There are three types of skill checks:
- Roll 5d20, take the highest
- roll 1d20
- roll 5d20, take the lowest
Roll 5d20, take the highest:
Insight, and Any knowledge check. Everyone rolls: someone is always going to roll high. Anyone who rolls low will retcon their failed skill check into a successful “aid another” check.
This is why knowledge and Insight checks always succeed. Once one player declares one, everyone else rolls for it too.
Roll 1d20:
A character uses Athletics.
Sometimes you jump over the pit, sometimes you fall in the pit. Just as it should be.
Roll 5d20, take the lowest:
The party tries to use Stealth.
Someone is always going to roll low. This is why no matter how clever their plan is, D&D groups can never avoid combat (that and their “clever plan” is actually really bad). 4e is better in this regard than 3e, where everyone had an opportunity to blow their Hide roll AND their Move Silently roll, but it is still bad.
Given the numerical variance between “roll 5d20, drop the 4 lowest” and “roll 5d20, drop the 4 highest”, it’s really hard to set universal, memorable DCs that can be used for all skills.
In the past, I’ve dealt with this problem by using a variation of the mechanism suggested in the 4e DMG: asking the party to do a single skill check, using their most knowledgeable PC for knowledge checks and their least stealthy character for stealth checks. It’s an improvement, but it’s sad that the guy with the second-highest Knowledge check is basically useless and has wasted his skill investment.
The DMG2 has some great advice about doing group checks, specifically Stealth checks. Everyone rolls their Stealth, and if at least half of party succeed, the group check is successful.
This is a totally great rule. It’s a perfect fix for Stealth. It’s a shame that it’s tucked in a sidebar in the DMG2, not in the DMG1, or the PHB skill section, so that more DMs won’t read it; the “one bad roll screws up the whole plan” syndrome is implied by the Stealth rules and, in my experience as a player, is almost universally used by DMs. This leads to “screw it, let’s do a frontal assault” syndrome, which eliminates a lot of possibilities for sneaky fun, or “let’s send the thief in alone” syndrome, which often leads to thief death.
I wonder whether the same rule shouldn’t be used for knowledge checks. Instead of everyone individually making a check, everyone pieces together their knowledge. If half or more of the party succeeds, the party remembers that beholders are evil, or whatever. This would mean that you don’t have to set knowledge DCs insanely high to challenge the wizard: you can use moderate DCs. The wizard is still helpful, pitching an automatic success, and the second-smartest guy still gets to contribute.
Usually I try to limit the everyone rolls 5 dice checks to things I kind of want the party to know anyway, since obviously the odds to go up greatly!
Also, for perception, remember that passive perception is usually good for stealth checks versus the party unless people note they are “guarding” or what have you.
Also, if the DC is high enough, only really skilled people are going to have a reasonable chance, so there is that to take into account. Like if I require a DC 25 perception check at level 1, only people with +5 in perception have any chance of making it.
The stealth rules definitely make a lot of sense.
As an alternative to the group knowledge check, you could do something where there are a bunch of facts about beholders and every success gives another fact.
Hi guys! Here’s what I do in my games: the PC with the highest score is the “leader”. Others can only aid him with an “Aid Another” action (remember it now gives a penalty if you fail, so not everyone will try their luck). The leader then rolls the check. Other players don’t feel useless. WIN.
This can be further house-ruled to give more weight to those who really invested on the skill (e.g. trained characters give a +3 bonus, +6 on a natural 20). It can be adapted to work with Stealth (e.g. check is mandatory and results in a -3 penalty if failed, 0 if it’s passed, +1 if it’s passed and the character is trained), and so on…
Like it?
I really like the idea of a natural 20 increasing the benefit of Aid Another. Rolling a 20 is exciting; every time a natural 20 is just another success, I feel like it’s a failure of the game system.
If one dude steps on a duck, then of course the whole party blows the stealth check. That makes sense. That is why you have a scout.
If everyone in the group tries a knowledge check, and it’s a toughie, then probably nobody even has a chance of knowing. Take seven doofuses and one aircraft engineer and ask the group an airplane related question. If it’s super easy, the others might know it, but damn that engineer is gonna nail it guaranteed. Ask a hard question like what are some of the difficulties in implementing a fly-by-wire system, and the normal folk are just gonna scratch their heads and look at the engineer. That makes sense, and it’s a good example of how 3E style skills work.
I don’t think there’s a problem here.
Aid Another is also misinterpreted. It says your help might not be beneficial and a limited number of people might be able to help. I’d also tell the players that they can choose to try to know it themselves or else Aid Another, but you don’t get both and you can’t do the other if you fail the first one. Otherwise a gaggle of 100 village idiots would on average have 10 AA successes and thus give +20 to the 101st idiot’s roll, making him the equivalent of a high-level expert. And as politics have shown, a million idiots do not have better ideas than one idiot.
In 2E we have a problem with people kicking down locked doors. The strongest dude would try, and when he failed the next, etc. It was possible for the weak M-U type to kick it down after everyone else failed. But that comes from a poor reading of the rules: the chance to open doors is only for stuck doors. The chance to break down a locked or magically held door is available only to the exceptionally strong (18/75 or something).
This is why I don’t call for “roll-to-know” checks. If someone wants to know something, they do. If they want to take an action that makes practical use of that knowledge, that’s when they roll the check. The knowledgeable person does well, the non-knowledgeable person doesn’t.
This is also why I subscribe to the “Consider failure, consider success” approach to skill checks, and to collaboration with the players. Why are they rolling Stealth? What does failure look like? What does success look like? A failed Stealth check can mean more than “the party doesn’t get away with it.” If I can’t find interesting success and failure, I ask the players. If they can’t either, we don’t roll, we just assume the interesting thing (either success or failure) happens.