Random Game Crap!
1. Fortune in the Middle Stunts: One of the things that bugs some players about stunt systems like Exalted is that it's essentially fortune at the end -- you either have to severely limit your stunts or stunt as though you'd already rolled successfully. So you describe the big cool thing, then roll to see if you can pull off what you already said you did. Now, I (usually) don't have a problem with this, but I can see the awkwardness.
The traditional response of "make the attempt the stunt, without stunting the results" is a solid one, but limited in the Exalted context. So, what Mo suggested was that you do a more typical FitM setup -- general action, roll, describe result. If you describe the result well, stunt it, then you get a bonus to your next roll. Stunts in this situation don't help you with the thing you're doing, they give you a "karmic" boon to cash in the next time you roll. It works quite well in Exalted where there is a "to hit" and "to damage" roll -- and its easy enough to stunt your damage to give you a bonus on your next hit roll as well. (I also considered making stunts more effective by letting them come after the roll and just giving flat bonus successes for good stunts -- 3 dice in Exalted is pretty small. 3 bonus successes after the roll, however, is brutal.)
2. Instinct reactions to stress in Unknown Armies. UA has one of the best madness/stress systems in RPGs today – but it is fairly narrow in its output range for the immediate scene. If you fail a stress check, any stress check, you either berserk, flee, or freeze. This is the same if the failed roll was from watching your mother get beaten to death or if you have a moment in which you're not sure why you just lied to your wife about where you were earlier. And once you start using UA for something other than its designed use (because you're a bad, bad monkey) it gets more startling – Superman beating the fuck out of someone until they die, for example, is probably not all that fitting.
So, I have considered that in future UA sessions I will steal a little bit I've heard about from Burning Wheel: the instinct. Each player would, at chargen, set up an instinct statement for each of their stress meters that details what they do when they fail at that meter. For high-trust games this can be a general mission statement (Bob has a hidden core of rage that makes him go nuts when he gets into the megaviolence, biting and tearing and going for blood from the throat… Bob's self alienation manifests in him becoming cold and callus towards everyone around him, saying deliberately hurtful things). For more standardized games it could be a triple threat of aggression/passivity/withdraw – three default choices you go to when you get screwed in the stress. Violence, for most people, would probably stay "kill it / run /freeze" – but the other meters (Self, especially) could do well with different options.
3. Heroic Stands in Truth and Justice. Truth and Justice is a pretty bad-ass Superhero game. It has one mechanic in specific that I love: the Revolvin Development. In a Revoltin Development the GM basically had the ability to bribe the PCs to accept a sudden turnaround / loss in the situation by giving them mass Hero Points if they take it. It often gets used in order to have the villain toss them into a death trap, escape with the dingus, or do the other things that happen all the time in comics but that normally drive players nuts when they happen in an RPG.
However, while the GM has a "I must win for the plot" mechanism, the players do not have a "I must win for my vision of the character / to have fun at this point" mechanism. Now there are some obvious reasons for this, but for some players such an absence is not a good thing. So, for those who like to be able to sacrifice character growth for the ability to win when it most counts for the player, my wife and I came up with the following idea.
The Heroic Stand: A player may declare that their character simply wins/succeeds at a contest. The cost of this is a number of MAX points equal to the HP cost of a similar "luck be a lady" purchase.
What this does is let the player know they are going to win, but at the cost of their characters advancement. Those that want to win will win, but the character won't grow from it. As with the comics, the time when a character is most likely to grow is when they put themselves out there, but when it isn't so important they can just pull it out in the end, wrap up the issue, and go home.
4. Heroquest is a badass system. However, I sometimes find extended contests drag on to long, or go to short. And simple contests are all over in one roll. But a recent post on the Forge Heroquest forum made me realize there is a way to do a "medium length" contest using variable augments. I like this idea, because it lets players and GMs together decide (through a slightly push/pull mechanism) how long they want contests to go on, based largely on how many interesting poses they can think up. If you have an interesting modification to throw in, in it goes – but if you're out of ideas (which can happen in the middle of an extended contest, leaving it as a bean counting exercise) then you bring the contest down.
Here are the rules I was tinkering with: As with all Heroquest contests, set the stakes and chose the primary abilities that will be used for the contest. On your turn you declare either "modifier" or "ender" for your action. A modifier contest either raises or lowers your or your opponents default ability for the contest. An ender brings the contest to an end, win or lose.
A modifier action uses a non-primary ability of yours to augment your primary ability, or your primary or non-primary attribute to lower the primary ability of your foe. You describe your action, including how it could help you or hurt them in the contest, and then they must resist with an ability that would let them counter what you are trying to do. If you win you get an augment to your primary ability or they get a penalty as shown on the following chart.
Complete Victory -- 1/4th of the attribute as a bonus to you or penalty to them
Major Victory -- 1/3rd of the attribute as a bonus to you or penalty to them
Minor Victory -- 1/10th of the attribute as a bonus to you or penalty to them
Marginal Victory -- +1 to primary attribute or a -1 penalty them
Marginal Defeat -- -1 to primary attribute for you or a +1 bonus to them
Minor Defeat -- 1/10th of the attribute as a penalty to you or a bonus to them
Major Defeat -- 1/3rd of the attribute as a penalty to you or a bonus to them
Complete Defeat -- 1/4th of the attribute as a penalty to you or a bonus to them
Note: Secrets that give greater benefits for augments are treated as 1 level more successful. A secret with a Complete Victory gives ½ of its rating as a bonus.
An ender works just like a simple contest – one roll between the current values of the primary abilities, winner takes normal results for a simple contest.
So when you want to show your abilities, you must use them and win with them to have them help you. It also makes a step between simple and extended contests – longer contests, but with out AP bean counting.
GMs and players can still use it to determine the length of the contest – mooks might always go for an ender, meaning they only get one roll (or players could penalize the big bad by hacking down his mooks, which he has to resist with their crappy combat score instead of his own massive one), while big bads may do multiple modifiers (with PCs doing the same) before the climactic ender. Similarly PCs can try to stretch things out for a longer fight if they are at a disadvantage, or go for the quick or lucky kill with a fast ender.