kelly_binary

qf_lib.common.utils.miscellaneous.kelly.kelly_binary(win_probability: float, win_size: float, lose_size: float) → float[source]

Calculates the value of the Kelly Criterion (the fraction of money that should be invested) for a bet that has two possible outcomes.

NOTE: This method should not be used to estimate the kelly value for a timeseries.

Parameters:

  • win_probability (float) – probability of winning. Assumes that probability of losing is 1 - win_probability.

  • win_size (float) – gain if we win. For example: 0.7 means that we get additional 70% of what we bet. (if we bet 10$ and we win we now have 17$) new_value = old_value * (1 + win_size)

  • lose_size (float) – lose if we lose. This value should be negative. For example: -0.2 means that we lose 20% of what we bet. (if we bet 10$ and we lose we now have 8$) new_value = old_value * (1 + lose_size)

Returns:

fraction of money that should be invested

Return type:

float