@gokulnc indeed, I tend to use the mode for basically any situation in which you need to weight a single-end bounded contiuous (or discretized for of a continuum) series.
To use your betting example since you cant bet less than 0 one might use the mode of a log-normal distribution to determine the bet amount that should receive the highest weight for consideration in an algorithm for playing a card game. since the mode of a log-normal is the peak and often the mode and mean are not at all near eachother on log-normal distribution as well as most single-end bounded distributions.