Prediction League

[Mnementh]

kuraobi said:
Hi,
I'm just submitting a bet system, donno if you'll like it but it may give you some ideas.

So let's say 5 people bet on the sales of a game.
Player 1 : bet 100 points on 5000 sales
Player 2 : bet 20 points on 3000 sales
Player 3 : bet 30 points on 6000 sales
Player 4 : bet 150 points on 8000 sales
Player 5 : bet 10 points on 14000 sales

Total pot : 310 points

The game sells 6200 copies.

Player 1 : 80,6% accuracy, he wins 0,806*100 = 80,6 shares.
Player 2 : 48,4% accuracy, he wins 0,484*20 = 9,7 shares.
Player 3 : 96,7% accuracy, he wins 0,967*30 = 29 shares.
Player 4 : 71% accuracy, he wins 0,71*150 = 106,4 shares.
Player 5 : 0% accuracy, he wins 0*10 = 0 shares.

That makes a total of 80,6+9,7+29+106,4+0 = 225,7 shares.

Player 1 gets 80,6 * 310 / 225,7 = 111 points, for a bet of 100 points and an accuracy of 80,6%.
Player 2 gets 14 points, for a bet of 20 points and an accuracy of 48,4%.
Player 3 gets 40 points, for a bet of 30 points and an accuracy of 96,7%.
Player 4 gets 147 points, for a bet of 150 points and an accuracy of 71%.
Player 5 gets 0 point, for a bet of 10 points and an accuracy of 0%.

So you can add parameters easily.
For example, if you want to give more money to people predicting earlier, you can set a multiplier for the shares they win.
If you want to give an advantage to people playing often, you can add a multiplier to the pot. For example, people bid 310 points, you give them back 341 points (10% more).

I was thinking in the same direction, but wasn't able to think about an formula/algorithm. Thanks alot for doing this, I like your suggestion.