as well. Though with a current sample size of only 20 I don't think it's indicative yet. I wish I knew how to reach more people.)

The other guy didn't really take kindly to the suggestion that it's better for players to go with high success low failure and "doesn't appreciate being told to copy other systems" (I didn't tell him this. I told him to research other systems) and that "if people just copied each other we'd still be using THAC0" (which is...wrong. At least I'm pretty sure it's wrong.)



Ultimately whether success happens on big numbers or small numbers is aesthetic.
Well, there's a bit of psychology that goes into but you're largely correct.


One thing though is that it looks like the game is using some number of dice, the example seems to have rolled 6D12. The more dice you're rolling at once the more valuable it is to use D6s because it's less mental load to look for the dot patterns you want rather than to care about the actual numbers.

Also they're cheaper and easier to get in bulk.
I don't think my co-developer has "mental load" in the forefront of his mind. I wish he did. It's a big problem with his development style.

I personally use d6s for the system I'm developing (which is a pretty shameless amalgamation of World of Darkness, Shadowrun, and Ars Magica) but my buddy is DEAD SET on using d12s.

[Amnestic]

I definitely prefer higher number=better, but I can certainly see the logic and don't dislike lower number=better. I would just ask that the system as a whole be consistent with it, and doesn't mix+match. If rolling low is good, I'd want it to always be good for that system, that way I can keep it in my head.

Also I like that he's attached to the d12 honestly. It's a nice dice.

[Anonymouswizard]

"Roll under" is actually pretty common. Percentile systems are an example, and if my memory is correct The Dark Eye uses "roll under" for d20 and is one if the most played games in Germany.

Also everything Steve Jackson Games seems to put out, which is not a minor portion of the Anglosphere TTRPG market (once we discount D&D's Monopoly).


I have a soft spot for rull under systems but it's hard for them to gain traction in the general player pools.

GURPS spent what, decades as one of the most popular RPGs out there? And is still going strong today? Admittedly a lot of that has to do with every other game having less detailed sourcebooks, but the game itself isn't exactly hated.

Which brings us to the actual answer: your core resolution mechanic matters less than everything else around it barring certain very specific instances (e.g. Dread). People don't play D&D for the roll over resolution mechanic, they play D&D because 1) they like fantasy and B) everybody else is playing D&D.

Anyone who can't grok the difference between things like success rolls and damage rolls is probably asking what die to roll for skill checks again. Really as long as you're more intuitive than d666 you should be fine (and it's generally easy to translate between roll under and roll over TNs anyway).

The big, big advantage to a roll under system is that they can be slightly more resilient to things like loaded dice (if you do 'roll under roll high'). This isn't important enough to truly worry about.

[Rockphed]

more intuitive than d666 you should be fine

I have to ask if this is a real roll instruction you have seen, or if you are plucking a horrible, pathological example from the aether? Aside from some obviously cursed implement of torture, how would your roll a d666 by combining other dice? 666 is 9 * 2 * 37. 9 and 2 are easy enough to get dice with that number of unique sides, but 37 seems impossible.

Edit: Honestly, I would stop arguing about roll-over vs roll-under and focus criticism on other parts of his system. Aside from the hassle of changing the instructions everywhere, changing from roll-over to roll-under is not hard. Try to convince him to include a tag everywhere he has roll-under math in the system so it is easy to double check that he did it right (it will also make it easy to find all of them to change if he eventually changes his mind, but sell it on the checking math).

[Vahnavoi]

The more dice you're rolling at once the more valuable it is to use D6s because it's less mental load to look for the dot patterns you want rather than to care about the actual numbers.

Also they're cheaper and easier to get in bulk.

Cubes also pack more neatly than dodecahedrons.

Of course, these only apply if you're using genuine physical dice. If using a digital roller or planning for a virtual tabletop, simple physics-based reasons such as these lose meaning.

As a side note, a point about physical dice: cheap hobby dice are often poorly balanced. One of the reasons to occasionally roll under and occasionally roll over with the same set of dice is to prevent such physical flaws from having an undue effect on a game. It can also mitigate or eliminate the incentive to cheat by using weighed dice.

I don't bring this up because I think it's very compelling on its own, I bring this up because it's the sort of thing you'll never catch if you're only thinking on the level of "roll over better because big numbers good". When it comes to dice as a physical system, roll over versus roll under isn't just personal preference or aesthetics.

[MoiMagnus]

I have to ask if this is a real roll instruction you have seen, or if you are plucking a horrible, pathological example from the aether? Aside from some obviously cursed implement of torture, how would your roll a d666 by combining other dice? 666 is 9 * 2 * 37. 9 and 2 are easy enough to get dice with that number of unique sides, but 37 seems impossible.

Edit: Honestly, I would stop arguing about roll-over vs roll-under and focus criticism on other parts of his system. Aside from the hassle of changing the instructions everywhere, changing from roll-over to roll-under is not hard. Try to convince him to include a tag everywhere he has roll-under math in the system so it is easy to double check that he did it right (it will also make it easy to find all of them to change if he eventually changes his mind, but sell it on the checking math).

I'm assuming that by d666 they're talking about "roll 3d6, pretend you rolled 3d10 and consider the result as a 3 digits number" of In Nomine Satanis and Magnita Veritas (though from what I read, the details of how it works changed in the newest edition, additionally I only ever played it in conventions so my understanding is very limited)

[LibraryOgre]

My preference is "Highest successful".

If I understand the example, with a base skill of 7, normal (non-penalized) would mean a 1-7 is a success; penalties would lower that.

Now, let us say that I have a 4 in Melee combat, and you have a 7. If I roll a 1-4, I succeed, but if you roll a 5-7, you beat me, no questions asked... because you succeeded at a level I cannot match. Even if I roll a 4, and you roll a 5, your roll is better than mine.

It makes for easy opposed comparisons, maintains "high is good", without going all the way to "highest die number is best." It also makes penalties on this a bit more intuitive, I think... if I have a -3 on my 7, I am now rolling under a 4.

Another alternative is a "DC 12" system... roll your die and add your skill. d12+7 means I need to roll a 5 or better to succeed.

[Anonymouswizard]

I have to ask if this is a real roll instruction you have seen, or if you are plucking a horrible, pathological example from the aether?

I mean the-


I'm assuming that by d666 they're talking about "roll 3d6, pretend you rolled 3d10 and consider the result as a 3 digits number" of In Nomine Satanis and Magnita Veritas (though from what I read, the details of how it works changed in the newest edition, additionally I only ever played it in conventions so my understanding is very limited)

Kind of, I'm talking about the system from the SJG translation/adaptation In Nomine, not the original French version.

Roll 3d6, two similar and one distinct sum the first two and compare to the target number (normally Characteristic+Skill/Song), equal or under is a success while over is a failure. Then check the distinct d6 this is the degree of success. Don't forget to apply the correct modifiers to your target number and degree of success. If you rolled all ones or all sixes check your alignment to see if you got a stroke of luck or misfortune.

I mean it works and is relatively fast, but you need to get it first and it's not entirely intuitive. That alternative version of d666 sounds blessedly simple.

[MoiMagnus]

I mean the-



Kind of, I'm talking about the system from the SJG translation/adaptation In Nomine, not the original French version.

Roll 3d6, two similar and one distinct sum the first two and compare to the target number (normally Characteristic+Skill/Song), equal or under is a success while over is a failure. Then check the distinct d6 this is the degree of success. Don't forget to apply the correct modifiers to your target number and degree of success. If you rolled all ones or all sixes check your alignment to see if you got a stroke of luck or misfortune.

I mean it works and is relatively fast, but you need to get it first and it's not entirely intuitive. That alternative version of d666 sounds blessedly simple.

Now that you're describing it, it does sound familiar, so I might have blocked some memories about how it worked, and maybe the initial version in French was not as simple as I remembered...

[Rockphed]

I mean the-



Kind of, I'm talking about the system from the SJG translation/adaptation In Nomine, not the original French version.

Roll 3d6, two similar and one distinct sum the first two and compare to the target number (normally Characteristic+Skill/Song), equal or under is a success while over is a failure. Then check the distinct d6 this is the degree of success. Don't forget to apply the correct modifiers to your target number and degree of success. If you rolled all ones or all sixes check your alignment to see if you got a stroke of luck or misfortune.

I mean it works and is relatively fast, but you need to get it first and it's not entirely intuitive. That alternative version of d666 sounds blessedly simple.

Somehow In Nomine never fails to disappoint. That is at once the most pathological example of dice rolling I have ever seen and a real example. It even has different high vs low behavior depending on which character makes the roll. I need to get the In Nomine books and get together a group. I don't think I can sell it to my wife as the first RPG to get my children playing, both from its dissonance with our theology* being confusing for children and from its mechanical complexity.

*Technically almost all RPGs have significant dissonance with the theology of most religions. In Nomine just makes that dissonance a central fixture of the system and setting.

[Anonymouswizard]

Somehow In Nomine never fails to disappoint. That is at once the most pathological example of dice rolling I have ever seen and a real example. It even has different high vs low behavior depending on which character makes the roll. I need to get the In Nomine books and get together a group. I don't think I can sell it to my wife as the first RPG to get my children playing, both from its dissonance with our theology* being confusing for children and from its mechanical complexity.

*Technically almost all RPGs have significant dissonance with the theology of most religions. In Nomine just makes that dissonance a central fixture of the system and setting.

I absolutely adore In Nomine. None of that is to do with the mechanics, or even that I'm able to talk about here.

[Mastikator]

I don't think my co-developer has "mental load" in the forefront of his mind. I wish he did. It's a big problem with his development style.

I personally use d6s for the system I'm developing (which is a pretty shameless amalgamation of World of Darkness, Shadowrun, and Ars Magica) but my buddy is DEAD SET on using d12s.

I recommend playtesting, preferably with newbs and your cocreator. Make a note of everything that isn't intuitive for the players, every time playing forget a rule that they already used, etc. It might bring mental load to attention, it would definitely find weak spots too.
A playtest shouldn't involve a story, just the stuff you want to playtest, like character creation and then combat, then a skill challenge.

[gbaji]

I tend to favor "roll under" as a resolution mechanism. As a couple people have pointed out, it's simpler and more intuitive. Assuming we have stats on our character sheet, and assuming we have skills on our character sheet, those are generally represented as positive numbers in which the higher the value the better. If the game system also normalizes the actual stats/skills to the dice being used, then a very simple "roll under your stat/skill" mechanism can be used. No additional math need be employed (though, of course, there may be situational bonuses/negatives applied by the GM).

All "roll over" systems require some additional math in the process. You take a target number, you roll the dice, you add your skill/stat to the dice, and need to roll higher than the target number is the most common method. The biggest point is that there is *always* a target number that has to be determined first, then the dice are rolled and the skill is then used as an adjustment to the die roll (or precalculated via subtraction from the target number, but the result is the same). This flips some concepts around from roll-under where the "default" is simply "roll under your skill/stat/whatever", and additions/subtractions may be applied (but do not need to be).

The classic form is pecentile based systems. If I have a 57% in my skill, I literally just roll the dice. If I roll equal or under the skill value written on my sheet, I succeed. And additional modifications are optional, so the "simpliest case" is easily and quickly handled by default.

The same basic concept can apply for point based skills. You have X points in a skill. You roll under X, you succeed (obviously dice and skill point range need to be calibrated for the system). As your skill increases (more points), your chance to succeed just goes up. No need to do any additional calculations. Roll-under has the advantage that "my stuff" adds to the target number, while "the other guys stuff" subtracts. Meaning you can calculate the roll needed ahead of time. To do the same pre-calculation with a "roll over" you are subrtracting your skill from a target number. Most players like to know what they need to roll before rolling, so that's usually what they do. The only way to retain a "My stuff adds", is to just roll the die first, then add your skill. Which tends to not be as comfortable for players.

I think the aestetics are pretty similar in either direction. I've used both, so it's not that big a deal. But I think most of the "roll high" positions are based more or less just on what someone is already used to (D&D most commonly). And IME, to make "roll high" work for all things in a game requires either doing more math at the time you roll the dice, or precalculating some additional values and writing them on the sheet ahead of time.


Having said all of that, roll-high does work better in opposed roll situations though. A very simple "roll die and add to value and highest result wins" is a very very simple mechanic (Talisman uses this, as do lots of games). So there's something to be said about that. I will note, however, that the Talisman also frequently uses "roll under stat" as a mechanic as well though (cause that's a lot easier than "subtract stat from 6 and roll over"). So if you're looking for one method to use consistently, that's maybe not the greatest example. It does highlight the point I made earlier that "higher stat/skill is better" does tend to synergize with a very simple "roll under your stat/skill" resolution method.

[King of Nowhere]

The other guy didn't really take kindly to the suggestion that it's better for players to go with high success low failure and "doesn't appreciate being told to copy other systems" (I didn't tell him this. I told him to research other systems) and that "if people just copied each other we'd still be using THAC0" (which is...wrong. At least I'm pretty sure it's wrong.)


ah, but by rejecting the roll over he's the one basically copying thac0.
i think the biggest contribution thac0 gave to the rpg community is showing that roll over works better



I don't think my co-developer has "mental load" in the forefront of his mind.
my buddy is DEAD SET on using d12s.

you are not painting a flattering picture of your buddy. from your description, he comes across as a sort of virtuoso trying to make a needlessly complex system to appease his own sense of art, disregarding any practical consideration.



All "roll over" systems require some additional math in the process. You take a target number, you roll the dice, you add your skill/stat to the dice, and need to roll higher than the target number is the most common method. The biggest point is that there is *always* a target number that has to be determined first, then the dice are rolled and the skill is then used as an adjustment to the die roll (or precalculated via subtraction from the target number, but the result is the same). This flips some concepts around from roll-under where the "default" is simply "roll under your skill/stat/whatever", and additions/subtractions may be applied (but do not need to be).

On the down side, this means that in roll under, all checks have the same chance of succeeding. in d&d breaking a normal door is 10, a sturdy door is 15, a normal reinforced door would be between 20 and 25, a bank caveau would probably be between 40 and 50.
in roll under, you always need to roll under your stat. normal door? roll under your stat. reinforced door? roll under your stat. reinforced, barred, barricaded door? roll under your stat. I'm sure the dm will step in and say breaking a bank caveau is downright impossible except maybe for a big dragon, but the point is, what you describe as a strenght of the roll-under system actually encourages to set a fixed difficulty for everything instead of making provisions for how difficult something actually is. I'd consider that a flaw.
of course you can make adjustments to account for the door being reinforced, but then you're better off with roll+modifiers.

[GloatingSwine]

Well, there's a bit of psychology that goes into but you're largely correct.


People like it when number big, but if you use roll-under then good stats are big in a way that directly interacts with your roll instead of via derived modifiers.

Which people like because number big in directly applicable way.


ah, but by rejecting the roll over he's the one basically copying thac0.
i think the biggest contribution thac0 gave to the rpg community is showing that roll over works better


THAC0 was still roll-high, the annoyance of it was that you used a modifier based on a property of the target on a dice rolled by the active player to hit a target based on the acting character which was ass backwards compared to modifying a roll of the active player with a property of the acting character to hit a target based on the target.

Roll-over sometimes means that small number good though, like with 2e's version of saves. Smaller target to roll over = better save.


in roll under, you always need to roll under your stat. normal door? roll under your stat. reinforced door? roll under your stat. reinforced, barred, barricaded door? roll under your stat. I'm sure the dm will step in and say breaking a bank caveau is downright impossible except maybe for a big dragon, but the point is, what you describe as a strenght of the roll-under system actually encourages to set a fixed difficulty for everything instead of making provisions for how difficult something actually is. I'd consider that a flaw.

The system described appears to include dice pools, so a harder check requires more successes. As long as you math this out and make the numbers required appropriate it's fine to do.

[False God]

I personally think "roll high" with a static DC where everything over it success is more intuitive.

Roll Under works best IMO, with a percentile die system and your skill is the range in which you have to roll within. The greater your skill, the more it expands into the higher numbers, the more difficult the check, the more it reduces your range.
-Sure, you can apply it to any die, but IMO, the sheer range of numbers on a percentile die and the likelihood of rolling more average numbers rather than extremes makes it work better.

However, you may consider that rather than a "number" based success system, a d12 with a visual success key may be more appropriate for this game. On a single die, it doesn't matter if it's 1,2,3 or 10,11,12 are successes. They each have the same chance of coming up as any other number. You could replace them with success symbols, rather than numbers, place them at any location on the die, and have everything else be failure.

The upside to a symbolic die is you now only need to remember two(really, one) things, rather than which unique numbers are success and failure. You only need to remember the success symbol, and frankly you can leave the rest of the die blank.

[Buufreak]

He gets angry and attempts to attack his friend...

I understand it is just an example. I understand it is in effort to show all the things and ways the dice rolling can do and is used for. But by gods, this is perhaps the silliest of bits that makes the least sense of it all.

Like, how often do we tell stories about the time you went out with your friends, have a great time, drop your keys by accident, and instead of picking them up, you shoot Carl?

[Thrawn4]

I would suggest to focus on making the rules clearer.
Why does the 1 and the 2 count as success in the first example? Is 2 the default difficulty?
Why does it matter that the thief has two wounds? No reason is stated.

Although I prefer roll high, roll low can make more sense, for example when it is based on a given value (you have lore 3, now you succeed on 1,2,3 ).

[Vahnavoi]

ah, but by rejecting the roll over he's the one basically copying thac0.
i think the biggest contribution thac0 gave to the rpg community is showing that roll over works better.

GloatingSwine beat me to it, but still: this is completely wrong. THAC0 is roll over and the system described by the OP bears no resemblance to it.

What sets "To Hit Armor Class 0" of 2nd edition AD&D apart from "(Base) Attack Bonus" of d20 is how it displays information. On a basic level, it attempts to show the number you need to roll on the physical dice on the character sheet. This isn't a bad idea. What makes it difficult for people is that AD&D uses descending armor class that goes into the negatives. This leads to situations where a player needs to subtract a negative number from the number on their sheet.

By contrast, d20 (B)AB focus on displaying and adding up modifiers is part of why the modifiers eventually overtake die rolls and rolling becomes more of a formality than anything.

Note: it would be super easy to set up a more intuitive THAC0 just by using ascending armor class (which would mean a player only ever adds enemy armor class to the number on their sheet) or using descending armor class without negatives (wich would mean a player only ever substracts enemy armor class from the number on their sheet).

[GloatingSwine]

I understand it is just an example. I understand it is in effort to show all the things and ways the dice rolling can do and is used for. But by gods, this is perhaps the silliest of bits that makes the least sense of it all.

Like, how often do we tell stories about the time you went out with your friends, have a great time, drop your keys by accident, and instead of picking them up, you shoot Carl?

Carl knows what he did.



By contrast, d20 (B)AB focus on displaying and adding up modifiers is part of why the modifiers eventually overtake die rolls and rolling becomes more of a formality than anything.

Part of it, but you can correct that by having someone in the office whose job is to smack people on the back of the knuckles with a wooden ruler whenever they try and add a new modifier.

It's also because D&D has strong vertical progression for things that aren't known spells, you get increasingly better at a few things rather than horizontal where you don't get much better at individual things but become proficient in more different things which other systems encourage, and because it's generally all designed to hinge on a single roll of one die which means everything that modifies success needs to apply all at once to one axis instead of eg. a dice pool where you can modify number of dice or the target number or have a floating modifier or partial reroll or many many other ways to fiddle with the outcome)

[Anonymouswizard]

Part of it, but you can correct that by having someone in the office whose job is to smack people on the back of the knuckles with a wooden ruler whenever they try and add a new modifier.

It's also because D&D has strong vertical progression for things that aren't known spells, you get increasingly better at a few things rather than horizontal where you don't get much better at individual things but become proficient in more different things which other systems encourage, and because it's generally all designed to hinge on a single roll of one die which means everything that modifies success needs to apply all at once to one axis instead of eg. a dice pool where you can modify number of dice or the target number or have a floating modifier or partial reroll or many many other ways to fiddle with the outcome)

Yeah, another effect of roll under is that the system tends to soft cap stats/skills at roughly the maximum die result (higher if negative modifiers are common, lower if there's automatic failure).

I find your dice pool example interesting, because oWoD had the multiple axes of difficulty and EVERY successor has been considered to have better rules for reducing it to one. Sure it's more interesting in theory, but in practice generally just slower.

[LibraryOgre]

Like, how often do we tell stories about the time you went out with your friends, have a great time, drop your keys by accident, and instead of picking them up, you shoot Carl?

You are clearly not from Texas.

[warty goblin]

The differences are entirely aesthetic from a mathematical point of view. Any roll under system has a one to one mapping to a roll over system with exactly the same probabilities, and vice versa. In the most trivial case, roll under difficulty d on a dS is equivalent to roll over S - d, e.g. roll under a 7 is the same as roll over a 5 on d12.

In generality, for target x, modifier d, and dice size S, Pr(Y + d <= x) = Pr(S - x <= S - Y - d), where Y is your roll, just by basic algebra. This formula also lets you convert dice pool systems, if you want to for some reason.

The only things that really matter mathematically in a binary success/failure dice system are how granular you want the degrees of separation between different skill levels, the shape of the underlying dice cumulative distribution function (cdf), and how changes in character skill/circumstance alter that shape.

So every roll single dice, compare to a number system differs from any other system only in granularity of skill. A d12 system measures competence in quanta of 8.333%, d20 does it in 5%, and d100 in 1%.

Any 2dS + d system has roughly the same behavior as any other, because the cdfs all have the same form; there's always 1 combination for the lowest outcome, 2 for the next, and so on, up until a maximum of 2S ways to obtain the average result of S+1, after which there's always one fewer combination for the next highest outcome. This makes the impact of a modifier d non-constant, but fairly easy to calculate, in a roll under system increasing the threshold from x to x + 1 adds S - |S - X - 1| additional successful outcomes. Using larger dice, i.e. bigger values of S, increases the spread, and allows for more granularity, but the behavior is the same.

Using more dice gives increasingly curved cdfs, and so even sharper rewards for each +1 to the threshold as it approaches the average.

The other major thing you can do are dice pools, either in count the number of successes or sum the k best dice. These can behave fairly differently, as direct count success system use an underlying binomial cdf, which can be pretty skewed for small n, and the sum the best is a sum of order statistics cdf, which again is skewed. Indeed the attractive thing about these systems is that they offer increased consistency with competency (i.e. more dice) but are generally easy to threshold above, particularly the sum the best system.

Lastly, the Dark Eye uses a genuinely novel in my experience success calculation, where you do 3 roll under checks, using your skill points to essentially raise your attributes if needed to get the rolled number under the threshold. Mathematically this reduces to Pr(success) = Pr(max(0, Y1 - x1) + max(0, Y2 - x2) + max(0, Y2 - x2) <= skill points), which is a really weird function.

[Jay R]

Mathematically, either makes sense. In practice, players can and do get used to either one.

I've heard people maintain that it was "confusing" to sometimes want a high number and sometimes want a low number, but back when this was common in D&D, I never heard or saw anybody get confused.

It really makes no difference.

[gbaji]

On the down side, this means that in roll under, all checks have the same chance of succeeding. in d&d breaking a normal door is 10, a sturdy door is 15, a normal reinforced door would be between 20 and 25, a bank caveau would probably be between 40 and 50.
in roll under, you always need to roll under your stat. normal door? roll under your stat. reinforced door? roll under your stat. reinforced, barred, barricaded door? roll under your stat. I'm sure the dm will step in and say breaking a bank caveau is downright impossible except maybe for a big dragon, but the point is, what you describe as a strenght of the roll-under system actually encourages to set a fixed difficulty for everything instead of making provisions for how difficult something actually is. I'd consider that a flaw.
of course you can make adjustments to account for the door being reinforced, but then you're better off with roll+modifiers.

Correct. But that has a bit of a psychological differerence in how we view the rolls themselves.

In roll-over, the skill is basically the modifier/adjustment to the roll. The target number is what the die being used is compared to (someone with 0 or 1 skill is basically rolling against the difficulty determined by the GM, not their actual skill, right?). So if you are using a D20, you have a range from 1-20. So you determine how difficult something is by how high the target number (DC) is relative to that 20 point roll-range. The actual skill is almost secondary in that calculation.

In roll-under the skill is what is directly compared to the die being used. So in the same D20 scenario, your skill is compared to the 20 point range of possible rolls first. Then the difficulty is used as an adjustment to the roll (or to the skill). This makes things easier when there are no modifications. You can't do roll-over without generating a difficulty value (it's what you are rolling against). It's possible for "normal things" to just "roll under your skill" in roll-under and not have to do the extra math step at all. Making it, sometimes equally complex, but sometimes easier.

Mathematicaly, these can both work exactly the same. But mentally, we tend to see that there is more focus on the target number in the first case, and the actual character skill value in the second. There's also a secondary scaling effect from a game mechanic point of view, where roll over tends towards "skill vs difficulty", and in level based games can lead to a "race to the high number" scenario. It's not uncommon at higher levels in D&D for someone to have 30 points in a skill working against a 40 difficulty target number (AC, resistance, whatever). The scale of those two numbers have outstripped the range of the die itself (though that's still resulting in a 50% probability so still workable from a game mechanic point of view). Um... As a couple people have pointed out though, this leads to vertical skill development rather than horizontal.



Roll Under works best IMO, with a percentile die system and your skill is the range in which you have to roll within. The greater your skill, the more it expands into the higher numbers, the more difficult the check, the more it reduces your range.
-Sure, you can apply it to any die, but IMO, the sheer range of numbers on a percentile die and the likelihood of rolling more average numbers rather than extremes makes it work better.

Percentile dice have no different probability spread than rolling a single D20 does (or any single die). Granularity increases, but the odds of rolling a 100 is exactly the same as the odds of rolling a 50.

As I stated earlier, the primary advantage perceptually is that the skill you have is directly correlated to the die you are rolling. So if you are rolling a D20, you can immediately see how your skill stacks up based on how high it is within that 20 point range. Same exact concept works with a D6, or D8, or D12, or D100.

In roll-under the player knows at a glance how skilled they are because the number on their sheet directly correlates to the die they roll. In roll-over, they must compare the number on their sheet to some other number, which the GM will give them when attemping their skill roll (the target number). Depending on the game system, and how normalized (or not) the target numbers may be, how "good you are" may vary wildly between different skills even with the same number displayed. Heck, in D&D there are some things that you do which may encounter very high difficulties, and others which will not. So having a few skill points in some skills will make you "very good" at that skill, while other skills the same exact number of points may be "woefully inadequate". This difference actually gets very large at higher levels.



Although I prefer roll high, roll low can make more sense, for example when it is based on a given value (you have lore 3, now you succeed on 1,2,3 ).

Yup. Makes for quicker/easier "basic roll" mechanics. Lots of games have skill point/buy systems, where skills may often be reflected in a 1-6 point range. Those systems practically scream for simple roll-under D6 mechanisms.


Yeah, another effect of roll under is that the system tends to soft cap stats/skills at roughly the maximum die result (higher if negative modifiers are common, lower if there's automatic failure).

Absolutely. There are ways to make roll-over less "race to high number", but it's difficult. When the primary comparison is to the die itself, then it does tend to soft-cap things at that die value. You also tend to have modifiers (both postive and negative) that are smaller (and also scaled to the die size). So in the above mentioned d6 system, you might have skill points that range from 1-6 as your base mechanic and roll-under as the resolution. Modifiers might merely range from plus or minus 1 maybe 2 points. This gives a good resolution method, that is easy to use, doesn't reqquire much math most of the time, but also really incentivizes people to spend EXP points on increasing other skills within that range rather than increasing the ones they are already good at. In such a system, there would be no reason to ever increase a skill higher than 8 points (and that would be someone really trying to make sure they always succeed even under the worst conditions possible).

As I mentioned earier this leads towards horizontal skill development. And heck, if we add in a increase to point cost based on skiill level, we provide even more incentive (a few games do this sort of thing).

While you *can* do the exact same thing in roll-over, IME there is a tendency for increased scale of difficulty numbers. As I mentioned earlier, the die itself becomes disconnected from the calculation and it's really about the skill versus the difficulty. Which lends itself to scaling the difficulty up, and requiring the skills to scale up with them. Yes. This is 100% psychological from a game design pov. And yes, you can certainly see the same roll-under system with the GM appying -12 difficulty modifiers, thus requiring people to have super high skill levels to have a chance to succeed. I've just noticed that the higher scaling difficulty numbers tend to exist most in games with roll-over mechanisms.


I find your dice pool example interesting, because oWoD had the multiple axes of difficulty and EVERY successor has been considered to have better rules for reducing it to one. Sure it's more interesting in theory, but in practice generally just slower.

I said this in another thread, and I'll repeat here. I actually like the dice pool concept. It provides an additional axis to use, and can make things like target numbers, skill levels, modifiers, etc a bit more "interesting" mathematically. It's not "simple", but can introduce situations where the mathematical advantage of using a spell/ability that adjusts the roll/difficulty/bonus versus one that changes the number of dice being rolled can be very different. Without die polls these are all typically going to be linear addition/subtractions to a single die roll. You may have multiple things that may adjust the die results, but you're always just making an adjustment to the same roll.

So yeah. There's something about die pools that tickles my game mechanic funny bone (in a good way). Um... And I care less about roll-over/under when dice pools are involved, because usually the dice are rolled against a difficulty/target (which tends to be very normalized to the die size, with only very small adjustments possible), and skill levels and other stuff tends to increase/decrease the number of dice. Dunno. Works for me.

[BRC]

Where do the target numbers come from?

Generally, the intuitive thing is that High Number=More, so in D&D, A More Difficult Test is a higher DC, a more skilled character has a higher bonus.

as somebody already mentioned, in my experience Roll-under shows up more in systems where the characters skill is the TN, rather than a bonus to their roll. That way the more intuitive "High Skill=Better" thing comes into play, and you skip having to add bonuses to rolls. (I'd say high skill number=better is more important than high dice number = better).

I'm not quite sure how your skill system works, it looks like you're using skills to increase die pools and requiring a certain number of successes (Which is a fine way to do it), but in the attack example A skill of 7 + unspecified wounds and a friends defense somehow turns into needing a 3 to hit, implying that skills influence what you need to roll (Like a bonus) rather than how many dice you get to roll.

The most important thing IMO is that a higher number on your skill should be better, and that the math should be as simple as possible. If you're comparing to a target number, your players should be able to give you a number to compare quite easily. If you are counting successes, your players should be able to identify a success on sight without having to do additional math.



Something like "A success means rolling below your skill value - the test's difficulty, and then count the number of successes and compare that to some other number" that's probably workable, but is kind of clunky, especially since it requires the exact difficulty to be known for every roll, rather than the player being able to give a result that the DM just compares.

[Pauly]

Broadly speaking roll above or roll under are fine, as long as they are consistently applied. Mixing roll over and roll under is annoying and requires more cross checking during play to check whether this subsystem is roll over or roll under.

In roll over systems the modifiers are applied to the dice roll. In roll under the modifiers are applied to the target number. An alternative is to increase the number of dice rolled and count successes (sometimes called dice pool, but there are a few ways it can be implemented)

With roll under you have to be a bit more careful with terminology, at least in English. If for example you say you “lower the DC by N” you could lead the reader into thinking that lowering the DC is making it easier rather than harder.
Generally, most people will think positive modifiers are good and negative modifiers are bad.

As far DnD is concerned if you want to remind people of similarities your system has with DnD then roll over is the way to go. If you want to emphasize that it is a different system with different processes then roll under is the way to go. Honestly this would only apply to a high magic fantasy setting.

[gbaji]

In roll over systems the modifiers are applied to the dice roll. In roll under the modifiers are applied to the target number. An alternative is to increase the number of dice rolled and count successes (sometimes called dice pool, but there are a few ways it can be implemented)

It's semantics, but that's not really correct (or at least, not completely correct maybe?)

In roll-over, the "difficulty" is the target number. The characters skill is actually then the modifier to the roll. So the character attempts to do something. The GM decides how difficult "something" is, and assigns a target number (let's say 15). The player then rolls the dice, adding the characters skill to the die, and if the result is higher than the target number, the skill attempt succeeds. Now, yes, you can also look at that as "subtracting" the characters skill from the target number to "math ahead" the result. So if you had a skill of 10 in that sittuation, with a target number of 15, you could think "I need to roll a 5 or higher". But in either case, the character skill is always either a modifier to a pre-existing (and GM generated) target number *or* a modifier to the die roll result.

In roll-under, the characters skill value is the target number. If you have a 10 skill, and you're rolling a D20, then you succeed on a 10 or lower. Done. In additioin to this, the GM may also apply a modifier (positive or negative) based on how easy/hard the thing you are attempting is. The benefit to this system is that in the case of "standard actions", there may be no modifier at all, so there's less math involved (a target/difficulty number must always be present in roll-over systems). There is a negative here though, in that to work the "basic skill" has to be somewhat normalized to the die being used. So in a D20 system, you have to have some amount of "starting skill level" that is higher than say zero. Which puts more focus on the skills (and potentially requires more space on the character sheet). If you don't do this, then you have to have a "base number" that represents an "unskilled person trying to do this", which frankly starts to look a lot like the target number in the roll-over method (if you assume that rolling a 5 or lower even if you have no skill, succeeds, then that's no different than having a target number of 15 in the above example, righrt?).

The math can actually work exactly the same in either direction. But, as I pointed out previously, there are some psychological differences in how both players and game designers view those things, and thus how it affects a host of things in the game itself. It's also why it's not actually so easy to just swap one for the other in the same game system. There's a number of structural changes involving skills that will tend to be present depending on which resolution methodology is being used.


With roll under you have to be a bit more careful with terminology, at least in English. If for example you say you “lower the DC by N” you could lead the reader into thinking that lowering the DC is making it easier rather than harder.
Generally, most people will think positive modifiers are good and negative modifiers are bad.

Sure. But typically, in roll-under, the concept of a DC just doesn't exist. You roll against your skill. So a higher skill is better, since it's easier to roll under. Period. What would be a DC value in roll-over is a modifier applied by the GM, which is also a "plus/minus". And has the advantage that a negative value is "bad", and a positive one is "good". So if I have a skill of 15, then rolling a 15 or lower succeeds. If the GM appiles a "negative 5 situation modifier", then that's "negative" and "bad". My number is lowered from 15 to 10, and now I have to roll under a 10 to succeed (well, usually it's equal or lower, but whatever). If the GM determines that the task is easy, and applies a "+5 bonus", then now I succeed on a 20 or lower (15 skill + 5 bonus).

The difference is that you generally don't think of the modifier as affecting the die roll (like skill does in roll-over), but a modifier to the skill you are using. So you have a skill of X, and a modifier of Y, so you apply Y (adding positive, subtracting negative) and result in the number you need to roll under.

Both skill value and modifier run in the "positive and large is better" direction in roll-under, since they are actually just added together to generate the "target number" to roll against'.

In roll-over they are opposed values. Either skill is added to the die roll and match against the DC *or* skill is subracted from the DC and we roll against the result (which, yeah, results in the "we're subtracting our skill from something" condition).


As far DnD is concerned if you want to remind people of similarities your system has with DnD then roll over is the way to go. If you want to emphasize that it is a different system with different processes then roll under is the way to go. Honestly this would only apply to a high magic fantasy setting.

Eh. As I mentioned above, if you want to go roll-under, you have to make a significant additional number of changes to your games base skill system. Otherwise, you'll have everyone simply failing all but "so easy the GM is giving you a massive bonus" attempts at most things. There's a fair amount of overhead to make this work, and it does requires a game design that is much more skills focused than a game like D&D is.

The roll-under game system I use most often (RQ) basically puts every single standard skill that exists in the game on the character sheet. And every single skill starts at a base value (often based on the starting race), and includes skills category bonuses as well (which are based on character stats). So a starting character, never having used a skill (or spent any time/points during character creation on it), might still have many skills in the 25-40% range. But that's a degree of emphasis on skills that most games don't have (and front loads the work, while simplifying things at time of play). Most just have characters decide to take a skill, and only track the skills the character has taken, and the points they've put into it. Then they include in the skills section or descriptions themselves rules/charts for DCs and some rules for folks attempting to do things without having taken the skill.

That's a dramatically different approach, and changing that affects a whole heck of a lot in your game system. Eh. And I'm also descrbing a non-level and non-class based game system. Wedging skill-under into class/level based games (and especially skill selection and point allocation systems like D&D) requires doing some additional fudging of "default skill values" to make things work for "unskilled" versus "folks who took this skill and put points in it". So it's not really as simple as just swapping the numbers around.

[ngilop]

It is a roller under or a roll over sysyem?

It is counterintuitive if it is a roll under sustem for the 12 to be an automatic success.

Conversely, if it is a roll over then a 1 as an automatic success does not lend itself to easy.

Roll under: target number is x; you try to roll x or below.
EX: GURPS. HERO

Roll over: target number is x; you try to roll x or above.
EX: D&D.



Most people if you are trying to get a higher number would not 'get' why a 1 is a success.

Likewise, those same people would not 'get' why a 12 is a success if they are trying to roll a lower number.

The only time I can thi k of that that makes sense if this is some kind of price is right thing where the closer one gets to failing while still succeeding is somehow the best outcome.

[MoiMagnus]

With roll under you have to be a bit more careful with terminology, at least in English. If for example you say you “lower the DC by N” you could lead the reader into thinking that lowering the DC is making it easier rather than harder.
Generally, most people will think positive modifiers are good and negative modifiers are bad.

For that point, it's the d20 system which require to be careful (but we're just used to it):
"Given the circumstance, I've added a modifier of +2 to this test", "Wait, +2 to my roll or +2 to the DC?"

In roll under systems, there is no ambiguity. The number obtained on your dice is never modified (outside of rerolling). The only thing that ever change is the target number (usually a skill value).

Another way to say it is that roll-over systems are fundamentally "with two sides", they both try to get the biggest number and the one getting the bigger number wins (one of the side might be "the universe itself"). While on the other end, roll-under systems are focussed on the "one side doing something", and if there is any opponent or ally, it is just yet another modifier to the skill check. Opponents give negative modifers to the target number, allies give positive modifiers to the target number.

But that does mean that in roll-under systems, it's a little awkward to use the result of an action to set up the DC for an opponent party(you can't easily make a skill check for stealth and say "the result is the DC for peoples trying to see it").
[Well, you can with some more math, like saying "make a stealth check, the difference by which you succeed will be a modifier to peoples trying to detect you", but it's more awkward. I don't remember any roll-under system that care about the difference by which you succeed.]

[gbaji]

Another way to say it is that roll-over systems are fundamentally "with two sides", they both try to get the biggest number and the one getting the bigger number wins (one of the side might be "the universe itself"). While on the other end, roll-under systems are focussed on the "one side doing something", and if there is any opponent or ally, it is just yet another modifier to the skill check. Opponents give negative modifers to the target number, allies give positive modifiers to the target number.

Yup. One of the best ways to visualize roll-over systems is to think of them always as opposed rolls. There are two "sides" and each is trying to defeat the other, so bonuses to each side add to their total. Well, except that usually one "side" is a little higher by default (or has some default starting "roll") that the other side (the one played by the player) rolls a die against. In a game like Talisman, enemies have a strength/craft value, that you have to over come with your strength or craft value. But in that case, each rolls a die (another player rolls the enemy you are fighting), so the numbers start out even. In Dungeon of Dorukan, the opponents dont roll, but start with somewhat high values, the player then takes their attack or defense value, adds in any bonuses from help (loot is key), and then rolls their D12 (cause the D12 gets some love in this game!). The opponent value has to be higher if we're going to resolve this with just one roll.

And yeah, it's very similar with D&D style skill checks. The value you have to exceed is designed to be high enough that it requires rolling the D20 high enough to over come, and more difficult tasks will require having some skill bonuses involved to increase that total to be high enough to succeed. It's basically the same concept. But yeah, it can cause confusion about what you are adding to and I find thinking in terms of "two sides" can make this easier.


But that does mean that in roll-under systems, it's a little awkward to use the result of an action to set up the DC for an opponent party(you can't easily make a skill check for stealth and say "the result is the DC for peoples trying to see it").
[Well, you can with some more math, like saying "make a stealth check, the difference by which you succeed will be a modifier to peoples trying to detect you", but it's more awkward. I don't remember any roll-under system that care about the difference by which you succeed.

Yeah. Interestingly enough, while roll-over matches up well (and even feels more natural) in an opposed roll scenario, roll-under has some difficulties when doing opposed rolls (or at least, it becomes more complex with more math). It's very easy to do against static things (take lockpick skill, add bonuses for situtation/tools/whatever, subtract minuses for difficulty/conditions/whatever, roll dice). But managing two actual opponents with opposed skills requires more work. Does the person using the stealth skill roll against some value from the opponents perception skill? Or vice versa? Does just one person roll, or both? Given the issue I mentioned above of having to scale things such that "single roll" resolutions require that the opposed value (difficulty) must be set higher (with higher being a function of the die rolled) to make things work, you can't just match skill against skill and resolve in a single roll. Doing that will result in rolls being too easy in roll-over and too hard in roll-under.

I've seen a few different means to manage this in roll-under, and none of them are great. The easiest is "highest roll that succeeds wins". Which, erm... "works" but feels strange because you want to roll high, but not too high, and it actually causes some "odd" statistical outcomes. It's easy. But not terribly functionally accurate. Also, it fails completely when/if skill values exceed the range of the die being used (which is a common thing in percentle skill based games). Another method (which is the one I tend to use), is to have each roll, and whomever makes it by the most succeeds. This maintains the "low roll is good" dynamic, but requires more math (you can also think of this as "one rolls", then the amount they make (or miss!) it by is applied as a modifier which subtracts from the other's chance to succeed). It's a more mathmaticall "correct" way of doing things, but yeah, requires a lot more math. Um... But to be fair, it's not much different than that used in a full opposed check using roll-over, but players tend to find "add bonuses to sideA or sideB" to be easier to manage than "subtract from either".


So yeah. Both methods have positives and negatives (and they do really need to be "baked in" to the game system to some degree, or they wont work). I tend to like roll-under, but only because in the most common situations a GM will request rolls for the players, it's easier and requires less math. Well, and that skill values area always equalized. Having an X value in a skill always reflects the same relative capability in doing that thing, because the value of X is always compared against the same thing (the die roll). In a lot of roll-over based systems (and yeah, D&D is possibly one of the worst offenders here), the DCs for different skills under different situations can vary so wildly as to make X to Y skill level comparisons meangingless. But the negative to roll-under is that opposed checks absolutely get a bit messy.

[Anonymouswizard]

'Both parties make a skill check, highest success wins, mutual failure is a stalemate' works for opposed rolls. Default to defender wins if applicable. You can get more intuitive, or declare that a high failure beats a low failure is ties are unacceptable, but it's not hard.

You can get more complicated if you want of course, I have a d20 roll under system I've written there attacks rolls are 'you must roll over your opponent's defence but under your skill value'. Defence is calculated from relevant stats and skills in a way that intentionally keeps it relatively low.

[Vahnavoi]

Counting the difference by which a roll succeeds is fairly common in roll under systems, which is typically used to measure degree of success that can then be used to solve opposed checks. Alternatively or additionally (to break ties), higher skill wins if opposed parties all succeed their checks.

There's another way to set up opposed roll under, which I'm just going to tell you:

X - Y + 0.5D = Target number

Where X is the value for whatever is being attempted, Y is the value for whatever opposes that attempt, and 0.5D is a constant that's half of the randomizer's range. So, for example, in a percentile combat system, it looks like:

attack% - defense% + 50% = rollunder%

The point for setting it up so is that between equal opponents, the chance is fifty-fifty.

None of these are particularly hard to use.

[gbaji]

'Both parties make a skill check, highest success wins, mutual failure is a stalemate' works for opposed rolls. Default to defender wins if applicable. You can get more intuitive, or declare that a high failure beats a low failure is ties are unacceptable, but it's not hard.

That works well if we assume all skills are always in a range within the range of the die. But that assumes that no one is ever past 100% (on a percentile system). Which gives us the odd case of everyone having a chance to fail "just because" every time they try to do something, even under "normal" circumstances. It's quite common to expect that highly skilled/experienced people will have skills over 100%, specifically because this allows them to succeed with a high degree even under adverse circumstances (and well, thematically, if everyone is capped at the same value, then everyone stops at "equaly competant", which is not terribly heroic).

In those conditions, "highest success" just becomes a toss up roll. A character with 100% in an opposed roll against someone with 150% is 50/50, when realistically, that extra 50% should count for something (and a significant something).

I've seen systems attempt to use "level of sucess" then "highest when tied at success level" (I'm looking at you Mongoose), but that gets "weird" really really fast. I want to roll low to get a 1/5th or 1/20th threshold, but within that threshold, I want to roll high? Er... No.

While it's a bit of extra math (well, not really considering the above alternative), just eyeballing "how much did you make it by" works pretty easily. I have a 150%, you have a 100%. I roll a 56, you roll a 32. I made it by "a bit less than 100", you made it by "a bit less than 70". We can quickly eyeball and see who succeeded most of the time, without too much difficulty (and can be more precise with our math if it's "close"). Yes, in that example "high roll wins" would work. But it wouldn't if we reversed the rolls. If I roll a 32, I've made it by "about 120", while you rolling a 56 have made it by "les than 50". I have clearly rolled better relative to my skill than you.

What this method does is mean that the person with the lower skill has to roll that much better (ie: lower) than the other person by exactly the number of points difference between their two skills. And it has the added value that it scales to infinity. Any two opposed skills, if they are within the numerical range of the die used, will normalize the relative chance of success in an opposed roll.

I go back and forth on the pros and cons of roll-over and roll-under, but I do think that the worst are "roll one way some times, and another way other times". So if I'm doing roll-over, I want "highest roll is best" to always be the case. if I'm doing roll-under, I want "lowest roll is best". Always.

[Kish]

Ejariak finishes his betrayal by rolling damage, normally 2d6+4; he rolls 2 and 6, doing a total of 9 damage that ignores armor due to critical strike.
Huh?

How you get to 9 from those numbers? Add them together and add +4, it should be 12. Just add them together (perhaps the +4 is included in the 6 and the example is meant to indicate that both dice came up 2?), it should be 8.

How difficult is that lock supposed to be? "That's an amazing roll, with a critical success, a success, and four other numbers in the better half of all possible numbers. You fail!"

(And why are you rolling 2d6 for damage if you're rolling 1d12 to hit? It would be simpler to have the damage roll be 1d12 too.)