What is your "Fight Me" thread?

[Airk]

TOn the flip side, I've also become skeptical of variable die pools (either added up, or counting "successes") due to the huge range of possible results and the difficulty balancing no skill, low skill, and high skill inside the same range of "difficulty".

The range of results in a success based dice pool is actually usually smaller than the range of results on 3d6 (Specifically, you would need 19+ dice in your dice pool before you can generate more than 18 meaningfully distinct results), so I'm not really sure where you are coming from here.

[Max_Killjoy]

The range of results in a success based dice pool is actually usually smaller than the range of results on 3d6 (Specifically, you would need 19+ dice in your dice pool before you can generate more than 18 meaningfully distinct results), so I'm not really sure where you are coming from here.

It's not about the number of results, it's about how variable the possible results can be.

Using oWoD as an example, depending on the applicable Attribute and "skill", you could have a pool anywhere from 1d10 to 10d10 in most instances, against a standard target number of 6+, meaning that one character has a 50/50 chance of getting a single success, while the other character could roll anywhere from 0 to 10 successes.

On 3d6, every roll has the same range of 16 possible results, but centered around the middle of the curve at 10.5 (technically rolling 11 or more is the 50/50 point.)

[Airk]

It's not about the number of results, it's about how variable the possible results can be.

Using oWoD as an example, depending on the applicable Attribute and "skill", you could have a pool anywhere from 1d10 to 10d10 in most instances, against a standard target number of 6+, meaning that one character has a 50/50 chance of getting a single success, while the other character could roll anywhere from 0 to 10 successes.

On 3d6, every roll has the same range of 18 possible results, but centered around the middle of the curve at 10.5 (technically rolling 11 or more is the 50/50 point.)

You're not comparing apples to oranges here, because no system uses 3d6 without modifiers for all rolls. You have to compare your 3d6 system plus any modifiers that might come into play. So how do you handle 3d6+4 vs 3d6-1? 3d6+6 vs 3d6-3? You will discover a massive shift in potential outcomes at that point. Some characters will be literally unable to do some things, and other characters will have extremely high success rates on some things.

Dice pool systems have all the same features as a fixed set - if you have 5 dice, and you need 6 successes, that is impossible for you, in much the same way that if you have 3d6+1 and you need a 20, it's not possible for you. Dice pools systems present, basically, a floating curve - it always centers around exactly half as many successes as you have dice, and as you get more dice, the extremes of performance increase, but get less likely. The only real difference between them and a fixed pool with modifiers is that it is impossible to "skill yourself" out of a chance of failure - theoretically, even rolling 30 dice you could get zero successes even though it's vanishingly unlikely, whereas you literally cannot roll a 3 on 3d6+1. Whether that's a feature or not is up to you.

I don't think it presents any meaningful "balancing" difficulties.

[Knaight]

The only real difference between them and a fixed pool with modifiers is that it is impossible to "skill yourself" out of a chance of failure - theoretically, even rolling 30 dice you could get zero successes even though it's vanishingly unlikely, whereas you literally cannot roll a 3 on 3d6+1. Whether that's a feature or not is up to you.

This feature is a large part of why I've been experimenting with the reversed dice pool - it gives a very clean way of skilling yourself out of a chance of failure, while also making it theoretically possible to succeed at anything. Exploding dice do the same thing, but the probability curve on exploding dice is really wonky.

[Max_Killjoy]

You're not comparing apples to oranges here, because no system uses 3d6 without modifiers for all rolls. You have to compare your 3d6 system plus any modifiers that might come into play. So how do you handle 3d6+4 vs 3d6-1? 3d6+6 vs 3d6-3? You will discover a massive shift in potential outcomes at that point. Some characters will be literally unable to do some things, and other characters will have extremely high success rates on some things.

Dice pool systems have all the same features as a fixed set - if you have 5 dice, and you need 6 successes, that is impossible for you, in much the same way that if you have 3d6+1 and you need a 20, it's not possible for you. Dice pools systems present, basically, a floating curve - it always centers around exactly half as many successes as you have dice, and as you get more dice, the extremes of performance increase, but get less likely. The only real difference between them and a fixed pool with modifiers is that it is impossible to "skill yourself" out of a chance of failure - theoretically, even rolling 30 dice you could get zero successes even though it's vanishingly unlikely, whereas you literally cannot roll a 3 on 3d6+1. Whether that's a feature or not is up to you.

I don't think it presents any meaningful "balancing" difficulties.

What is your opinion of something like HERO's figured "target numbers" that center everything around the peak of the 3d6 curve (at the 50/50 point of 11+)?

[Airk]

This feature is a large part of why I've been experimenting with the reversed dice pool - it gives a very clean way of skilling yourself out of a chance of failure, while also making it theoretically possible to succeed at anything. Exploding dice do the same thing, but the probability curve on exploding dice is really wonky.

I don't really understand how exploding dice help you skill out of failure? Don't you still need to roll 6's or whatever for it to apply?

[BRC]

I don't really understand how exploding dice help you skill out of failure? Don't you still need to roll 6's or whatever for it to apply?

Exploding Dice make it theoretically possible to succeed at anything. Even somebody needing to get 24 on 1d4 can theoretically pull it off.

[Guizonde]

This feature is a large part of why I've been experimenting with the reversed dice pool - it gives a very clean way of skilling yourself out of a chance of failure, while also making it theoretically possible to succeed at anything. Exploding dice do the same thing, but the probability curve on exploding dice is really wonky.

are you talking about crit systems like dark heresy? based on luck, how unbalanced a d10 is, and general shenanigans, crits get stupid fast. i mean, it's logical for a mugger to land a lucky hit with a rusty pocket knife on a knight in full plate (ie, under the gorget and through the carotid), that'll kill anyone dead. iirc, according to whfrp2e, that's bf-2 +d10. for a normal human, that equates to d10+1. you roll a 10, confirm, reroll, land another 10, reroll without needing to confirm, land a 9, add 1, that's 30 wounds dealt. a bretonnian knight should be around the 4 endurance mark, 5 armor, 12 wounds, so 30-9 is 21, minus 12, the dude is firmly in the negatives. you roll on the crit table (in this case i'd wager head wound?) and being at -9 wounds will most certainly kill you dead in a spectacular fashion, akin to getting your head lopped off by a butterfly knife. is it likely? no! is it realistic to die from a knife wound through plate armor? it's actually one of the surest methods. for your head to fly off? mwahaha, no.

my personnal record pulling exploding dice hijinks is 291 wounds of armor piercing damage (i think, the dm lost count, but it's definitely over 285 and under 296). yes, i was throwing multiple d10's, but when you're regularly critting on 3 dice and adding the total... it may be a homebrew system that's voluntarily unbalanced to increase lethality, but keep in mind that the all-time high score for wounds on a pc was 17, and my weapon was hitting at the d10's value in wounds... it made the dm shed a few tears, ask for a glass of something medicinal, recess the table for 10 minutes, and banned me from ever pulling another crit that took a calculator to solve.

best. system. ever. or worst, depending on your stance on overkill.

[daniel_ream]

Answering the OP:

Game balance is a canard, and point costs don't mean anything.

[Max_Killjoy]

Really not a fan of exploding dice, personally.

[Guizonde]

Really not a fan of exploding dice, personally.

to each their own. only downside i have about it is that i have way too much luck with the things. i lose count. then again, when you're talking about the difference between an anti-tank mine and several hundred pounds of dynamite, the point is moot anyway. for me, it illustrates perfectly what a critical hit should be in an action movie fashion. a dope-slap that rips off a head is so over the top that it both fits my game world and pleases my inner looney tunes fan.

for gritty realism, though, i admit it's really not a good mechanic.

[Knaight]

I don't really understand how exploding dice help you skill out of failure? Don't you still need to roll 6's or whatever for it to apply?

They make it theoretically possible to succeed at anything, although exploding dice systems fit well enough with roll and add that skilling out of failure is still a feature in a lot of them in a way that it isn't in most conventional dice pool systems. It's just that the way they make possible success happen involves a downright bizarre set of probability curves.

[kyoryu]

Answering the OP:

Game balance is a canard, and point costs don't mean anything.

You're arguing that those statements are true?

Because I won't fight you on that at all.

[Quertus]

It's been a pretty solid system, with one oddity in that there's no intuitive way to pull of opposed rolls. There are a few methods of varying degrees of unintuitiveness, and I quite like the one I ended up using for the specific game I was developing the reverse dice pool for, but rolls take a while to resolve and it wouldn't play well with a lot of players (the specific game is tailored towards two players and a GM).

Ever-scaling difficulty, with the option to back down? Sounds like a fun minigame to me.

[CharonsHelper]

On 3d6, every roll has the same range of 18 possible results, but centered around the middle of the curve at 10.5 (technically rolling 11 or more is the 50/50 point.)

To be nitpicky - there are only 16 possible results when rolling 3d6.

Really though - I'm not sure what you're looking for. You don't like d20 because it's not a bell curve, and you don't like 3d6 because it's not random enough? Maybe I'm missing something.

[Max_Killjoy]

To be nitpicky - there are only 16 possible results when rolling 3d6.

Really though - I'm not sure what you're looking for. You don't like d20 because it's not a bell curve, and you don't like 3d6 because it's not random enough? Maybe I'm missing something.

Typo on the number range.

I wasn't saying 3d6 was "not random enough", I was comparing it to variable dice pools.

[PhoenixPhyre]

I'm afraid that I don't really understand the strong opinions about the different dice mechanics' probability curves. RPG sessions (in my experience) are still firmly in the tyranny of small numbers regime, not the nice bell-curve large numbers regime. I just don't roll enough during a session to be able to notice a difference between the mechanics, and I forget (along with hind-sight and other cognitive biases) about past sessions.

IMO, any mechanic can work as long as you're not rolling for stupid things. If something's impossible, don't roll. If something shouldn't fail (or failure would just mean rolling again), don't roll. Only roll when it's meaningful. That makes the numbers even smaller and removes the binaries as a possibility.

For example, I DM a 5e D&D group. The rogue has an ability that floors his roll at 10 on skills he's proficient in--anything below that counts as a 10. For stealth, for example, he can't roll below a 23. That's high enough that he can't be found unless someone's actively looking for him (except for a very few creatures). Doesn't help his party roll well on stealth though, so they still have to roll. He just generally gets a turn being hidden if he has a conducive situation (which he uses to great effect).

I figure that for adventurers doing adventuring things (most of 5e), they can do the simple stuff. All of them. No rolls for getting out of bed, no rolls to remember to put on pants. Even climbing (unless the surface is unusual) doesn't take a roll--it's just slower. Only the unusual things or things that ordinary people would struggle at get rolls at all.

All of this together makes it so you're not rolling enough to get a real distribution off the dice.

I can understand liking different mechanics for other reasons, but not really the distributional mechanics. But that's just me.

[Max_Killjoy]

I'm afraid that I don't really understand the strong opinions about the different dice mechanics' probability curves. RPG sessions (in my experience) are still firmly in the tyranny of small numbers regime, not the nice bell-curve large numbers regime. I just don't roll enough during a session to be able to notice a difference between the mechanics, and I forget (along with hind-sight and other cognitive biases) about past sessions.

IMO, any mechanic can work as long as you're not rolling for stupid things. If something's impossible, don't roll. If something shouldn't fail (or failure would just mean rolling again), don't roll. Only roll when it's meaningful. That makes the numbers even smaller and removes the binaries as a possibility.

For example, I DM a 5e D&D group. The rogue has an ability that floors his roll at 10 on skills he's proficient in--anything below that counts as a 10. For stealth, for example, he can't roll below a 23. That's high enough that he can't be found unless someone's actively looking for him (except for a very few creatures). Doesn't help his party roll well on stealth though, so they still have to roll. He just generally gets a turn being hidden if he has a conducive situation (which he uses to great effect).

I figure that for adventurers doing adventuring things (most of 5e), they can do the simple stuff. All of them. No rolls for getting out of bed, no rolls to remember to put on pants. Even climbing (unless the surface is unusual) doesn't take a roll--it's just slower. Only the unusual things or things that ordinary people would struggle at get rolls at all.

All of this together makes it so you're not rolling enough to get a real distribution off the dice.

I can understand liking different mechanics for other reasons, but not really the distributional mechanics. But that's just me.

Is there some hyperbole involved, or are there really groups out there where it's normal to roll for putting on pants or getting out of bed?

Part of the reason for that concern about the distribution (curve, range, etc) is that it affects those other reasons to like or dislike the dice mechanics.

If you want a game where skill matters most, that's a different dice setup than if you want a game with random wacky luck dominating. If you want a game where "median" results are more likely, that's a different dice setup than if you want a game where all results are pretty much equally as likely or the dice are really swingy.

[PhoenixPhyre]

Is there some hyperbole involved, or are there really groups out there where it's normal to roll for putting on pants or getting out of bed?

Part of the reason for that concern about the distribution (curve, range, etc) is that it affects those other reasons to like or dislike the dice mechanics.

If you want a game where skill matters most, that's a different dice setup than if you want a game with random wacky luck dominating. If you want a game where "median" results are more likely, that's a different dice setup than if you want a game where all results are pretty much equally as likely or the dice are really swingy.

Those examples are hyperbole, but I've seen groups roll for all sorts of things (like remembering who the guy who had sent you on the mission was).

But as for the rest of it, with only a few rolls, you're not going to see the statistics. That takes ~30 sample points. In my experience, no one but the DM rolls that many in several sessions combined. Especially if you save rolls for the important things where everyone has some relevant skill. Sure, if you want systems that are, in essence, binary (or where the dice roll only determines how well you succeed, not whether you succeed) you'll need to either stack modifiers on a flat system (D&D 3.X's approach, where skill checks were basically pass-fail--if you weren't guaranteed success you will probably fail) or modify the dice mechanic. It just doesn't matter that much to me either way. If I'm gonna roll dice, I'd rather that they have a relatively large effect on the outcome (otherwise, why roll at all?). Combine this with discretion on auto-success/auto-fail, this makes skill matter (a truly skilled person isn't going to have to roll for much in his specialty, while someone with strong penalties probably can't even succeed) while also allowing the dice to matter.

The big issue I see with 5e's system is people using the methods of 3.X or earlier and simply using the new numbers while rolling for the same sorts of things. This gives swingy events with little skill, it's true. Most of those rolls shouldn't have happened at all.

Again, this is purely academic for me. I just don't have a strong emotional attachment or antipathy to any system.

[daniel_ream]

You're arguing that those statements are true?

Per the OP, these are reasoned statements I've made that provoke extreme emotional responses, and that I don't seem to be able to walk away from.

I can start the flamewar right now by providing specific criticisms of Champions and/or Mutants & Masterminds.

Wild Talents took all the fun out of it by providing a sidebar right up front that says "point costs don't matter and the game isn't balanced", right before launching into a hundred pages of complicated power build rules.

[Max_Killjoy]

You're arguing that those statements are true?

Because I won't fight you on that at all.

So you'd say that there's no way to at least approximate a relationship between mechanical impact of character elements, and their cost in some character-build currency (points, XP, whatever)? And that furthermore there's no way to get characters built on the same amount of "currency" to be at least close enough in "impact" that they the players will be neither over- nor underwhelmed relative to each other in terms of their character's ability to affect the game via the mechanics?

[daniel_ream]

So you'd say that [...]

I do, and have.

[Vitruviansquid]

Simple: The random factor of the rule is too large in magnitude compared to the differences in character capability, so your chance of success is far more driven by random chance than it is by the actual capabilities of your character. Bumbletorp McTwoLeftFeet with a -1 dex mod and no stealth skill can outsneak Steathly vonSneakypants and his +5 pretty easily and these are pretty close to the natural extremes of competence in 5e.

The alternative, of course, is the "I have so many modifiers that it doesn't really matter what I roll" side of d20 (Mostly Pathfinder, though it happens "naturally" a high levels of play in many editions, though those often end up being in category A, above, because the target number is so high that it's functionally the same as +5 vs -1.) which has its own problems.

And maybe it's "easy" for someone to compute the "chance" of doing something, but it also presents plenty of chances for someone to get frustrated when their supposedly good at something character fails at it repeatedly.

How come it's a bad thing that your expert can fail where your novice succeeds? Maybe vonSneakypants is doing everything right, but at the instant he's trying to sneak past, the guard turns around for a completely unrelated reason. The point of rolling in the first place is to simulate the uncertainties with certain tasks. And if there is no uncertainty, you shouldn't be rolling in the first place - after all, you don't have games where people need to roll to put on pants.

If you say there's an alternative of jamming a character so full of modifiers the roll doesn't matter... then why not just have them auto-succeed?

I asked about the problem with a flat distribution, but the only thing you've shown me here is the problem with rolling at all. If you have a bell curve distribution, which is a fairly popular alternative to flat, your expert can still fail at doing something you're good at. You can still attempt to fight this by jamming too many modifiers onto that roll, which can still have its own problems.

[Max_Killjoy]

How come it's a bad thing that your expert can fail where your novice succeeds? Maybe vonSneakypants is doing everything right, but at the instant he's trying to sneak past, the guard turns around for a completely unrelated reason. The point of rolling in the first place is to simulate the uncertainties with certain tasks. And if there is no uncertainty, you shouldn't be rolling in the first place - after all, you don't have games where people need to roll to put on pants.

I'm not even sure "the guard just happens to turn around" should be part of the consideration on the input or output side of the roll made by the character trying to sneak past. The guard's random act isn't part of the character's capability that's being tested, it's an external factor.

[RazorChain]

It's not about the number of results, it's about how variable the possible results can be.

Using oWoD as an example, depending on the applicable Attribute and "skill", you could have a pool anywhere from 1d10 to 10d10 in most instances, against a standard target number of 6+, meaning that one character has a 50/50 chance of getting a single success, while the other character could roll anywhere from 0 to 10 successes.

On 3d6, every roll has the same range of 16 possible results, but centered around the middle of the curve at 10.5 (technically rolling 11 or more is the 50/50 point.)

This is one of the reasons I like Gurps. The bell curve is "centered" around 9 to 12 and you can implement degrees of success

[Vitruviansquid]

I'm not even sure "the guard just happens to turn around" should be part of the consideration on the input or output side of the roll made by the character trying to sneak past. The guard's random act isn't part of the character's capability that's being tested, it's an external factor.

Where does it say that the dice simulate the character's capability?

That'd make no sense as the character's capability is constant. The skill modifier you have already simulates the character's capability.

The dice can only simulate the external factors in situations.

[Max_Killjoy]

Where does it say that the dice simulate the character's capability?

That'd make no sense as the character's capability is constant. The skill modifier you have already simulates the character's capability.

The dice can only simulate the external factors in situations.

Consider that even a basketball player taking free throws, an uncontested and very repetitive / similar action, will not make 100% of the attempts over time. The internal variables are still enough to make the outcome uncertain, with a percentage made over time reflecting their relative skill at the task.

[Max_Killjoy]

This is one of the reasons I like Gurps. The bell curve is "centered" around 9 to 11 and you can implement degrees of success

Likewise, I consider that one of the strengths of HERO.

[kyoryu]

So you'd say that there's no way to at least approximate a relationship between mechanical impact of character elements, and their cost in some character-build currency (points, XP, whatever)? And that furthermore there's no way to get characters built on the same amount of "currency" to be at least close enough in "impact" that they the players will be neither over- nor underwhelmed relative to each other in terms of their character's ability to affect the game via the mechanics?

Yup!

In practice, I have *never* seen it actually work.

The best balance I've seen is either pure class-based systems, where there's no choices beyond "class", and Fate, where you don't have points, but just get to slot limited things in slots.

[Vitruviansquid]

Consider that even a basketball player taking free throws, an uncontested and very repetitive / similar action, will not make 100% of the attempts over time. The internal variables are still enough to make the outcome uncertain, with a percentage made over time reflecting their relative skill at the task.

Alright, I guess then that the dice simulate a combination of internal and external factors.