General mechanics

[Murtak]

I thought it might be useful to gather a repository of general mechanics, for use in whatever homebrew systems we are working on. There are so many dice systems, resolution mechanics, stat systems and the likes out there, it is getting hard to keep track of even the obviously useful ones.

Resolution mechanics
Dice vs DC:
Roll one or multiple dice, add them, add some number (stat, skill, etc), compare to a fixed DC. Advantages: Simple, one roll only. Disadvantages: Defender feels uninvolved.

dice vs dice:
Roll one or multiple dice, add them, add some number (stat, skill, etc), compare to another similar roll. Advantages: Everybody is always doing something, still fairly simple. Disadvantages: Double the die rolls you strictly need. Other: Inherently produces a bell curve (May or may not be desirable).

Die mechanics
d20 (D&D):
Produces 20 equally likely results. Very variable results due to the large die and large standard deviation.

3d6 (GURPS):
Produces a bell curve. With just 15 possible results (and 4 of them very unlikely) opposed rolls drop off the RNG very fast.

xd6, count successes (Shadowrun):
Produces a bell curve which keep scaling into infinity. Has issues at low numbers of dice.

xd6, keep y after rolling (L5R):
Rolls are in the form of RollX/KeepY - you first roll x dice and then you pick and keep y of them, then add them together. Produces a very flat bell curve, with an everpresent possibility of rolling really low or really high (especially with exploding dice).

Exploding Dice
A system may opt to have the player reroll any die that rolled the maximum possible number, then add the result to the original roll. This second roll may or may not explode again. A similar mechanic is of course possible for minimum results (imploding dice?).
Exploding dice allow system with otherwise fixed maximums to a possible for even the most unskilled characters to potentially succeed at any task. Note that exploding dice add sharp edges and plateaus to a possibility distribution. (I.e. with exploding d6 a 6 is twice as hard to get as a 5, but a 7 is not harder than a 6 at all, and a 8 or even 9 only marginally harder than a 6)


I'll post more on attributes, levels, skills, wounding, death, spellcasting/resource systems and so on if/when I get some more free time, unless someone else beats me to it of course. Also if someone can do some probability distributions on the various dice mechanics and post them as images that would be great. I'm finding it hard to model the L5R mechanics without brute forcing the solution.

[Avoraciopoctules]

Eh? What about poker dice and their incorrigible incomprehensibility?

[RandomCasualty2]

I sort of like the shadowrun system as a general mechanic. The only real drawback is that you're rolling armloads of dice constantly, so it's rather slow to do at the table, especially at larger amounts of dice. The system can only scale so much before you're rolling too many dice to keep track of.

Also another interesting feature which can be applied to the Shadowrun system is changing the target number to alter the randomness of a task. I'm not talking about SR3 where both sides in an opposed contest are rolling against different TNs, but rather just changing the TN for both sides to change predictability of something.

This is most prevalent in strength tests to break something or arm wrestle, where for instance, a TN of 2 would make a lot more sense, because it's a case where the strong guy almost always wins.

[Sir Neil]

xd6, count successes (Shadowrun):
Produces a bell curve which keep scaling into infinity. Has issues at low numbers of dice.

What is the recommended lowest number of dice?

[Murtak]

Enough to get a decent curve - say 3ish. This also avoids the weird point at two dice, where, if you use glitch rules (50% ones = glitch) going from one to two dice doubles the likelihood of a glitch.

That reminds me - i forgot the Feng Shui mechanic of rolling a positive and negative die and adding a base value. Again produces a nice curve and lends itself well to glitch mechanics.

[CatharzGodfoot]

That reminds me - i forgot the Feng Shui mechanic of rolling a positive and negative die and adding a base value. Again produces a nice curve and lends itself well to glitch mechanics.

Much like Fudge dice, which (IIRC) are [+1, +1, 0, 0, -1, -1].

[Draco_Argentum]

Eh? What about poker dice and their incorrigible incomprehensibility?

I think he was after systems that have at least one person believing they're good.

[Thymos]

I think there needs to be 2 additions.

Reroll, and roll another die and drop lowest.

Some systems embrace these, others don't, but how much you include them very much affects the game. So while these two don't necessarily stand by themselves, or are must haves (hell, the add a die and drop lowest is actually incompatible with some of these), they are definitely something to consider.

Basically I think these two can be tacked on to some die systems, without being separate die systems themselves.

[RandomCasualty2]

Enough to get a decent curve - say 3ish. This also avoids the weird point at two dice, where, if you use glitch rules (50% ones = glitch) going from one to two dice doubles the likelihood of a glitch.

Yeah, the glitch rules in general have a few problems. the most notable is that your explosive ammo is more likely to blow up in your gun when you're firing blind.

[IGTN]

A variation on Roll x keep y: Roll x keep k.

Basically it works like the 3d6 bell-curve system, although with rerolls built in (and bonuses being, essentially, partial rerolls). This (and roll x keep y) also has the possibility of adding "drop highest z" to your pool.

So, basically, you could have a system with k=5, making a roll where you have 3 bonus rerolls and a penalty reroll. So you roll nine dice, take out the highest one and the lowest three, and add up the rest. You could also have dice added that skip the drop (so you could have a roll where you keep the highest, drop the second-highest, take the next four, and drop the rest).

This has the downside of a very tight curve if you're using convenient dice, especially with a bunch of dropped dice added; any results that are at all in question are probably going to end up marginally on one side or the other in such a situation.