One of my rules of thumb for evaluating RPG combat systems is the number of times you have to roll to resolve an attempt to hit, on average. For example, in D&D, you have to roll twice — once to hit, and once for damage. In Vampire, you roll at least three and often four times — once to hit, maybe once to dodge, once for damage, and once for soak. In Feng Shui, you roll once — the roll to hit is also the roll for damage.

My assumption is that (assuming a standard combat system, rather than something more narrativist) fewer is better, because it make combat flow more quickly. There’s an orthagonal concern, which is getting the feel of combat right; for most games, you don’t want to say “roll 1d6 and if you get a 4 you hit, and if you get a 5 you hit and kill.” That’s quick and simple but most likely not satisfying.

However, I recently decided that this is too simplistic. While playing Mutants & Masterminds, I found myself getting all antsy about the combat system. Which is weird, because it’s simple: one roll to hit, one roll for defense.

But it’s a different person for each roll!

So, the addendum: you have to take information transfer into account. Go back to D&D. Roll to hit, tell your opponent what you rolled, roll for damage, tell your opponent what you rolled. Compare and contrast to M&M — roll to hit, tell your opponent what you rolled, tell your opponent what your Damage bonus is, your opponent rolls a Damage save, your opponent tells you if you did damage.

Same number of rolls, but you keep having to pass information back and forth. Feng Shui, the ruling champion of quick combat systems, is way simple: roll to hit, tell your opponent what you rolled, opponent tells you if you did damage. Hero is on par with D&D — roll to hit, tell your opponent what you rolled, roll for damage, tell your opponent what you rolled.

Of course, you really ought to figure in math complexity. It’s easier to do the math in Feng Shui than it is in Hero, and Hero is noticably more complex than D&D (since you’re counting BODY and STUN from the same roll, and adding a lot more dice).

The key observation, though, is that information transfer matters. I’ve heard more than one game designer talk about giving the defender a chance to roll to “involve him in the game” and so on, but I begin to think that’s a misguided concept.

## 7 Comments

You forgot Tunnels and Trolls. Anyway…

I do think that allowing the defender to roll is a worthy goal, and I think the best way to do that is to put all the math before the roll, so that both players can roll simultaneously, and anyone inspecting the dice knows what happened.

You’re also forgetting RuneQuest, where the attacker made a weapon skill (minus the defender’s Defense) roll, and the defender then rolled to parry; you then had a list of success levels (critical, special, success, failure, fumble) to compare, and THEN you rolled damage. So this approach has a long and honorable history.

HeroQuest, for that matter, has a similar mechanic–opposed rolls, with a list of relative success levels to compare–with the interesting twist that ‘damage’ isn’t rolled directly, but *bid* before dice are rolled–it’s an abstraction of how much risk the character takes. Haven’t played HQ yet so I don’t know how well this works in practice.

Unknown Armies, of course, hybridizes RuneQuest’s success levels with Feng Shui’s one-roll philosophy.

There are billions of variations. Good point on UA, and T&T, for that matter.

I like/find interesting the idea of doing the math before the rolls, then giving both players dice. Not sure how it’d work but it’s intriguing.

There’s a few other variables here, I think…

Complexity of rolls you somewhat address with the Hero comment, but it’s clearly true that some die rolling schemes (say, just roll a d20 as in d20 or HQ) are simpler than others (say, roll 8+ dice, group them by like value and compare the groups, as in Whispering Vault, or counting “height” and “width” as in Godlike). So just reducing the number of rolls may not be a perfect measure if the new rolls are each significantly harder to evaluate. I’m a big Feng Shui fan, but I’ll still admit that open-ended rolls are, on occasion, time-eaters. Rolls that are just “roll one die and read result” are much quicker than “roll lots of dice, think, reroll, etc.”, and this has to be factored in somewhere.

Also significant is who does the rolling. I recall at least one game (I believe it was called Chaotik) where all die rolling was done by the players, leaving the GM free to do other things. Thus, players rolled to-hit when attacking and to-dodge when defending, the GM never had to pick up a die. An interesting idea, and it could be reversed to GM-only dice use if that’s what you want, all without moving to the “extreme” of diceless play.

A real advantage of both-participants-rolling, IMO, is that when two people roll, you can use a simpler (perhaps even linear) distribution on both rolls and still get a bell-shaped result distribution… That appeals to the probability math nerd in me. This is one of the things that appeals to me about Hero Quest, as it uses fairly simple rolls (d20 vs. rating) but achieves a nicer distribution by making most all rolls contests (effectively making all rolls 2d20 for distribution purposes).

I was kind of hoping you’d chime in, Jack. ðŸ™‚

I agree — some rolls are more difficult to make than others. If I had infinite money with which to design a game, I’d do some real UI testing on this issue. I think it’d be very interesting to time a number of people on various types of rolls.

I’m moving my reply over here from Bryant’s thread, as I’ve wandered a bit from his original points. The essential topic here started as how many times you need to roll dice to resolve an action (usually a combat action)…