PDA

View Full Version : Low Variance RPGs



ZenBear
2015-08-30, 07:13 PM
I've been contemplating the effect lowering the size of the die in an RPG like D&D. My experience outside of D&D is very limited, only having played Call of Cthulhu, Paranoia, Exalted and Gamma World. I have been watching Titan's Grave by Will Wheaton and the episode of Tabletop where he plays Dragon Age, and the use of d6s was a change of pace. Still, 3d6 is not much different from 1d20 and still allows for someone with STR as a dump stat beating the STR based character in an arm wrestling match, which I hate.

I'm considering something along the lines of D&D 5E where there's a cap of +5 to individual abilities scores but rolling only 1d6 for tests. Several tests would be straight numbers with no dice (ie arm wrestling) where variance doesn't occur unless the scores are tied, in which case you do roll. Are there RPGs out there that do something like this I could look at?

I'm also interested in alternate ability scores. Some games I know have a 3 score system like the Fable Series' Strength/Skill/Will, and I like that to a point. Perhaps even as narrow as 2; Body/Mind. What RPGs use these systems? Are they good, bad, apples to oranges different?

Thinker
2015-08-30, 07:49 PM
I've been contemplating the effect lowering the size of the die in an RPG like D&D. My experience outside of D&D is very limited, only having played Call of Cthulhu, Paranoia, Exalted and Gamma World. I have been watching Titan's Grave by Will Wheaton and the episode of Tabletop where he plays Dragon Age, and the use of d6s was a change of pace. Still, 3d6 is not much different from 1d20 and still allows for someone with STR as a dump stat beating the STR based character in an arm wrestling match, which I hate.

I'm considering something along the lines of D&D 5E where there's a cap of +5 to individual abilities scores but rolling only 1d6 for tests. Several tests would be straight numbers with no dice (ie arm wrestling) where variance doesn't occur unless the scores are tied, in which case you do roll. Are there RPGs out there that do something like this I could look at?

I'm also interested in alternate ability scores. Some games I know have a 3 score system like the Fable Series' Strength/Skill/Will, and I like that to a point. Perhaps even as narrow as 2; Body/Mind. What RPGs use these systems? Are they good, bad, apples to oranges different?

3d6 is actually very different from 1d20. You have a similar range of possibilities, but the actual die rolls will gravitate much more toward the middle of the range. The modifier is much more important when rolling 3d6 and will result in much more frequent victories for the character with the higher score. If it still bothers you, you could also just say that the higher score auto-wins if it is more than 5 higher than the lower score.

Rainbownaga
2015-08-30, 08:33 PM
Wasn't arm wrestling given as an example of strait stat comparison in the 3.5 dmg? Or was it an older version of d&d.

The variance is only really supposed to come up in things where they are complicated enough that luck can realistically play a major part, such as in a fight, picking an unfammiliar lock, noticing something out of the ordinary, etc. (which is why jump tests are pretty silly).

Using a d6 for fights or complicated skill checks would be annoying as the degree between automatic success and automatic failure becomes way too narrow.

Tobtor
2015-09-01, 11:19 AM
Using a d6 for fights or complicated skill checks would be annoying as the degree between automatic success and automatic failure becomes way too narrow.

Why is that annoying?

If its complicated/difficult, the chance of auto success should be small... similarly if you are really good at something a 5% fail chance at a rutine manoeuvres is a bit silly, but an unskilled person might fail so you cannot forego the check altogether. A 3d6 reduces the auto fail/succes to about 0,5% (if 3 or 18 is chosen), much more reasonable in my opinion.

A 3d6 also puts more important into circumstance modifieres, as changes around average means a lot, while in a d20 a +/-1 is always a 5% increase/decrease. In a 3d6 environment a "gifted" person (strong, clever, dexterous, wise etc) person still have a decent chance at doing something difficult, while a person around average can fulfil simple task but not difficult ones. Seems more realistic and fair, that someone who has focussed on an area of expertise, gets rewarded over someone who has not.

Lvl 2 Expert
2015-09-01, 11:37 AM
I've done some dice calculations recently, and I think I've fallen in love with a system I call n-1 D3. You throw n D3's (n starts at 2) and detract the lowest throw. Getting more skilled in something gives you another D3, rather than a plain modifier.

One cool thing about it is that you generally perform closer to the best of your ability than to your worst, especially in untrained checks. If you try to lift something heavy but within the limits of what you can lift, you'll probably succeed, only so many ways to screw that up. As you get better at a task, the amount of room you have to screw up in grows, yet at the same time you'll still be performing better on your worst day than an untrained person on his best. The probability curves end up looking like this (bottom lines are the chance of throwing that number, top lines of throwing that number or higher).

http://s3.postimg.org/pt767gdlv/probability_curves_n_1_D3small.jpg
(click for bigger) (http://s14.postimg.org/d6u95cyqp/probability_curves_n_1_D3.jpg)

It's just a shame that the whole thing is too complicated and slow to ever actually use in any form of real play. Why add up 6 D3's all the time when throwing a D20 works too, even if the probability curve is less elegant?

kyoryu
2015-09-01, 11:46 AM
Amber Diceless and its successors remove randomizers *entirely*

3d6 has much lower variance. Also, most 3d6-based systems (that I know of) handle skill increases in smaller numbers than D&D does - the two factors you're dealing with are variance in the roll, and variance in the skills themselves.

Fudge/Fate dice (4d3-4) are also pretty consistent. Fate puts skills on a 0-5 range (typically), and I think most Fudge variants do the same as well.

Knaight
2015-09-01, 12:17 PM
Amber Diceless and its successors remove randomizers *entirely*

3d6 has much lower variance. Also, most 3d6-based systems (that I know of) handle skill increases in smaller numbers than D&D does - the two factors you're dealing with are variance in the roll, and variance in the skills themselves.

Fudge/Fate dice (4d3-4) are also pretty consistent. Fate puts skills on a 0-5 range (typically), and I think most Fudge variants do the same as well.

Fudge generally uses -2 to +3 for starting skills, though the ones you actually put points in are from -1 to +3. There's then room for +4 later. As for the variability, while small it's still pretty noticeable, unless you use a dF.1, which is a d6 where a 1 is a -1, a 6 is a +1, and 2-5 are all blank.

ZenBear
2015-09-01, 01:05 PM
Why is that annoying?

If its complicated/difficult, the chance of auto success should be small... similarly if you are really good at something a 5% fail chance at a rutine manoeuvres is a bit silly, but an unskilled person might fail so you cannot forego the check altogether. A 3d6 reduces the auto fail/succes to about 0,5% (if 3 or 18 is chosen), much more reasonable in my opinion.

A 3d6 also puts more important into circumstance modifieres, as changes around average means a lot, while in a d20 a +/-1 is always a 5% increase/decrease. In a 3d6 environment a "gifted" person (strong, clever, dexterous, wise etc) person still have a decent chance at doing something difficult, while a person around average can fulfil simple task but not difficult ones. Seems more realistic and fair, that someone who has focussed on an area of expertise, gets rewarded over someone who has not.

If I decided to replace all d20 rolls in D&D 5E with 3d6 would that screw up the game do you think?

Takewo
2015-09-01, 01:27 PM
If I decided to replace all d20 rolls in D&D 5E with 3d6 would that screw up the game do you think?

Nope, it would just mean more average results.

Bobbybobby99
2015-09-01, 02:57 PM
Unearthed arcana from DnD explicitly has rolling 3d6 instead of 1d20 as an option, so this has been a thought for a while. Apparently it's fairly simple, and you take 16 instead of 20, and can take 18 by taking 100 times the regular time.

The reason that it doesn't have common use is probably because it can easily cause ridiculous amounts of extra rolling, counting, having to pick three times as many dice off of tables, etc. It just isn't as practical.

If it was an explicit option in 3.5, then it should probably work in 5. They aren't that different.

Madeiner
2015-09-01, 05:16 PM
The reason that it doesn't have common use is probably because it can easily cause ridiculous amounts of extra rolling, counting, having to pick three times as many dice off of tables, etc. It just isn't as practical.

If it was an explicit option in 3.5, then it should probably work in 5. They aren't that different.


3d6 probability is so much nicer than 1d20 that i was tempted to try and make a custom dice with 3d6 probabilities or close to it.

I guess it would take a 30ish sided dice and repeating some numbers.
Even a d20 with skipping numbers would be better than 1d20.

Knaight
2015-09-01, 05:19 PM
One other option for lower variance is dice pools. They generally follow binomial probability distributions, which are possibly tweaked to some degree or another by particular mechanics. Nonetheless, the variance is drastically reduced compared to things like d20+x.

Mendicant
2015-09-01, 05:48 PM
Nope, it would just mean more average results.

It would change all kinds of things. 3d6 means that even very small differences in modifiers have very big effects on what is possible. For instance, imagine you're dealing with an enemy with an AC of 20, and you have two characters: one with an attack bonus of 4 and the other with an attack bonus of 7. The character with a +4 is locked out, basically, with a roughly 6% chance of hitting. The +7 has a 26% chance, which is a pretty huge swing. Add +1 to that and the hit chance goes up by more than 10%.

DnD is not built with such huge swings in mind, and it'll get weirder and weirder the higher in level you go.

Mr Beer
2015-09-01, 05:54 PM
3d6 is very different to 1d20 and provides much more plausible outcomes IMO.

That's one of the reasons why I use GURPS instead of D&D as a system.

TheOOB
2015-09-01, 07:28 PM
As people have mentioned, rolling a smaller dice doesn't reduce randomness, it just makes each point of bonus a lot bigger.

There are tons of dice pool systems out there, that is systems where you roll multiple dice based on your skills, to the point where they may outnumber systems that just roll a single die(so long as you count all d20 games as one system). Shadowrun, World of Darkness, 7th Sea, Exalted, ect. Dice pool systems have much less randomness to them, which is a huge boon when rolling for non combat encounters, and tends to lead to quicker combat encounters.

Just remember that reducing the randomness of a system will make the system favor the players more(an conversely, increasing randomness favors the NPC's). Just switching from d20 to 3d6 and changing nothing else makes the players more powerful than they were before.

Takewo
2015-09-02, 08:49 AM
It would change all kinds of things. 3d6 means that even very small differences in modifiers have very big effects on what is possible. For instance, imagine you're dealing with an enemy with an AC of 20, and you have two characters: one with an attack bonus of 4 and the other with an attack bonus of 7. The character with a +4 is locked out, basically, with a roughly 6% chance of hitting. The +7 has a 26% chance, which is a pretty huge swing. Add +1 to that and the hit chance goes up by more than 10%.

Which is exactly what happens when you get more average rolls (sorry, I meant rolls when I wrote 'results' in my previous post). The rest is no more than the outcome of those rolls.

Knaight
2015-09-02, 09:19 AM
Just remember that reducing the randomness of a system will make the system favor the players more(an conversely, increasing randomness favors the NPC's). Just switching from d20 to 3d6 and changing nothing else makes the players more powerful than they were before.

Not really. It will make the system favor those who are at present better more. That could easily be the NPCs in a lot of situations.

Thinker
2015-09-02, 10:23 AM
Not really. It will make the system favor those who are at present better more. That could easily be the NPCs in a lot of situations.

The wisdom of TheOOB's statement is that it reduces the chances of extreme rolls. In cases where critical hits are a thing, an NPC critting a PC can be devastating for the whole adventure, but a PC critting an NPC just ends the encounter quicker.

Bobbybobby99
2015-09-02, 11:50 AM
One should note that play by post has a large advantage as far as the ease of using this particular change goes. While it is about three times as difficult to roll 3d6 as 1d20, it is actually slightly easier to type 3d6 than 1d20 with a dice roller.

Just an observation.

NichG
2015-09-02, 11:58 AM
My advice would be, start with a zero-variance system and figure out how that will work. There are several examples out there - Amber Diceless, Nobilis, etc. When you go to zero variance, it requires having some other way to develop uncertainty - usually in the behavior and decisions of the players and GM combined with private information (e.g. 'I don't know how good he is at this... if I use only a bit of my strength, then I'm likely to lose but at low cost; if I use all of my strength and still fail, the cost is much bigger' kinds of things).

If you can make sure that even at zero-variance the system doesn't break or misbehave or end up being boring, then you can add however much variance you want back in and it should still be pretty stable.

TheOOB
2015-09-02, 02:27 PM
Not really. It will make the system favor those who are at present better more. That could easily be the NPCs in a lot of situations.

The basic idea goes like this. When an encounter happens, there are two rough results, the PC's win, or the PC's lose. PC's tend to experience many encounters over the course of a campaign, while most NPC's only experience one encounter, or a small number of encounters at most. Generally speaking, the PC's are assumed to win most encounters by default, or at least lose in a way that doesn't stop the campaign, most campaigns don't have every encounter having the PC's win only by the barest margin and the luck of the dice.

Increasing randomness can lead to one side winning the encounter that wasn't intended. While this can cause the PC's to win an encounter they otherwise wouldn't, it is more likely to cause them to lose en encounter they otherwise would, as before the random element is thrown in the PC's are expected to win most of their encounters. It can also cause NPC's to lose fights they might otherwise win, but since we expect NPC's to lose anyways, that's not a big deal.

Basically, randomness can make you win when you would lose, or lose when you would win, but since players are expected to win most the times and it's a much bigger deal when they lose, randomness tends not to benefit the PC's as much as the NPC's, who are expected to lose, and winning is a much bigger deal for.

Lvl 2 Expert
2015-09-02, 05:21 PM
So if you want more uncertainty, you need to look at a larger variance in your throws.

I recommend two D6's, their results not added up but multiplied. The average number thrown is pretty close to the average number thrown on a D20, but there is also the option of throwing 24, 25, 30 or 36. Extremely critical success? ;p

Knaight
2015-09-02, 08:11 PM
The basic idea goes like this. When an encounter happens, there are two rough results, the PC's win, or the PC's lose. PC's tend to experience many encounters over the course of a campaign, while most NPC's only experience one encounter, or a small number of encounters at most. Generally speaking, the PC's are assumed to win most encounters by default, or at least lose in a way that doesn't stop the campaign, most campaigns don't have every encounter having the PC's win only by the barest margin and the luck of the dice.

On the other hand, in plenty of systems it's really common for the PCs to generally be individually weaker than their adversaries, on whom they are ganging up (particularly among the subset of fights that are actually dangerous). What edge they have is due to them getting more rolls, allowing them to overcome statistical disadvantages they individually have. Similarly, there are plenty of skill checks and similar that only need one person to succeed, and on which the odds of getting a success for each individual character is somewhere below 50%. Both of these disadvantages are highly exacerbated if low variance die methods are introduced.

Mendicant
2015-09-02, 08:25 PM
Which is exactly what happens when you get more average rolls (sorry, I meant rolls when I wrote 'results' in my previous post). The rest is no more than the outcome of those rolls.

Right, but the question was "will I screw things up" which requires that you do more than just note the general trend of roll results. You also need to look at how that it going to practically affect play. The answer is that it will affect play pretty drastically, messing with pretty much everything in the game, big and small.

The range of monsters you can run effectively will shrink for any given player level, as once manageable disparities between defenses and offenses become much more pronounced. Intra-party balance will get swingier. You'll need a whole new system for crits. It'll get harder for players to eyeball their chances of hitting a specific target number. You'll have to rework items and WBL, as the bonuses provided by various items will be far more significant. Getting rerolls will be less significant. So will the ability to take 10 under duress. A RNG with an even distribution of results is an assumption that is hard-baked into the game's design. Going to an RNG that's more like a normal curve will not be kind to that design.

Takewo
2015-09-03, 09:56 AM
Right, but the question was "will I screw things up" which requires that you do more than just note the general trend of roll results. You also need to look at how that it going to practically affect play. The answer is that it will affect play pretty drastically, messing with pretty much everything in the game, big and small.

The range of monsters you can run effectively will shrink for any given player level, as once manageable disparities between defenses and offenses become much more pronounced. Intra-party balance will get swingier. You'll need a whole new system for crits. It'll get harder for players to eyeball their chances of hitting a specific target number. You'll have to rework items and WBL, as the bonuses provided by various items will be far more significant. Getting rerolls will be less significant. So will the ability to take 10 under duress. A RNG with an even distribution of results is an assumption that is hard-baked into the game's design. Going to an RNG that's more like a normal curve will not be kind to that design.

Now I remember someone saying that an argument Giant in the Playground forum is basically a bunch of people who pretty much agree on the subject, but have minor differences in their point of view.

You feel it is necessary to state everything that follows more average rolls, I feel that it follows from the very statement. You seem to believe that all those consequences mess excessively with the game (you seem to, I might be wrong), on the other hand, I think all those things make the game more desirable.

So, basically we agree on the topic, but one feels more explanation is needed and doesn't like the new flavour it brings to the game; the other one feels that a short explanation is enough and sort of likes the outcome of more average rolls.