Simulating wealth over time while playing a biased coin toss game
Question
Suppose we have a biased coin toss game that can be played an infinite amount of times, with the following attributes:
-
The coin can return either heads or tails on a toss
-
The player wins when heads is returned
-
The coin is biased, and heads is returned with a probability of 0.6, while tails comes up with a probability of (1-0.6) = 0.4
-
If heads shows up and the player wins the turn, they will win the stake that was bet
-
(e.g. if you bet 1 dollar, you'll win 2 dollars while if you were to lose you would lose that original dollar).
Given this information:
- Write a function that can simulate n runs of this bet, returning the average payout
- Visualize the average payout across your simulations using a histogram
- Revisit your initial simulation, revising your bet amount (expressed as a function of your wealth) to suggest a roughly optimal allocation
- You can read more about deriving an optimal allocation mathematically here, but in this case just showing the sensitivity through simulation will do