It works like so.
The game revolves around GAMBLES.
Each GAMBLE has a RISK, and a REWARD.
Every player has a unique GAMBLE, specific to them; this is their role PM. (The game as I envisioned it would be a closed, thus, these would not be known in advance.)
However, there would also be a list of generic GAMBLES that every player, at specific times, would be able to take.
Said GAMBLES are not inherently one time use only; a player can take some of the same GAMBLE multiple times, and a non-player-specific GAMBLE could be done by more than one player.
GAMBLES are submitted privately.
Each GAMBLE has three possible outcomes:
- Obtained the REWARD, without suffering from the penalty of the RISK. (Total success.)
- Obtained the REWARD, but sufferer the penalty of the RISK. (Partial success.)
- Fail to obtain the REWARD, and suffer the penalty of the RISK. (Total failure.)
Inversely, every time you don't incur a RISK PENALTY, your chance of taking one increases, AND the penalty of the RISK is increased; every time you take a RISK PENALTY, your chance of incurring another lowers, and it'll be less severe when you take it.
So, say that you had a base REWARD of having one extra Vote, with a base RISK of taking one less vote to lynch.
Say that the risk/reward chance for that GAMBLE starts at 50/50.
If a player has succeeded in obtaining a REWARD once, then the REWARD of this GAMBLE would be having Two Extra Votes. However, the success chance would be lower, say, at 25%.
If a player has taken a GAMBLE, but not incurred its RISK PENALTY, then the RISK PENALTY would increase to being taking two less votes to lynch, and the chance to incur the penalty would increase, say, to 75%.
If a player has taken a RISK PENALTY, then this PENALTY would be reduced. In this case, it's already at the minimum, but the RISK of incurring it is lowered to, say, 25%.
You can similarly imagine the inverse GAMBLE:
Say, base REWARD of taking one more Vote to lynch, yet a base RISK of having one less Vote at your disposal. Both at 50/50 to start.
Succeeding at a prior GAMBLE would increase the REWARD; taking two more votes to lynch. Two prior successes, Three more Votes to lynch.
But the chance of success would drop accordingly to, say, 25/12%.
Taking a prior GAMBLE without incurring the PENALTY, RISK increases to 2 less Votes at disposal; 3 less, if having avoided two prior PENALTIES. The chance of incurring the penalty would increase, to, say, 75/88%.
Incur 1/2 prior PENALTIES, and while the PENALTY in this case is at the minimum, the chance of it being inflicted is lowered, to, say, 25/12%.
In other words, the worse you suffer, the lower the chance of suffering worse, and the less you will suffer; the less you suffer, the greater the chance you will, and the worse the suffering will be; the more that you succeed, the greater the future rewards are but the chances of obtaining the reward lowers; the less you succeed, the greater your chances of future success are, but the lower the reward is. (To a certain minimum level.)
Like I said, I feel like there's potential in the mechanic, but this is as far as I could refine it.