image Have you ever had this happen at a game?

“I roll to see if I know anything about ancient ankthyrian architecture and where they might have hidden a secret door. Crap, didn’t get it. Let me just try that again…”

or how about?

“Ok, I know I’ve rolled 8 times to try to force the door, but the 9th time is the charm!”

or what about this golden oldie?

Ok, we want to avoid a fight. I’ll try to talk them down since I am the only one who knows their language.  Crap. If they don’t immediately attack, can I try it again?”

Rolling the dice in most games denotes an attempt at an action by our characters, if the dice come up in our favor then the action is successful, if they don’t we are usually screwed. But is the failure scenario always such that the action can’t be attempted again? Since some actions can be attempted again we, as players,  sometimes feel like other actions can be attempted again and will try for it. Sometimes this is true, sometimes it isn’t, but it is often fairly subjective as to when a reroll is appropriate.

So that is the hot button for today. When do you allow players to reroll actions? Does it depend on the type of action? Physical, remembering something, diplomacy? Does it depend on the fail scenario? What about when the players are in a jam and really need to get a win scenario? Will you let a reroll occur then, even if you haven’t before? Do you consider the reroll a separate action? What if enough time has passed between rolls for something like remembering info or figuring out a puzzle, will you allow a reroll then?

To me, this is one of those issues right up there with fudging the dice as a GM. I’ve got players who will try to push the bounds and reroll anything. Other players will consider themselves doomed and walk themselves to the guillotine, even when I would have given them a chance to work their way out of the situation. Like so many questions and issues, I find this very subjective. So how does it play at your table? What issues do you have with it or what situations will you absolutely disallow it in? When do you consider it acceptable to reroll a failure?

IMG BY hatchibombotar | CC 2.0

33 replies
  1. mrtopp
    mrtopp says:

    For me, players can re-roll if one of two conditions has been met:

    1: There has been a consequence to their previous failure.
    2: They have done something that might alter their chances.

    An example of scenario #1 is a situation in which the PC is stuck in a trap. If the trap does ongoing damage, then each failure has consequence (another round of damage), so they are permitted multiple attempts. If it does not, however, and they have failed their roll they are stuck there until someone comes around to “free” them.

    Scenario #2 is well-illustrated by a failed attempt to bash in a door. In this scenario the PC could not *just* try again — if they keep rolling until they get a success, there is no consequence (so why the dice?). However, if they go find a battering ram, they can make a new (different) attempt.

    The two guidelines are basically insisting that such a die roll have consequences. If you (as the GM) are unwilling to dish out the consequences of failed rolls, then those rolls should not be needed.

  2. MaW
    MaW says:

    I consider it on the basis of what the check is. If it’s plausible for them to try again and maybe get a different result, then they can roll again (assuming something else doesn’t come along and distract them). If it’s something you only get one chance at, or where you’ve completely exhausted your options within one roll, then there’s no point rolling again because it doesn’t simulate anything in the game world.

    Things that generally aren’t rerollable are knowledge checks, you either know something or you don’t (or you’re only rolling for detail because the GM knows that your character inevitably knows some particular (ir)relevant fact). Likewise things that happen under extreme time pressure, and probably lockpicking. Another character might come along and try their lockpicking, assuming the first one didn’t break off a pick in the lock or something…

    I would allow, in-world events permitting, characters to retry anything at all if they come up with another way to attempt it. Lockpicking failed? Okay, shoot out the lock – but that’s not a reroll of lockpicking. I would generally prefer that people think of a different way to do it.

    But then, since I’m currently putting together a Paranoia group, failure’s probably destroyed their ability to try the same thing again anyway, and they might need to wait for their new clones to arrive too.

  3. shadowacid
    shadowacid says:

    I’ve always liked the idea of re-rolling failures. This was one of the things that I really loved about Star Wars Saga Edition. They had a nice solid mechanic as either a racial or class ability to re-roll checks.

    I’ve found that the general feeling of failing a check is “crap, that wasn’t fair” but if there is a re-roll you get a chance to redeem that failure, and if you fail the second roll it really feels like “ah crap, you really got me that time!” and there is a lot less angst about failing rolls.

    I liked the re-roll mechanic so much that I actually wrote a short superhero rpg where the primary function of powers is to re-roll checks.

  4. JonathanS223
    JonathanS223 says:

    When it comes to the adventures, I usually decide to allow them a re-roll on the basis that the action could be repeated if the situation was real. Like diplomacy could deal with a re-roll as they can try to speak again while falling and grabbing a hold of a branch, then missing would really not need a re-roll as they will hit the ground.

    I personally think it’s dependent solely on the situation at hand.

  5. Christoph
    Christoph says:

    I don’t like rerolling failures; it negates the importance of the roll. It seems to me that if a player can continually reroll failures without consequence then they shouldn’t have needed to roll at all. Perhaps the failure should merely indicate the amount of time it takes to succeed. A great system is one where both success and failure have interesting outcomes.

  6. BrianLiberge
    BrianLiberge says:

    I’m in a similar boat to MaW. Knowledge checks are not treated as a test of in and out memory. You either know it or you don’t. If the player should know it no matter what, like his old master’s name, then there is no check.

    Same is true with failure’s that have immediate consequences. I like the example of breaking a lock, but the other strong one for me is a roll of a 1 on a social skill. Other character’s may be able to jump in and redeem the situation as a whole but the duke is gonna remember for a while that the party cleric called his beloved kingdom of Cradathan, Crapathan.

  7. Dunx
    Dunx says:

    One of the many, many things I like about Savage Worlds is the benny system: each player gets a small number of chips they can spend to reroll check results they don’t like (damage is excluded from this usually – there is an edge to permit that, though). That and the wild die means that the number of gross failures is very small.

    The upside is that as players we very rarely feel the need to reroll gratuitously, and when bad failures do happen we more often take the licks from it.

  8. deadlytoque
    deadlytoque says:

    I’ve always said if you’re allowing re-rolls, then you’re doing something wrong. I only call for a roll if a) I think there is a likelihood of success; AND b) I can come up with a way for the person rolling to have failure that is as interesting (or more!) than success.

    So, never. I never allow re-rolls. If a player makes a die roll and fails, then it’s up to me to interpret that failure in a way that keeps the game on-track and is exciting.

    Really, if this die roll is absolutely essential to the story moving forward, then (IMO) that’s one of two mistakes on the GM’s part: either creating a “bottleneck” where there’s only one right path to success, or not being (lack of preparation) or not being prepared to diverge from the session-as-envisioned (lack of improvisation).

    I know you can only learn so much from a book, but I’m willing to bet that most GMing chapters and GMing books agree with me.

  9. lomythica
    lomythica says:

    I used to allow refills for really important situations, but over time, I learned that the problem was that I was putting too much pressure on a single roll. I really hadn’t planned for the “what happens if they fail the roll” outcome. Now, I always include options for failure that Llow the game to continue, albeit differently when a failed roll occurs.

    I also really like success and failure ladder mechanics. They make it easy to determine the outcome of a failure or success, which I feel makes my job easier, and makes it easier for the players to deal with, as it is part of the mechanic. We are currently playing Doctor Who Adventures in Time and Space, amd the success ladder has really added a lot of fun and depth to the stories.

    Lastly, I really like the bennie/story point mechanic model. It gives some extra chances, and some cinematic flair without making failure a nonexistent option. Works great for cinematic games. I have yet to have a player ask for a retool since we started playing games that use story points.

  10. lomythica
    lomythica says:

    Sorry for the typos above…. Using my iPad.. Auto correct is not always correct. Got to check things over better.

  11. Roxysteve
    Roxysteve says:

    I think it all depends on what the context of the given action is. Sometimes I treat failure as a success that takes an inordinate amount of time to achieve (example: Library Use to find critical clues in Call of Cthulhu, Computer Use vs moderate difficulty to discover traces of perfidious MJ12 activity in my D20 Delta Green game).

    In D20, the “re-roll until everyone is heartily sick of it” factor is catered for by Taking 20 of course. The rules for allowing that are good guidelines for re-rolls under other game systems.

    I agree with Dunx (me being a newly-minted Savage Worlds cheerleader) though I’ve found that there are some very common misconceptions about what is allowable as a Benny Reroll. Making a player trade-off something for that re-roll if they want it badly enough is a brilliant innovation.

  12. John Arcadian
    John Arcadian says:

    Wow. Lots of great comments early in the day.

    @mrtopp – 2: They have done something that might alter their chances.

    I wholeheartedly agree with this one, but I think the door bashing example you give could still be done without changing the situation. The door would (likely) weaken with each attempt and sometimes good old stubbornness could solve a physical situation like that. I do agree with the doing something to alter the chances though. When my players do “Would i know about x…” rerolls, I ask them to find a reason why they could make a reroll. The knowledgy characters have taken to carrying around books they can reference to justify the reroll. They usually also try to humor it up: “Aww crap, I was spelling it with a K! The demon’s name is actually spelled with a Q!”

    @MaW – “But then, since I’m currently putting together a Paranoia group, failure’s probably destroyed their ability to try the same thing again anyway, and they might need to wait for their new clones to arrive too.”

    Paranoia opens a whole different can of worms! :)How do you handle knowledge rolls that occur later? i.e. Character rolls for knowledge of particular monster and fails. 2 weeks in game time pass and they encounter a similar situation, rolling again. Do you mark down they failed the first one? Do you assume they might have learned new info in the 2 weeks?

    @shadowacid – I must admit, I had never really thought of it that way. Rerolling does offer a chance to retry and little room for argument if you failed the roll twice.

    @JonathanS223 – I sometimes let rerolls happen in “grabbing branches” situations like that because they make for a good story. I admit to being influenced by players who roleplay actions out more. If a player pantomimes the frantic scrabble and entertains the group, then I’ll be more likely to give a reroll for something like that.

    @Christoph – Rerolling can undo the significance of a roll. Do you find room for it in limited situations in your games or just say no rerolls? I picked up the “high or low” game from a buddy when questions like this arise. I roll a dice (hidden) and ask the player “High or Low”. If it comes up how they guessed I am generally more lenient, if not then I go hardcore.

    @BrianLiberge – Are you more lenient for things that the character would know but the player might have a hard time remembering? I’m bad at remembering NPC names as a player, thus I always write them down, but I always feel my character would have a much better grasp on the situation.

    @Dunx – Bennies and other plot altering points systems tend to negate this debate. They players can spend a resource to alter the game and try again. I use “Mulligan” points in some games (especially when I need to speed up things for a short session) to allow: reattempts of actions, preventing from getting ambushed or done over in a particular scene, and to smooth out a situation or puzzle.

    @deadlytoque – That’s a good way to look at a failure consequence. It has to keep the game moving and exciting.

  13. Roxysteve
    Roxysteve says:

    @lomythica – Well, you should be using your iPad in a reality that allows rerolls.

    It’s ironic we’re discussing re-rolls on a forum that provides no “preview” capability.

  14. evil
    evil says:

    The only time I allow rerolls is if the game cannot continue along the path unless the players get past the obstacle. If they fail the roll and cannot figure out any of the other ways I’ve set up to allow them past the obstacle, then I’ll allow a reroll, but usually I won’t allow more than one. At that point if the characters still cannot get past the obstacle, I’ll introduce a new piece (ex., some burly guards walking through a locked door and bashing the PCs).

  15. BrianLiberge
    BrianLiberge says:

    @John Arcadian – Absolutely. There are often weeks or months in real life between in game hours or days.

    Obsidian Portal has helped with that, where players can quickly look up past events and, if I’m lucky, review before the game.

    But like I said, if there’s no reason a character should know something, there’s no check involved. If the player can’t remember I’ll fill it in.

  16. Rafe
    Rafe says:

    Why ask for a roll if you or the player won’t accept failure as a result? Since the only other option is to succeed. . . .

    Roll the dice or say yes. If you’re rolling the dice, the roll has to matter, and a failure is a failure and a success is a success. Punish failure appropriately within the context of the scene (and system being used), or reward success.

    The end. 🙂

  17. mrtopp
    mrtopp says:

    @John Arcadian – “I think the door bashing example you give could still be done without changing the situation. The door would (likely) weaken with each attempt and sometimes good old stubbornness could solve a physical situation like that.”

    This rests on the assumption that “bashing in the door” consists of a single bash at the door. This seems unlikely (outside of a critical success) — if the door is relatively likely (say 30% chance) of breaking down on any given kick, it should be an automatic success. Any roll should be whether it goes on the first attempt, or after repeated (loud) bashing.

    Each roll should have two potential outcomes — one for success and one for failure. In the door scenario, I see two possibilities:

    1. Success bashes open the door. Failure means you are not strong enough to bash in the door, and the PCs need to take a different approach (or abandon the idea of getting past the door).

    2. Success kicks in the door on the first attempt. Failure means that repeated bashing is needed to break through the door, giving whatever lies behind it warning and time to prepare.

    There could be others, perhaps your stubbornness could be described as:
    Success means the door is bashed open.
    If the roll fails, repeated bashing takes at least one hour per point that the roll was missed by, and requires a successful follow-up roll. Failures stack.

    So if the player requires a 12, but only rolls a 10, they can take the stubbornness approach … but it requires two hours of game time and has no guarantee of success. Such an approach could be useful in a time-sensitive adventure, where the PCs have only 24 hours to complete some mission.

    The important thing is that success and failure are defined ahead of time, and that failure means something other than “keep rolling until you get a success”. In the bog-standard dungeon scenario, where other definitions were not made part of the forward planning, failure has to mean that the character cannot break through the door in the manner attempted. Period.

  18. Shaun Welch
    Shaun Welch says:

    @Christoph – “A great system is one where both success and failure have interesting outcomes.”

    Exactly! Failure should rarely ever mean “nothing happens.” If the PCs are going to roll to bash down a door, it should be because something will happen if they don’t. The question isn’t “Do we kick down the door?” but “Do we kick down the door fast enough to surprise the baccarat-playing gnolls?” (Gnolls all love baccarat. This is a FACT.) On a success, the door comes down and the gnolls are surprised. On a failure, the door still comes down, but all the gnolls have doffed their razor-rimmed hats and are ready to throw them!

  19. lomythica
    lomythica says:

    @Rafe – Personally, I would phrase it that the die roll is what it is. But not all successes are the same amd not all failures arw the same. Especially with a success/failure ladder. I was more prone to offer rerolls before I learned about success ladders. The black and white success and fail made me nervous, partly because I am a storyteller style GM and don’t like a dice roll to dictate big elements of the game (which may be a hot button issue all its own).

  20. John Arcadian
    John Arcadian says:

    @lomythica – Don’t stress the autocorrect. Over here we constantly hear about Chrome Spew when Martin sends via iphone.

    @evil – That’s a good point about when the roll is important to continuation of the story. I’ve seen players get stuck because they failed a roll and no longer perceived the situation accurately or couldn’t figure another way around.

    @mrtopp – That gets into a big player perception issue. What does one roll actually mean in the game world. It is highly important to define that. If the roll represents one or two kicks to the door then a reroll seems appropriate. If it represents the entirety of the door opening action, then failure definitely should mean failure. I tend to fall to the side of one roll equals one attempt at something in a short period of time unless otherwise specified. If I call for a roll to perform a big action containing many other smaller sections(launching a boat into the water, translating an entire book, etc.) I usually define how long it will take if successful and will make failure more drastic than if it is a smaller standalone action.

  21. BishopOfBattle
    BishopOfBattle says:

    @mrtopp – Stated far more eloquently than I would have put it. I’m the same. I typically will not allow a reroll unless they have suffered the consequences of their failures or the situation has changed.

    @mrtopp – “Each roll should have two potential outcomes — one for success and one for failure. In the door scenario, I see two possibilities:

    1. Success bashes open the door. Failure means you are not strong enough to bash in the door, and the PCs need to take a different approach (or abandon the idea of getting past the door).

    2. Success kicks in the door on the first attempt. Failure means that repeated bashing is needed to break through the door, giving whatever lies behind it warning and time to prepare.”

    I also tend to run sessions with the above two options in mind. This comes out of running Shadowrun as my first game I GMed, which has rules for running “extended tests” which calculate “successes over time”.

    I will tend to divide skill checks into one of the two above scenarios. If the former, then they get one shot at success (unless they can do something to change the scenario). If fast talking the bouncer doesn’t get them in, its unlikely that continuing to jabber at him will change anything.

    If a scenario falls into the latter, then they roll and either succeed or “fail”. How badly they fail determines how long it takes to get a success. So if they’re trying to search a room for an item and they get a complete success, the almost immediately find the item they’re looking for. If not, then it takes them a number of minutes / hours / days (as appropriate to the task) longer for them to complete the task.

  22. BryanB
    BryanB says:

    It really depends on the situation.

    What I won’t do is have the characters spin their wheels endlessly rolling one failed die roll after another when it comes to searches and investigative type of stuff. I’ll let them roll once for quick success and then if they choose to take a lot of time, they can take twenty and get a max result. They pay a cost in time taken or possible discovery by enemies for getting the take twenty (or max roll).

    In the case of searches, failure can be a fun thing. If the group fails to gain the crucial information in one way, they can regroup and try and to get it another way. No clue or item that an entire adventure hinges on should be narrowed down and held to to one failed die roll. That is crap adventure design and I won’t do that.

    I recall a Shadowrun where we were after a scientist who knew too much. Our employer wanted him for his information. Another faction wanted him dead. We had a shootout in the scientists apartment building. It got ugly and a fire elemental killed the scientist. No one’s dice were being nice that night. Was it game over and total failure because the scientist was dead and we had failed to protect him and his information? No. I suggested that the scientist must have had a backup copy of his notes and data results. If it was this important, he would have had a backup somewhere.
    We regrouped, found the backups at the scientist’s girlfriend’s condo and got it to our employer, facing down the opposition along the way.

    The point is that our horrible failure set up our eventual success and that success was that much more exciting because of our failure. Not that it helped the scientist or his girlfriend all that much. 🙂

  23. mrtopp
    mrtopp says:

    @John Arcadian – I just don’t see the point to the second roll in that situation.

    To me, either the PC succeeds on their check (bursts into the room quickly, in your scenario) or fails (slowly wears down the door to enter the room, in your scenario). That’s a single roll.

    Or, they succeed (breaking down the door, in my initial scenario) or they fail (not being capable of breaking down the door). That’s a single roll.

    Another take would be that it takes 3 successes to knock in the door (or one critical success), but the guards arrive in the time it takes to make 5 attempts.

    As far as player perception goes … if each roll means an individual attempt in a non-time-sensitive skill check with no consequence for failure (and I use “skill” loosely in this case), then the roll is effectively meaningless. I really don’t see the point of telling somebody “roll a d20 until you get a 15 or better”, and watching them roll six times. That’s not a mechanic, that’s a waste of time.

  24. Roxysteve
    Roxysteve says:

    I want to agree that a missed roll can mean “all attempts fail forever”, but then I’m drawn to the specific wording of the “Take 20” rule in 3.5-era D20 rulebooks …

    And then there’s that Dresden files RPG that I’ve been wrestling with for weeks, in which the GM advice does indeed urge a “don’t roll at all unless failure provides an interesting story line” approach, and yet the game has a “Benny” mechanism for buying rerolls. If the people who write a game don’t have the courage of their convictions, why should anyone else? (grinning big when I wrote that).

    I had no idea people took this issue so seriously.

    I never gave this much thought before, but it just occurred to me that I have a different attitude to this issue when I’m playing as opposed to when I GM. I’m loathe to re-roll when I play and really don’t sweat it as a GM.

    Does anyone else have this same bipolar view, or something similar?

  25. derobim
    derobim says:

    Been reading for a while, never bothered to register to comment, before, though. Almost did, back during the article about Burning Wheel, but I was too lazy then. Well, you finally hooked me in, congrats 🙂

    I’m a big fan of the Burning Wheel system, and one of the cool things about it is the Let it Ride rule. That is, a result of a roll remains until the circumstances behind it have changed. For instance, if you roll to pick up a giant boulder, and fail, you can’t pick it up. It is assumed that the failure continues to be the result no matter how many times you continue to try, so re-rolls aren’t allowed. However, if you change the circumstances, such as use a lever to get it started moving, you get to re-roll and take the new result.

    Another good thing about Burning Wheel is that every result, success or failure, should move the plot forward. As was suggested in an earlier comment, if you’re rolling to break in a door, it’s not success: you do it, failure: you don’t. It’s success: you break it in on the first try, failure: you eventually break it in, but have made so much noise in doing so that you’ve likely attracted enemies. Either way, the plot moves forward, the group is through the door. But failure brings complications.

    These two rules aren’t system-specific, so I’d try to use them in any game I ran.

  26. Patrick Benson
    Patrick Benson says:

    In regards to my earlier comment I should be clear that if there is a mechanic for re-rolling (bennies, fate points, etc.) I have no issue with that. Also, some tasks may have multiple rolls allowed such as bashing a door down, but the number of attempts allowed should be defined before the first roll is made.

    The problem is the infinite number of re-rolls that may result from ignoring a failure. Your PC can’t pick the lock? Well I don’t care that your PC has spent another five minutes studying the lock. Your PC failed, so let’s move on.

    We don’t re-roll successes when they happen. Why re-roll failures?

  27. Damocles
    Damocles says:

    I think that some sort of re-roll mechanic is a great addition to any game. It allows the players to have a small amount of control over the situation, which empowers them to feel stronger.

    However, I think it is also important that the players try and fail sometimes. Without failure, success would be meaningless. That being said a good GM should always try and make failure just as interesting as success.

    EX: Star Wars Saga: My players had bluffed their way into a comm center at the spaceport, and one of them was trying to wireless-hack the computer while the others made some sort of stink about lost baggage. He got the first roll fine, but (I required two checks) on the second roll failed. At that point I ruled that the system went into lock down and alerted the authorities. Now all the doors that the PC’s got led through on the way in, became major inferences in their escape. They still got the info they were after, they just had to do it at gun point, instead of bloodlessly.

    That, IMHO, is the way one should handle failures, since you never really want anything that is going to stop your game cold… that just sucks.

  28. Shaun Welch
    Shaun Welch says:

    @derobim – “I’m a big fan of the Burning Wheel system, and one of the cool things about it is the Let it Ride rule.”

    I (finally) picked up my copy of Burning Wheel two weeks ago. Let It Ride has forever altered the way I look at success or failure in RPGs.

    I’m not sure how fond I am of action points/benes being used for re-rolls. I like their use in 4E where you get another action. I especially like it when they give players narrative control (and players get a lot of them). A re-roll seems kind of weak for such limited resources, though. Auto-success, now, that’s something I enjoy.

Trackbacks & Pingbacks

  1. […] they keep wanting to retry the check. There is a whole discussion about this going on over at Gnome Stew as I write this, so feel free to read up on that. Suffice it to say, I think rerolling is a bad […]

Comments are closed.