I think average winnings does not determine a unique risk of ruin, some more information about the distribution of winnings is needed.