I was playing in a sci-fi game and my hacker wanted to set up a dead man’s switch on the environmental systems in a space station: When the melee characters launched their assault, I’d pull the dead man’s switch. The GM had me roll to set up the dead man switch… and then roll every single round to maintain it… and then roll again to throw it even after I’d set it up successfully.
Eventually, given the endless series of checks, I inevitably failed.
Laying aside the fundamental misunderstanding of what a dead man switch is, the negative effect of this sort of thing can be quite severe in an RPG session. In this case, the group quickly realized that we should never, ever try to make a plan: If we just improvised something, it would be resolved in a single roll and we’d have a chance to succeed. If we actually put together a plan, on the other hand, it would just invite lots and lots of dice rolling until the plan failed.
The solution is fairly simple: Let It Ride. Have the character make a single skill check that determines the ultimate success or failure of their endeavor.
Another solution is a complex skill check (making multiple checks until X successes are achieved). These tend to be very elegant in dice pool systems, and when you want multiple checks to be made they’re an effective framework for allowing that without the all-or-nothing of a single check ruining the entire attempt.
GM DON’T #2.1: FAILURE IS POINTLESS
The flip-side of rolling to failure is the “roll pointlessly until you succeed” thing. For example, you’ll often run into games where the PCs need to unlock a door: There’s no time pressure and no consequences for failure, and yet the GM will sit there and have the PCs roll over and over and over again until they finally succeed.
One way to deal with that is something like the Take 20 mechanic: If you can eventually succeed at this, then we can assume that you will eventually succeed at it and we can move on. Letting it ride can also solve this problem by providing the opposite outcome: Your failure on this Open Locks check tells us you are simply not good enough to pick this lock at this time and in this way. (The single check determines your relationship with the lock and until you can substantially change the situation, your character is going to be stymied by that lock).
GM DON’T #2.2: TOO MANY SEARCH CHECKS
A somewhat related problem is when a multi-step action resolution gets broken down into too many discrete parts. This can take many forms, but the most cancerous form I’ve seen in the wild came from GMs who took the Search guidelines for 3rd Edition D&D way too seriously. Those guidelines specified that it took a full action to search a 5-foot square. That’s a useful guideline for combat (when you might want to know how much area you can search during a single round), but some GMs took this to mean that you needed to make a separate Search check for every 5 foot square. So if you searched a 10-foot-wide hallway that was 40 feet long, you’d have to make sixteen (!) separate Search checks.
This isn’t rolling to failure because each chunk is a legitimately separate task. (Failing to search in Square #1 doesn’t mean you won’t find anything in Square #2.) But it murders pace – which is either directly undesirable or undesirable because it discourages players from using the specialties affected by the problem.
The solution here is to collect the tests into meaningful chunks: Searching an entire room (or even suite of rooms) is obvious. Alternatively, if they want to search the dungeon hallways as they move along, let the result of the check ride until they either make a meaningful choice to do something other than search down the hall OR until that check result produces a result (either success or failure) that they can recognize as such (i.e., until that check either finds a trap or secret door or until it fails to do so and the trap happens to them or the ambush pours out of the secret door they missed).
Sometimes you’ll end up with a player who demands multiple checks. In some cases this is because they, too, are following bad mechanical advice (like the “make a check for every 5 feet” misinterpretation of the Search rules). In other cases, it’s a manipulation of metagame information (“I know I rolled poorly, so let’s have that only apply to this one specific area and then I’ll make another check”). Often it’s because they’re irrationally trying to manage risk (“I’ll only search this little chunk so that if I roll poorly the effects will be minimalized” — which doesn’t make sense because your odds of discovering any given hazard remain unaltered, but that doesn’t mean people don’t do it).
Most of the time your response to this is fairly simple: You tell them no.
The exception would be the rare instance where it’s actually effective pacing to stretch out the mechanical resolution. Like a slow motion shot in a film, these are the times when specifically highlighting each small, discreet, tension-filled moment serves to escalate the crisis and leave the table on the edge of their seats.
Identifying these moments is a gut-check, not a science. For example, I was just about to say that it would never be Search checks down a dungeon hall… but then I realized that there actually was a time that I followed a player’s lead in the Tomb of Horrors to separately search every single inch of corridor because that mechanical resolution was so completely right in capturing the paranoia and terror the group was experiencing in that moment.
LET THE GOOD TIMES ROLL
On a closing note, let’s be clear that not every series of sequential rolls with a non-discrete outcome is rolling to failure. We’ve already discussed the situation where each individual check is a separate, meaningful accomplishment. But it’s also true that, for example, combat isn’t a roll to failure even though it involves multiple checks culminating in a single outcome of life-or-death.
It’s also useful to note that rolling to failure can be an effective choice if you’re actually looking at a situation where failure is assured and the interesting question is how long a character can stave off that failure. For example, how long can you say conscious in a vacuum? How long can you hold the door against the werewolves pounding on it from the other side?
I launched the GM Don’t List back in August 2015. It got waylaid almost as soon as it began because, as I’d described to my Patreon backers awhile ago, while I was writing the second installment I realized that what I was actually writing was a huge chunk of the middle section of a completely different essay series.
Reading this revamped version of the second installment, you can probably figure out that what I was talking about was the Let It Ride and Vector portions of The Art of Rulings. I believe when I abandoned the original attempt at this second installment I had written 4,000+ words and most of them were focused not on the ostensible precept of what a GM shouldn’t do, but instead analyzing the nuances of letting it ride in great detail.
So, basically, I realized I needed to finish writing The Art of Rulings (at least up to the point where this material would be covered) before I could return to the GM Don’t List. Then it took a bit of finessing to figure out how to make this entry actually valuable as a stand-alone contribution.
That finally accomplished, I anticipate that you’ll be seeing more of the GM Don’t List series in the near future.
“That finally accomplished, I anticipate that you’ll be seeing more of the GM Don’t List series in the near future.”
Great, I can’t wait! thanks!
And a fine stand-alone contribution it is.
2.1 happened happened to me with the Curse of Strahd: Death House adventure.
We knew there was a secret door behind a certain mirror, it just made too much sense (our mapping was on point). Plus we *had* to find it to advance the story. But we weren’t allowed to discover it through role playing, only rolling. It was so frustrating and pointless.
Player: There’s a secret door behind here, I know it.
DM: Roll perception.
Player: 12.
DM: You don’t find anything.
Player: Can I… can I try again?
DM: Sure.
Player: 8.
Dm: Nope. You need a 15.
Player: Can I keep trying?
DM: Sure.
Player: 11.
DM: Nope.
Player: 17!
DM: Ah! You find a secret door.
The decision to get rid of Take 20 in 4E and 5E was, IMO, one of the dumber moves in those games. I understand that it’s difficult to objectively set a universal standard for when Take 20 can be used and, therefore, you have to rely on the GM making a ruling — which, of course, was anathema to WotC’s designers — but it’s just such a ridiculously useful mechanical concept.
(Synchronistically, that link also deals with Ravenloft adventures.)
I liked the Take 20 rule too, it was just a logical extension of the laws of probability.
But even without it, if the players are pointing to exactly the spot where the secret is, just give it to them. No rolls. It means they’ve been paying attention to the DM’s careful scene descriptions, and the DM should reward that. Rolling dice shouldn’t supplant deduction.
DM: You follow the trail of blood up the stairs, and into the master bedroom, where it goes under the bed.
Player: I search the room.
DM: Roll investigation.
Or.
DM: [[same description]]
Player: I look under the bed.
DM: You see a scared goblin, cowering with a bloody knife.
Absolutely. Player expertise can trump character expertise.
Great article, Justin. Looking forward to more of these now!
Take 10, Take20 and ‘Take0’ are some of my favourite rules in Pathfinder just because of how neat and concise they are. They are relatively easy concepts to explain and is so useful both from a player and DM perspective.
I incorporate Take 20 into my rulings. As long as you’re capable of completing the task, then the failed roll just indicates an amount of time until you succeed.
The main reason I like this better is that it still requires a roll, the roll is still meaningful, and the amount of time is more variable than just taking 10 or 20.
So Justin McGuire’s example becomes:
Player: There’s a secret door behind here, I know it.
DM: Roll perception. (Actually, I would also ask you to describe what you’re looking for, but barring that…)
Player: 12.
DM: (Knowing they need a 15) OK, you move the mirror out of the way and poke and prod at the wall, it takes a few moments (3 rounds), but you find a secret door.
If there is a time element (like they are trying to escape something), then I’ll play up the time, round by round.
Incidentally, for things that have a catastrophic failure (like a trap), I generally have separate rolls. That is, failing to disarm a trap doesn’t mean you automatically set it off. I have a separate DC for disarming and triggering the trap.
This article could use a link to the 3rd in the series.
I’m enjoying this very much, thank you for posting your work.
When it comes to multistage checks, I have a pretty simple solution.
Perception checks are searching for what’s visible in the room. It won’t trigger any traps (unless the player walks over to look at something and there is a trapdoor there, but that’s the movement component and not the perception)
Investigation checks are sticking your hand in something, looking for air holes or buttons or whatever. This will trigger traps and only effects the five foot square you are investigating, but has generally lower DC.
4th Edition did remove Take 20, but it retained Take 10 and introduced “passive” checks. Not everyone seemed to agree with how to use passive checks, but one intent seemed to me to be that it wasn’t always necessary to call for Perception (or Insight, or knowledge) checks, because sometimes the information would just be obvious to certain characters.
4th Edition also encouraged DMs to get away from stand-alone “gotchas” like traps and locked doors. If you’re hit with darts while walking down the hallway and take some damage, and there’s nothing else going on… you heal the damage and move on. Or you take all the time you need to unlock a door.
If, on the other hand, a trap is shooting darts while you’re being attacked by skeletons, or a locked door is keeping you from reaching an ally who could use immediate help, then deciding how or whether to deal with the trap or the door becomes an interesting choice for which every dice roll matters. Does the rogue deactivate the trap, or help take out the skeletons? Does the rogue pick the lock, or does the fighter bash the door down?
4th Edition also raised for some the idea of interesting failure. While skill challenges were widely panned, one key aspect of them that many people missed was that failing a skill challenge shouldn’t lead to a dead end, like a lethal trap, or an impassable door, but to a new situation informed by that failure. Dealing with traps and doors could be an extended skill challenge for which rolls are made now during travel through a dungeon, with success meaning that you arrive before the evil ritual ends, and failure meaning that you arrive after it has been completed. You’re not stuck behind a door, but you still failed and have to deal with what that means.