I tried something at my weekend Beyond the Wall game that may need to have a few more iterations before I’m totally happy with it, but seemed to work well enough to mention here.

I’ve never been a fan of using margin of success in D20. For one thing, it makes skills work differently than the other reasons you roll a d20 in the system. When you attack or save, you care about whether you met the target number or not, and often what number the die displays (for auto-miss or crit), but you don’t typically get any benefit from rolling significantly better than the target number. So having to track how much your result exceeded the DC for increased success immediately makes the skill system feel bolted on, like it came from another game.

And, in general, those other games that use margin of success for skill results have some kind of weighting to the roll, such as a dice pool or adding together multiple dice. In those systems, there is usually one level of success that’s much more likely than the others based on how the dice are weighted (e.g., in Fate, you’re very likely to get a margin of success equal to how much your skill exceeds the target, and much less likely to get four higher or four lower than that). But when you use a d20, there’s a 20-point range of margins of success that are equally likely. Particularly for non-iterated checks (like most Knowledge checks), the results can wind up feeling very swingy (e.g., “Sorry, you missed out on getting really useful clues because you rolled low and only just made the DC; you would have gotten much more information if you’d rolled higher.”).

So I was very interested when I noticed (via Shieldhaven using it in his game) that 5e had added* the concept of the group skill check. In the base rules, it’s something you can do when the whole group is trying to accomplish the same thing that requires a skill (e.g., stealth, climbing, etc.). If at least half the party succeeds, everyone succeeds (the higher-skilled individuals are assumed to cover for the lower-skilled).

As written, this is a useful addition that solves a lot of standard issues (such as always having to leave the armor-wearers behind when trying to sneak around). But the variation I tried goes even further:

  • Virtually anything that the whole group could work together on can be a group skill check (e.g., perception, knowledge, persuasion, etc.).
  • Instead of rolling, a character can Help another character, and share the results of that success or failure (in BtW, helping is a specific action that can only be done if you have the skill or spend a Fortune Point, but I don’t think it would break anything if you allowed your D20 variant of choice’s version of helping). You can’t combine helping in this way (i.e., you can’t pile help on the person with the highest skill check to push her to no chance of failure; at least half the party needs to actually roll).
  • Instead of requiring a simple pass/fail based on party size, before the roll the GM has in mind the general spectrum of what it means if no one is successful up to everyone being successful. Very difficult results may require the whole party, easy ones may only require one success, and more successes might grant a better result over the minimum pass.

This essentially winds up splitting the difference between a dice pool roll and a 4e skill challenge. And it allows the GM to give out better results for more successful rolls without any actual roll caring about anything other than pass/fail (and maybe crits, if you use skill crits). Importantly, it doesn’t incentivize low-skill players to avoid participating the way 4e skill challenges did (because each player only gets one roll, so you can’t sit out to let someone else go multiple times, and because helping is as good as succeeding yourself). My players seemed to dig it, so I’ll probably keep experimenting with it. I welcome thoughts on possible improvements in the comments.

* This is the first place I saw it; apologies if it originated somewhere else.