By Robert Wachtel
In earlier articles about money management, I introduced the mathematical betting-optimization tool known as the “Kelly criterion,” and demonstrated that anyone with an emotional makeup which inclines either to optimism or pessimism will find that rule almost impossible to use. Now I would like to explain why it is such a daunting task — even for normal, emotionally stable individual — to “follow Kelly.”
The Kelly criterion admonishes us to bet exactly X% of our bankroll when we have an X% advantage. No other strategy will increase our net worth faster than this one. But the objective that most normal people have in mind when they gamble is hardly that of making as much money as possible! I know that this claim sounds crazy. After all, gambling is all about the money — isn’t it?
Winning is Everything?
My experience in money backgammon games persuades me otherwise. Having observed the typical punter in action for more than thirty years, I can testify that there is something he values far more than money: winning! True, most people want to come out financially ahead at the end of the day: but the exact amount by which they do so is not crucial. It is the thrill of victory and the bragging rights that accompany it that draw them to the game.
This everyman’s desire to win at all costs is no secret. Indeed, he expresses it quite consciously and manifests it continuously in his behavior. He loves to quit when he is ahead, to hit and run, to “book a win.” And the next time that he plays, he makes a point of bringing only his profits to the table, leaving his original stake stashed away somewhere out of temptation’s reach. “I’m playing with the house’s money,” you will then hear him boast.
His attitude is entirely different when he is behind. Now, instead of “getting out of Vegas” he will continue to play for hours, even days, in a vain attempt to avoid the damage to his ego that “being a loser” would visit upon it. And if he does succeed in climbing out of the hole … well, that is something, he rationalizes, to be proud of: a feather in his cap, a testament to his grit and determination.
The Anti Kelly Backgammon Shark
The point is that this typical betting pattern is not only non-Kelly – it is positively anti-Kelly! The plain truth is that no one performs as well when losing in any competition – whether it is a test of wit, strength, or patience — as when they are winning. Even among seasoned, professional athletes, losing engenders a host of negative emotions: loss of confidence, impatience, anger, inability to focus, etc. Conversely, it is when you are ahead that your mind is at its sharpest. That is when the link between mental effort and reward has been re-established – and rather than the chore it becomes when you are being beaten, the contest then presents itself as a fascinating puzzle, one you positively enjoy solving.
Games of mixed skill and luck (like backgammon and poker) present innumerable practical examples of this principle in action. When you have experienced a run of bad results in games like this, the mathematical, “Kelly” thing to do is to quit … or at least reduce the stakes for which you are playing significantly. By the same token, it is precisely when you are winning (and, probably, your opponents are playing badly) that you should increase your bets and continue to play!
In the poker world, there is a species of shark that makes his living precisely off this glitch in the human psyche. This specialist sleeps all day, and wakes up at 1AM so that he can arrive at the poker club fresh and rested a couple of hours later. There he finds the tables full of easy pickings: exhausted, frustrated players who came in after work for a few hands of poker and have been chasing their money ever since. Once he sits down at the table, the chance that those chasers will ever get even becomes that much more prohibitive.
In the next article, I will introduce several golden rules of money management for those of you who want to gamble or play backgammon for money responsibly without neglecting the psychological defects that distinguish them from the calculated Kelly bettor.