Enabled per-comment hyperlinks, which should make things easier to follow. And to put them to good use…

Josh says:

“It’s an RPG, but the dice rolls themselves are arbitrary. On the middle of the table is a large target of concentric circles, and your performance is based on how near to the bullseye you get with your massive handfull of dice. Somehow managing to get all twenty of you d20 into the centre of the target (the size of a large coin, say) would probably destroy evil, all of it, with no saving throw.”

Kevan says:

“It’s marbles played with polyhedral dice - each player has a set of dice in their own colour, and they throw them into a circle to either knock opponents’ dice out, knock them onto different faces, or simply roll high scores in safe places. Whoever has the highest total values at the end is the winner.

Possibly some strategic dice-selection elements at the beginning, where lower-scoring D4s are more stable than D20s, but also less good at knocking other dice out. (Perhaps players start with 30 sides-worth of dice, of their choice.)”

Like I said in that thread, I think these are worth combining, under the mental tag of “tactile objects first, random numbers second.”

Here’s my pitch: Everybody brings his or her own set of dice; these dice must be all the same color, and that color must be unique in the group. When you roll, you have multiple objectives:

• Get a decent number.
• Knock a competing player’s high-scoring dice onto lower-scoring faces.
• Land your die in a more central ring, which means a higher multiplier or bonus.
• Knock a competing player’s dice out of those rings (or, alternately, help move a collaborating player’s dice further in).

I’m thinking maybe your effectiveness in the current situation, whatever that might be, is determined by your current-highest die / multiplier / bonus combination (rather than the sum of all your dice). This could help sharpen a players’ goals (”I regain control if I just knock out that one 8 in the second ring…”).

Advantages: Dice are fun to throw! Moreover, this would balance the old “randomness is king” RPG trope (and board game trope, for that matter) by making player skill a factor. The situation could be in constant flux as players pay [resource] to withdraw dice from the table, gamble more dice in an attempt to disable an opponent, et cetera.

Problem 1: As Kevan pointed out, the d4 would be a midget muscleman here–not a high scorer, but almost impossible to change faces. The d6 would have a similar advantage; you’d need something more unbalanced like the d8, d10 and d20 to get interesting rerolls. I’m inclined to suggest that this be played with the tubes of 10d10 that people buy for White Wolf games.

Problem 2: Anybody got a theme? Kevan’s “er, magically summoning demons into a chalk circle” is good, but why are we summoning demons? What are the consequences of failing your summoning (do they just poof, or escape and eat your pets)? Do the demons get to roll?

Ken says:

“how about a gambling/rpg???

like, for instance, if you’re really about to get into a fight, you dont really have a concept of hit points. if theres a car about to hit you, you dont start hoping “gee, i hope this car isn’t capable of dealing more than 4d8 damage, or i’m probably screwed cause i only have 24 hit points available.”

somehow players would have to have hit points, but the amount they had could be hidden from them. could be interesting, yet somewhat hard to implement and/or control.”

Like I said in the comments, this made me think of an RPG that uses craps mechanics the way some RPGs use card-game mechanics. I’m not aware of any published RPG that uses craps, because RPGs do such weird things with dice that they tend to consider six-sided dice and gambling passé. I’m sure someone has tried it at some point, but obviously it hasn’t made a big splash.

I think it could, though. I’m going to introduce a couple things real quick: how craps works (based on a five-minute read of the Wikipedia craps article) and task resolution vs. conflict resolution, which we’ll get to in a second.

In craps, you place a bet for or against the shooter. The shooter rolls two dice and hopes they come up as a total of 7 or 11. If they do, the “for” side wins. If they come up 2, 3 or 12, the “against” side wins. If they do neither, everybody takes note of the total (called the “point”) and the shooter keeps rolling. If the point comes up again before a 7 does, the “for” betting side wins. If the shooter rolls a 7 before rolling the point, the “against” side wins. The farther from 7 the point was, the greater the payoff.

There are a lot more weird rules about betting, but this is the essential mechanic. Obviously, if you can bet for or against somebody, there’s no advantage for the house; this is why most casinos “bar” either a 2 or 12 roll and keep all bets if that’s what’s rolled the first time. (Mean casinos bar a 3.)

To me, this mechanic seems tailor-made for conflict resolution. Vincent Baker explains this concept better than I can at Roleplaying Theory, Hardcore–scroll down to the fourth entry, and notice where I subconsciously got the name of this weblog–but it’s essentially this:

• Task resolution is success or failure, eg “do I leap across this pit?” or “do I hit him?”
• Conflict resolution is winning or losing, eg “do I beat you at tic-tac-toe (and therefore get your Skittles)?” or “does he run me over (and therefore injure or kill me)?”

There’s only one interesting side to task resolution; there are two interesting sides to conflict resolution. That’s why I think the opposed-betting mechanic of craps would be cool for it. One could consider it functionally equivalent to flipping a coin–except craps is more complicated, and so more fun to mess with.

First problem: the odds for the “against” side are actually a little better than those for the “for” side. Player Characters in an RPG should win more than half the time, unless they don’t want to; simply reversing the sides (”try not to roll a 7 or 11″) is counterintuitive, especially for somebody who already knows how to play craps.

Second problem: completely random mechanics produce completely random results. Both sides of a conflict should have more agency than they do in the standard craps model.

There may be a way to solve these problems simultaneously, with a good add-on mechanic. For instance, say that each side of the conflict has limited resources with which to buy bar points (normally the role of the casino). You can take 2 or 12 for free, and spend to bar out a 3 or 8 or whatever–the closer to 7, the higher the cost. The problem with that, in turn, is that bar points are useless after the first roll, and it once again becomes a matter of chance leaning toward the opposition.

How else can we make manipulating craps resolution interesting? Swapping out dice? Buying rerolls? Standing bonus or penalty points for every roll?