Savage Worlds : There's something funky with those dice...

Savage Worlds assigns characters values for their attributes and skills based on a die type, from the usual collection, d4 through d12. That always seemed like a bad idea, but I couldn't put my finger on why until I started doing the math.

Now, granted, player characters are, by default, special, and roll their skill or attribute die plus a d6, and take the best result. SW dice are also exploding, adding the result of another roll every time they roll their highest value. Most common tasks use a Target Number of 4 to indicate success, and every 4 that it's beaten by is one Raise, for a better result. Skills at default roll with a d4-2, and apply the same -2 to the Wild Die d6.

This leads to a situation where you have the following occur:

d4-2 + WD = 20% success + 13% raise = 8- on 3d6 (more like 8.5)
d4 + WD = 43% + 19% = 11- on 3d6
d6 + WD = 49% + 26% = 12-
d8 + WD = 57% + 25% = 13-
d10 + WD = 45% + 40% = 13- (a little more, but closer to 13 than 14)
d12 + WD = 38% + 50% = 14-

When you're performing normal tasks, even the most highly trained individuals in most games are going to be facing a 14- roll. If the TN was raised to 8, that expert drops to 10-.

So, my gut feeling that there were some oddities being introduced to the game system by reliance on our old standby dice seems to be a valid one (I won't get into the situations where it's actually better to be rolling a smaller die...) The system doesn't support as broad a range of ability as it might, and tends to hew quite closely to the middle ground, with even the most rudimentary of expertise only available at the highest end.


  1. I agree with your calculation. Something I should have said under "RPG Litmus Tests": if there's a funky die mechanic, especially if it involves rolling multiple dice and taking the best result, that's a bad smell for me because the probabilities are almost certainly not what the author expected. In fact I'd say that if you're designing a game, you should make quite sure that one of your playtesters is the sort of person who works out the odds.

    Savage Worlds shows another of the classic problems: if your open-ended system is "roll another die and add", then the odds of getting at least (die size) are the same as the odds of getting (die size + 1). Early editions of Shadowrun did the same thing.

    (And yes, it does look worryingly as though a target number of 6 is more achievable with a d4 (0.25 * 0.75 = 0.1875) than with a d6 (0.16667). Similarly at TN8, a d6 is better than a d8.)


Post a Comment

Popular Posts