Dice Hacking in Gumshoe

by Joshua Kronengold and Catherine Ramen

One of the most interesting features of GUMSHOE is that failure is frequently not an important part of the gameplay. Whether by finding a clue with an Investigative Ability, or spending enough points on a General Ability test to ensure success, GUMSHOE games focus more on the what (which Ability do you use, what do you do to invoke it, what you do with information after finding it) the why (Drives, at minimum), and how much (do you spend points now or hold them in reserve for a future effort? How many?)

The thing is, as much as automatic successes make PCs seem (and players feel) like badasses, in games where the characters have broad ranges of skills and large amounts of points to spend, the die rolls on General Ability tests become less interesting. Varying the difficulty can add some drama to rolls, but it has to be done carefully, especially if the Difficulty is not revealed to the players. Unknown Difficulties that run higher than the expected 4 can discourage point spends as players become conservative, or cause frustration as too many tests fail because of underspends or running out of points. Keeping Difficulties to a narrower and more predictable range lets players make more strategic choices about when to spend points–but also tends to make any rolls a foregone conclusion and the die roll a pro forma task. In this article, we look at several ways to give die rolls drama and keep the results interesting. At their best, dice do more than moderate between players; they provide interesting and surprising results that nobody in the session would have chosen, while still staying within the bounds of what people consider an acceptable result.

So, then, how do we open up die rolls?

The first method, which appears in a rough form in Night’s Black Agents and TimeWatch, is to add another possible result to the roll: you get a Critical success if you roll a 6 and beat the Difficulty by 5. (Note that this will encourage some overspending by players, but both TimeWatch and NBA make it relatively easy to refresh pools or find extra points when needed, and that mathematically it encourages what we consider optimal play–usually spending just enough to guarantee success) However, this tends to still produce only two possible results for a roll. If you didn’t spend enough to guarantee success, the possibilities are failure and success. If you did, it’s success and critical. But either way, it’s a pure binary result: yes or no; crit or normal. There are three possibilities, but only two of them are possible on any given roll.

Catherine has designed a system that opens the results even more, by creating a system of “Benefits” that can be accessed with a high enough roll. In her system, for every four points that you exceed a Test’s Difficulty by, you can choose a Benefit (you can choose the same benefit more than once). In her last campaign, the list looked like this:

  • Terrible Harm: + 4 damage
  • Armor Piercing: you negate the target’s armor
  • Speed: You succeed very quickly
  • Unnoticed: No one sees what you do
  • No Traces: No obvious signs of what you did
  • Safe: You don’t expose yourself to danger
  • Disable: You break or damage an object
  • Disarm: You knock a weapon (or other held object) free
  • Suppress: You stun, force under cover, or otherwise prevent someone from acting
  • Opportunity: The PC may immediately take a second, related task
  • Missing Materials: You can succeed even without the proper tools

The system also allows you to take a Benefit if you are willing to take a consequence; one of the implicit Benefits is “You Succeed,” and several skills (like Intrusion or Filch in NBA) include Unnoticed or some other benefit by default) so a player willing to let their character suffer a consequence or lose one of the benefits of success could succeed even on a failed roll. When deciding on a consequence, often the easiest thing to do is simply reverse one of the Benefits: so instead of being safe, a character might take damage; their effort might take a great deal of time, or be unable to conceal. Note that the Benefits list can also function as a quick way to reward an Investigative spend, for example a Library Use spend that finds the results in only a half a day’s search. A GM might even allow a player to purchase a Benefit using an appropriate Investigative spend if it fit the situation–using Intimidate during a fight to keep other characters from joining the fight, for example (using the Suppress Benefit).

This “Margin of Success” method makes the rolls more interesting, since any roll might result in a valuable Benefit, as well as rewarding players who overspend on a Test since they will probably receive a Benefit rather that effectively wasting any points beyond the amount needed to guarantee success. (This is the reason why the margin to receive a benefit was set at 4–it gives a 50% chance of receiving a Benefit on most tasks, provided the player spends at least 3 points. For a grittier feel you could raise the required margin to 5 points or even 6; lowering the margin will tend to produce very competent characters.)

The Benefits list has to be customized for the GUMSHOE game you’re using, the particular style of your group, and the skills that are available.

That said, there are limits to this approach. While the variable benefits allow for players to make many more choices about how to customize successes and failures, die results have gone from being binary to sometimes-ternary–depending on how much someone spent. At best, the possibilities are failure or success, or success with a benefit — or success, success with a benefit, or success with two benefits (etc). While player choices can customize this result after the roll, the results are still going to be strictly bounded.

To really make sure we don’t know what will happen, we have to make die rolls truly open ended. That way, you’ll never know exactly what result you’re going to get–and can let yourselves get a little excited every time the dice hit the table. There’s a simple and well tested way to do this — use the same system that quite a number of other games have used and have a 6 or 1 result in another die getting rolled (repeating this as necessary). The problem with that is that most of us (and the authors are certainly in this set) are attached to the way that GUMSHOE’s results are more predictable than those in most others, while the “exploding six and one” system is incredibly, unalterably random.

So Josh suggests that we tone it down a bit. Whenever you roll a 6, it explodes, but to keep it slightly flatter, use a d3 instead of a d6 (if, like us, you don’t have a d3, just halve the results on a d6 and round up). And if you roll a 3, keep going.

Similarly, we can have rolls of 1 implode (if you like surprise failures–If you don’t like it as GM, or the group doesn’t want it, don’t use it). If you’re using this rule, you should also allow players to “take one,” setting the die to 1 (and not risking it imploding) rather than rolling if they want; this means that, if they’ve successfully gauged the difficulty, they get an automatic success. As an optional rule, only allow taking 1 if they did, in fact, spend enough to guarantee success; if they declare this and it’s not enough, tell them to roll anyway, but only after they’re committed. This avoids players being unfairly punished for “taking 1” in a situation where the difficulty is unknown. Imploding is less fun for players than exploding, so add one to the d6 roll before subtracting, but otherwise it’s the same in reverse–roll a d3 and subtract the result, but keep going if a 3 is rolled.

We’d be fine with stopping there, but we know some people are going to want a fumble system (and some people are going to really, really, really NOT want a fumble system). One option is to invert the Benefits system so that for every 4 points you fail the test’s Difficulty by, the GM assigns a consequence which is the reverse of one of the Benefits. If you want fumbles to be more rare, you should make them only happen if a 1 was rolled — and that should probably be the default for defensive rolls, particularly when target numbers are unknown, so that players don’t roll flat against a difficulty of, say, 15 and suddenly — surprise! — they take multiple consequences just for playing. After all, it should generally be better to roll than not to roll, even if the odds are long. Even if you don’t want to lock yourself into a fumble system, this can be a good guide to estimating the consequences of failing a high stakes active roll. The PC doesn’t get the benefits of a success, and may (particularly if they miss the difficulty by 4 or more) end up exposed to danger or notice, drop or damage something important, or an opponent might gain an opportunity. Gumshoe is often about pretty competent protagonists, so you can do all of this without having to make the PCs look incompetent or foolish.

And that’s it — several modular, open-ended, still very GUMSHOE-Y systems that should add a bit of anticipation to every roll you make — and provide a few entertaining surprises. Try it out, and let us know what you think!

Postscript (by Joshua Kronengold):
Since we drafted this article, Robin published an article talking about how one can embrace failure in Gumshoe general tests (as well as another on automatic successes).  Our article, on embracing and shaping exceptional success and softening failures, could be seen to act as a complement to Robin’s thoughts on the subject.

On the one hand, we have automatic successes (or failures).  Those are the points where you don’t a particular result as interesting — so you don’t allow for it at all, instead simply ruling success to be automatic or impossible.

Then we have rolls where failure is acceptable.  There (as with the addition of my highly caveated fumble system), Catherine’s system allows for a  failure to be a negotiation, not a catastrophe, shaping the narrative rather than cutting it off.  Perhaps you fail at your overall goal, but salvage one particular element from it that was more important than the goal itself.  Or perhaps you succeed at that goal, but at the cost of excessive time, damage, or some other mischance–either way, the dice steer you into a result that falls short of your character’s goal–and instead steers you into a different avenue that might end up being more narratively satisfying rather than a wall that you feel compelled to
bang against until it falls.


Catherine Ramen has been playing role-playing games for almost four decades. She is the designer of the upcoming Red Carnations on a Black Grave, a story game about the Paris Commune, Rovers, a customizable space-opera rpg about loveable anti-heroes, and edited the English edition of Nerves of Steel, a film noir story game.

Joshua Kronengold lives in Queens, New York with Lisa Padol, surrounded by books, games, and musical instruments.  He is a decades-long contributor to Alarums & Excursions, and has contributed to Over the Edge (in Edgeworks #3), Reign, and Unknown Armies.

2 Responses to “Dice Hacking in Gumshoe”

  1. Steve Dempsey says:

    I’ve been trialling what to do with 1s and 6s in Fearful Symmetries. Mostly it’s some kind of “and” or “but” system. If you roll a 1 and fail, it’s “no and” what’s more something else uncomfortable happens, if you succeed, it’s “yes but” etc. With a 6, you get “yes and” something nice.

    I usually decide on the 1s but take suggestions from players if they like. With 6s I offer a few suggestions and am happy for players to make their own too.

    So, for example, when making a particular magical item, the player rolled three 6s so I offered some exciting extra benefits as the stars were right.

    • Joshua Kronengold says:

      That’s awesome! I think in many ways that’s a less mechanical approach to the same concept! Expanding out, you could say that a 6 on a success is “yes and”, a straight role is “yes”, a 1 on a die where you’d spent to guaruntee success is “yes, but”, a straight failure is “no, but”, and a 1 on a straight failure (where you hadn’t spent enough to guaruntee succss) could be “no, and”.

      With repeated 6s or 1s adding emphasis to the and or the but.

Leave a Reply