Dice and Probability vs Aesthetic

General questions, debates, and rants about RPGs

Moderator: Moderators

Post Reply
shinimasu
Master
Posts: 230
Joined: Thu Nov 20, 2014 7:04 am

Dice and Probability vs Aesthetic

Post by shinimasu »

So I know on an intellectual level that different dice mechanics in rpgs are about probability manipulation. d20 systems tend to have linear results and players can easily predict their chances based on what modifiers they have. I know if the monster has 20 AC and I have a +10 then I hit on a 10 or above making my chances roughly 50/50. Dice pool systems have a bell curve probability. Roll and keep has a bell curve but allows the player some control over what the arc looks like. They're also harder to mental math.

However I'm having trouble figuring out what the dice mechanics contribute to the feel of the game. I know rolling a lot of dice can feel satisfying but counting hits can feel frustrating. I know they all have different crit chances/problems which contribute to how often the player feels like they're getting kicked in the shins. But how different would shadow run feel if it was using l5r's roll and keep system instead of the hit counting it does now? How would it affect the kinds of scenes PCs find themselves in? Or what the players are likely to try and do?
User avatar
deaddmwalking
Prince
Posts: 3460
Joined: Mon May 21, 2012 11:33 am

Post by deaddmwalking »

That depends in part on how it affects the probabilities that you're concerned about. Some systems increase the rate of failure perhaps more than they'd intend if they math-hammered it more (d100 roll under, I'm looking at you).

Since 'beauty is in the eye of the beholder', you're not going to get a definitive answer as to what dice systems 'feel best', but you might get a better sense of which ones work for the system you want to create.
-This space intentionally left blank
User avatar
angelfromanotherpin
Overlord
Posts: 9745
Joined: Fri Mar 07, 2008 7:54 pm

Post by angelfromanotherpin »

Roll-and-keep weights its result towards the high end of the possible, so there's little room to roll above your expected result, and a lot of room to roll below it; also your average result is hard to calculate. That's basically the opposite of the Shadowrun mechanic. The result would probably be a lot less risk-taking, because the players are less sure just how risky something is in the first place and also rolling meaningfully higher than their baseline is less likely, and not even that much higher.

Also resolution would be slower and more annoying, because roll and keep is slow and annoying.
shinimasu
Master
Posts: 230
Joined: Thu Nov 20, 2014 7:04 am

Post by shinimasu »

deaddmwalking wrote:Since 'beauty is in the eye of the beholder', you're not going to get a definitive answer as to what dice systems 'feel best', but you might get a better sense of which ones work for the system you want to create.
Yes that's more or less what I'm interested in. I know mechanically what different dice rolling does from a probability standpoint but I don't think I've played enough games for enough time to really grok how it interacts with encouraging different sets of player actions.

Like angelfromanotherpin pointed out l5r mechanics make players more cautious because of where the weight is on the curve. And this was definitely something I noticed while playing roll/keep games in general. Action was considered and weighted more carefully. However the dice themselves are interacting with things like TN and HP. In games where players had more HP or TNs were on average lower (though still offering maybe a 45% chance of success as most) there was more risk taking.

So if I wanted, say, a game focused on riding beasts around a large open world with experience gain focused on location discovery/artifact collection as opposed to direct combat it's hard to know instinctively what set of dice rules suits that best. Or a game where all combat is social combat. Sure I could take a d20 ruleset and replace "fireball" with "scathing insult" but it wouldn't feel right. Charisma represented by a steadily increasing modifier has never sat well with me, or a lot of other people. I think socially based systems in general work better with dice pools of various types but I couldn't tell you why that feels "better" than +5 to talk pretty.
User avatar
deaddmwalking
Prince
Posts: 3460
Joined: Mon May 21, 2012 11:33 am

Post by deaddmwalking »

I think it feels better because you're more likely to get an average result. D20 is swingy - you're just as likely to have your 'best day ever' as your 'worst day ever'. In our experience, we usually do things the way we usually do things, and only rarely do we do exceptionally well or exceptionally poorly. Adding more dice makes extreme results correspondingly more rare.

At least, that's my guess. If you generally want people to be competent in their domain, dice pools are pretty effective. But usually calculating the odds are beyond your average user - especially if either the number of dice or the target number changes (or both).
-This space intentionally left blank
User avatar
erik
King
Posts: 5861
Joined: Fri Mar 07, 2008 7:54 pm

Post by erik »

A minimum requirement I insist upon for a RNG is that being better (having a better die) should translate into being better (succeeding more often).

This design failure usually only shows up on some exploding dice mechanics, since they have discrete points where having worse dice is more likely to be better.

Mechanics that can drive player decisions often come up if there are extreme consequences, like critical successes and failures. I've noticed players pick up on these and adapt their behavior quickly. If you get extraordinarily punished for bad failure then it will make players more cautious. If you get extraordinarily rewarded for high success then players are more likely to swing for the fences or make as many attempts as they can.
User avatar
momothefiddler
Knight-Baron
Posts: 883
Joined: Sat Feb 22, 2014 10:55 am
Location: United States

Post by momothefiddler »

I can't tell if you're asking about the effect of the literal dice chosen or the probabilities of success involved. Like... for the purposes of this thread, do you care about the difference between a 4dFudge and 4d3-8? Are you talking about how being more likely to get an average result affects players' risk-aversion, or how having different shapes of dice for different things makes it easier to sort your calculations, mentally?
User avatar
OgreBattle
King
Posts: 6820
Joined: Sat Sep 03, 2011 9:33 am

Post by OgreBattle »

However I'm having trouble figuring out what the dice mechanics contribute to the feel of the game.
When certain things share mechanics they'll feel more similar, when they have differing mechanics they'll feel different. A core complaint against 4e was fighter sword swings and wizard spells using the same dice rolling (and same encounter/daily format).

If you want certain things to feel different (like shooting an arrow vs hurling a fireball) having different dice rolls can help.

Some folks also get spooked by an 'automatic success' at something. Like Pathfinder had a feat intensive path for monks to nullify one melee attack a round, it got banned because the idea of a dirty martial bypassing dice rolls like a magic user terrified pathfinder society. Now if instead it was a die roll (even one with a low chance of failure) it would feel more like a martial maneuver (where the user rolls d20's) and not magic.

-----

"Dice feeling" was a core part of a "How would I do D&D4.5 and fix everything gloriously" idea I had. magic and swordplay would roll a d20 to hit but the decision making step would differ like...

martial maneuvers- roll d20 to hit, then you pick what martial power to apply (trip/bullrush/step away without OA and so on). The decision making is after the die roll so it feels more 'spontaneous'.

magic powers- choose what spell to use, then roll a d20 to see if it hits. Decision making is before the die roll so you must have a spell in mind before rolling.
Last edited by OgreBattle on Thu Dec 22, 2016 9:26 am, edited 3 times in total.
Blade
Knight-Baron
Posts: 663
Joined: Wed Sep 14, 2011 2:42 pm
Location: France

Post by Blade »

The psychological aspects are indeed important to consider. I think there are 3 aspects of this: the feel and the maths, the feel and the flow and the feel and the atmosphere.

Feel and Math
Since humans are pretty bad at probabilities, one of the big problem many systems have is that the "feel" is often different from the math. This is what leads people to support a system that's mathematically bad because it "feels" right.

For instance, consider the Savage World mechanism, where you use different exploding dice (from D4 to D20). Having a D4 somewhere means you're supposed to be bad at it, worse than someone who's got a D6. Instinctively it kinda feels right. But when you do the math, you realize that you're more likely to get results above 6 with an exploding D4 than with a D6.
But these "unexpected" successes from D4 characters give the impression to players that their character beat this odds, so they like the feel of that.

That's why a good system should have the math meet the expectations. Systems with variable TN and dice pools are great to allow for very fine probabilities adjustments and different effects of different variables, but they're not intuitive at all. Most players and GM won't be able to tell what difference adding or removing a dice will make compared to increasing or lowering the TN.

Feel and Flow

Rolling many dice feels cool, but counting many dice is slow. Some mechanism have added subtleties like optional re-roll that make people feel that they're making tactical choices (even if most of the time it's just a matter of probabilities), they'll also be slow but players will notice that less. Revealing a card doesn't take longer than throwing a die, but feels faster for many players.

Feel and Atmosphere

Using a deck of cards for a western or mahjong tiles for a game set in China can contribute to the atmosphere.
Rolling x die against a TN is the same as rolling x times one die against the same TN, but having the rolling be sequential makes for increased tension.

Having an obscure mechanics for a part of the game is most of the time a pain in the ass, but it also ensure that only players who are really into it will do it, and what they'll do will look as obscure to the players as what the characters are doing.

There is probably more to say on the topic, but that's what comes to my mind for now.
shinimasu
Master
Posts: 230
Joined: Thu Nov 20, 2014 7:04 am

Post by shinimasu »

momothefiddler wrote:I can't tell if you're asking about the effect of the literal dice chosen or the probabilities of success involved. Like... for the purposes of this thread, do you care about the difference between a 4dFudge and 4d3-8? Are you talking about how being more likely to get an average result affects players' risk-aversion, or how having different shapes of dice for different things makes it easier to sort your calculations, mentally?
The basic tldr I guess is: How do different dice systems incentivise different playstyles. For example I have a set of rules and conflict resolution mechanics based around exploration. I want a dice system that works with those mechanics to encourage players to make liberal use of exploration skills.

To take an example: l5r uses a roll and keep system. This dice system is weighted towards the high end of the bell curve, and is hard for players to mental math. l5r also has a "raises" mechanic that offers the players bigger rewards for voluntarily raising the TN of a challenge. The idea here is essentially gambling failure for glory. It has the potential to be kind of fun, but since the roll/keep system A) makes it hard to calculate chance of failure in your head and B) makes it unlikely you will ever roll above certain thresholds even if you do know your chances then the raise system hardly ever sees any use. Raises would have worked much better in a system with greater probability transparency, where the dice are a bit swingier.
pragma
Knight-Baron
Posts: 822
Joined: Mon May 05, 2014 8:39 am

Post by pragma »

I think deaddm hit the nail on the head: many die rolling systems have very high variabliity compared to what you'd expect from genre norms or everyday life. D20 is especially egregious, and I think it's pretty bad at modeling skill systems as a result.

Shadowrun dice pools are nice because the most common result is pool/3 hits, which is easy to understand and which players can grow to rely on. I find players are much less hesitant to use skills if they have reasonable assurance of a middling result.

However, Shadowrun has this weird side effect that your results get more variable as you get a bigger pool. Even though you're more likely to succeed on a 3 hit task with a 15 die pool than an 8 die pool, you'll see bigger swings in your number of hits, which my players report feels bad.

I like the idea of fudge dice as a really consistent randomizer which doesn't penalize high skill characters, but I don't know of any crunchy games which use them.
Post Reply