Checks, bonuses and probabilities

Keith J Davies 

FWIW, in Echelon attacks are contested rolls as well. Same number of successes = both hit (and both take damage). Otherwise, if I have more successes than you then I hurt you (and hurt you _more_; each success after a tie adds more damage), and vice-versa. 

But there are also complications. If you roll 3 successes and 1 complication, and I roll 3 successes and 0 complications, you get to pick: tie (and both cause damage) and suffer a complication, or pay off the complication and take more damage (because I rolled more successes now than you did). 

The active character does get to decide if there will be an exchange of attacks. If you attack me, it's on. As the target I could choose to ignore this (passive defense only), but that seems like an uncommon case so I largely ignore it. If I'm better at fighting than you are, or luckier, it might be that you end up getting hurt too (or even instead!).
Most games, you attack me and I have no recourse or decision at all. You decided to attack, I decide to... nothing. I get hurt or not entirely at the whim of you and the dice gods. 

The mutual attack thing has some rather interesting characteristics. I ran models that suggest that it does a lot to balance out the action economy. Being outnumbered is still bad, but it's a much more even fight than in most games. In fact, I appear to have hit a sweet spot in the design. Even with four levels difference (i.e. exactly one tier higher), beng outnumbered is bad. The higher-level fighter has better rolls and slightly more hit points, but the sheer number of attacks makes things more even. The more skilled fighter is likely to get more successful (tied or better) attacks than any one of the lesser fighters, but they get spread among all the lesser fighters, while the lesser fighters' attacks are concentrated on one. 

I forget if the approximate balance point is 2:1 or 4:1 (I think it was 2-3:1, to be honest). I know that 8:1 is very very bad, don't let that happen.

Matt Miller 

I also like the d20 'explode' rule, where 20 hits and 1 misses, better; there is always a chance of success or failure, regardless of other conditions. Opposed rolls (d20 vs. d20 or 3d6 vs. 3d6) are analogous to 2d6, where the resulting roll is very much a bell curve. 3d6 vs. 3d6 creates a very strong, very regular bell curve, where the difference in medians is decisive a very large 

Dan Felder Skill checks are one of those things that I feel like were reasonably solved a while ago. While it may be possible to improve them, I feel like the effort it takes compared to the tiny amount of improvement it provides is not worth it. Unless you are altering your skill check system thematically for a specific game design goal of course, but even then it seems iffy.

Rob Hicks Summary: die mechanisms that favor median rolls, (2d6, etc,) don't do well if you are using opposed checks combined with bonuses, because it so heavily favors the higher bonus. Underdogs will almost always lose, even with only a slight numeric disadvantage.

Keith J Davies unless, of course, that's the behavior you want. I can easily imagine a tiered system[1] where you have a 'tier bonus' that shifts your entire roll, simply because higher-tier _is better_.

[1] level 1-4 = tier 1, level 5-8 = tier 2, etc.

In a D&D-esque system you might then have "rolls are 3d6. You get +1 per tier, and your 'class skill' (fighter is combat, wizard is casting, rogue is skill) gets +1 more" [no, I wouldn't use these numbers, but they're good for discussion]

Fighter _always_ has an advantage over everyone else at the same tier in combat, and can fight evenly with a non-fighter above his tier... which means a non-fighter above his tier is no worse than he is. The higher-tier fighter will still have advantage over him. This might be exactly how you want the system to behave. I wouldn't say these systems "don't do well", I might say "have characteristics that must be understood".

Rob Hicks Dan Felder Still worth looking at this stuff, because tinkering is fun. Also, the d20 system isn't "solved" by any stretch. It still has some serious drawbacks that aren't answered by other prominent systems.

Like · Reply · October 1, 2015 at 9:39am

Dan Felder Tinkering is always fun. But note "reasonably solved". As in we know the benefits of various skill check systems and have solid enough options, to the point that putting in added effort has a significant opportunity cost compared to working on other aspects of design. smile emoticon

Like · Reply · October 1, 2015 at 9:41am

Matt Miller "d20 is too random"...which could reasonably be solved by moving to a d12 or d10, and simply adjusting numbers. ...but that would actually make it more random, as the high and low automatic hit or miss would be much more common.

Rob Hicks 
Which is not a bad thing. I like d12s. It occurs to me, that you could get interesting results from a d12 if you did 1's as an automatic miss, 10's as automatic hits, 11's as a crit, and 12's as a double-crit. Make the rest of the game built around 1-10 as standard results. "Turning the speakers up to 11," as a parallel. It would make hits more common, and crits much more common than a d20 system, but that is more interesting, isn't it?

Dan Felder 
Seems like it'd just make things more complex. If you want variety in your system, something akin to FFG's Narrative Dice system in their star wars RPG is a good place to start.

Matt Miller 
No good. Makes 10-11-12 into hits, so you automatically hit 25% of the time, even against impossible enemies. Higher hit rates need balanced with either lower damage or higher HP. To be sufficiently variable to be interesting, damage can only go so low (1d4?), and lower HP makes for easier mental math.

Paul Goldstone 
I think players like Bonuses - so the greater the deviance - the more bonuses you can give. It's been shown several times a +1 on 2d6 is worth more than in D20, when you look at +2 through to +5 the difference is startling.Many games reward players with bonus gear in some tangible form, from the classic magic sword to the self aiming blaster. When looking at the system some thought should be given to what rewards the game will give especially deep into a campaign as well as end game one-shot adventures with powerful pre-gens.  If the roll effectively becomes a given, and therefore trivial at the top power end, then player/gm enjoyment will suffer. Furthermore is disparate parties where either due to power level, stats, gear if one player is tougher than the other players game balance can be skewed significantly which could lead to power creep, as the encounters have to become much harder to compensate for the tougher/luckier players. Thematically you could look at players choosing what they roll, ie, d20 or 3d6 in a game, maybe based on skills, talents etc.  The choice means against a particular tough opponent, highly dramatic situations the players could choose the d20 - needing the higher numbers. In non dramatic situations, they could choose the 3d6 looking for more likely success.

Matt Miller

When we made Mayhem, one the design principles was directly opposed to providing more bonuses.I think there are 13 different compatible bonuses to hit in DnD 3.5 We called it the "Christmas Tree" effect, with character liberally strung with magical ornaments. Rob Hicks DM'd a long DnD 3.5 campaign where the characters managed to acquire...every single bonus possible.

But what happens is that you cease having a 'character', and start having a 'toon'. You no longer have an avatar of yourself, but rather a 'vehicle' whose stats you have optimized. Letting players choose 3d6 or d20 seems like an interesting mechanic, but I don't think it actually adds much. The more mathematically talented would quickly know which represented the better alternative...ie, anytime your target number is over 10, roll the d20.