Sunday, December 1, 2013

Regarding hidden rolls

In spite of the infrequent publications, I actually do intend to keep this blog.

Last week, while browsing some OD&D/OSR rule sets, I was thinking about player rolls for skills whose outcome the players shouldn't know immediately -- hiding, moving silently, searching, sensing motive (if you buy into that sort of thing) and so on. The problem with having the DM roll for all these sorts of things is that, in my experience, players would rather be holding the dice. After all, it's their character who's attempting the check, so why should someone else be rolling? But, the DM says, you can't know the outcome, so I'm going to roll. Lame.

In D&D variants, rolls are binary checks for success/failure, with success happening in some range. (That is, either 1-3 on a D6, per OD&D skills, or, say, 12-20 on a D20 after all the modifiers are all thought out in 3.5/Pathfinder, etc.) How do we hide whether such a roll is a success? Shift the success range randomly.

Whenever the player attempts a roll whose results shouldn't be known immediately, the DM rolls an identical die/dice and adds the result to the success range's smallest number, spilling over to the lower end if necessary by subtracting the number of faces on the die. (This is just modular or clock arithmetic, if you like.) For example, in the 1-3 on a D6 example above, if the DM rolls a 2, the player's roll is a success on a (1+2) = 3 to a (3+2) = 5. If the DM instead rolls a 5, the roll is a success on a (1+5) = 6 to a (3 + 5) = 8 -> (8-6) = 2. Thus, the player's roll is successful if it comes up 1-2 or 6. The odds of success are always the same, and the player still rolls the die, but there's no way to tell what a roll means without knowing the result on the second die.

Alternately and equivalently, one add the DM's roll to the player's. again wrapping around when the result spills over the maximum die value. Using the above example again, if the player rolls a 2 on the D6 and the DM rolls a 3, the result is (2 + 3) = 5, outside the success range. On the other hand, if the player rolls a 5 and the DM rolls a 3, the result is (5 + 3) = (8 - 6) = 2, a success.


  1. I'm not so sure about this one. The player who rolls a natural 20 and discovers (moments later, in the course of play) that this was a failure might feel like an exciting roll has been stolen from them. But then I come from a school of thought that just says, "Your character doesn't know what number showed up on the die."

    1. I always want to believe that I'm from the same school. In practice, I find that players have just enough difficulty discriminating between their own knowledge and character knowledge so as to damage honesty of play, even if it's only subconscious. If nothing else, it tends to kill narrative suspense of wandering int he woods if the players know they're going in the right direction because they rolled well.

      Still, one could repair the "natural 20 always succeeds" simply by never shifting it out of the success range. That is, consider the range one elements smaller (12-19) and leave 20 fixed when you shift everything else. However, since skill rolls in particular aren't auto-success in Pathfinder, I wouldn't necessarily feel the need to do so using such a system.