Win Thumb Lose Thumb

A few nights ago, I was watching a show about people that have been attacked by various forms of marine life. There was a man that was stabbed by a marlin’s bill, another man that was bitten by a beaver, that kind of thing. The best segment though, was the one where a giant moray eel bit a diver’s thumb off and ate it. It seems that the diver had built a ‘relationship’ with this particular eel by feeding it sausages. Flesh colored, thumb shaped sausages. The diver answered the question of why he was surprised by the, seemingly obvious, outcome by stating, “I was shocked because there had been no problems up to this point.” Fortunately doctors were able to replace his missing thumb with one of his toes. Unfortunately, there was nothing they could do for his tragic lack of imagination.

The diver made a classic blunder. He assumed that not having body parts eaten the first few times held some significance. He confused a winning bet with a good bet, which is often not the same thing. There are four types of bets:

  •  Good bets
  • Bad bets
  • Winning bets
  • Losing bets

A good bet is one where the probability of a favorable outcome outweighs the probability of an adverse outcome. A certain percentage of the time bad bets will produce winning results just as good bets will produce losing results. Sometimes you can have a bad bet that almost always produces a winning outcome, but the magnitude of the occasional loss overwhelms the benefit of the more frequent win. This effect is illustrated by our toe-thumbed eel diver or by the Financial Products division of AIG. When the inevitable happens, these people will tell you how unlucky they were or how it was just an unpredictable confluence of events that led to their unfortunate demise. Common sense tells us that the outcome of doggedly sticking with a bad bet is anything but unpredictable. Unfortunately our wiring makes it easy to stick with that type of bad bet; the frequent rewards feel good and mask the mounting danger. The opposite is also true; a good bet that pays off infrequently is difficult to stick with.

Anyone that has coached youth baseball will tell you that a big part of the job is managing expectations. It’s hard for a young batter to understand that going 0 for 6 and doing everything right can happen at the same time. Since boys in their early teens are not widely admired for their abundance of patience and good judgment, they will often try to fix something that doesn’t need fixing. Here’s the thing though. As a coach, even though you may not agree with it, sometimes you have to indulge a player’s impulse to make something happen in order to prevent a spiral of frustration and self-doubt. People differ in their ability to remain rational in the face of adversity, it’s important to acknowledge and make allowances for that reality.

Over the years I have backtested hundreds of investment hypotheses. These are simulations where you apply different rules to historical investment data to see how a portfolio would have performed during the time frame of the test. These tests can’t tell you what will happen in the future but, conducted properly, they can be a valuable aid in developing reasonable expectations for a particular approach. One of the most interesting and counterintuitive lessons learned from this testing is how being “correct” impacts the performance of a portfolio.

By adjusting the rules that govern buying and selling decision, you can generate results that have very different rates of accuracy, otherwise known as a portfolio’s batting average. The batting average of a portfolio is expressed as a number that reflects the percentage of transactions that were ultimately sold for a profit. For example, a batting average of .650 would mean that 65% of the transactions in the simulation were profitable. One of the interesting lessons learned over all of these tests has been that a high rate of accuracy, or generating a high batting average, ended up being among the least important factors in generating a high rate of return over time. In fact, the best performing simulations often had a relatively low batting average, sometimes less than .300.

Obviously you have to sometimes be correct to make money, but the impact of the rate of accuracy is far less important than the impact of magnitude. It’s the size of your gains relative to the size of your losses that makes all the difference. It is very common to see high performing portfolios where one or two holdings out of twenty account for most or even all of the gain.

The high batting average portfolios generally don’t do as well because the easiest way to get a high batting average is to sell the winners too early. If you put money into investments with positive upward momentum and place a relatively tight trailing stop, you stand a very good chance of winning more often than losing. However, a tight protective stop is likely to protect you right out of that rare but vital big winner due to nothing more than normal volatility.  These portfolios have too much symmetry between the size of the gains and losses to generate high returns, but the frequent rewards are a comfort to the investor.

The best performing portfolios had great annual returns because they not only generated wins of large magnitude, but generated a lot of them. A high frequency of transactions, regardless of batting average, combined with large magnitude wins often translates into stellar performance.

Here’s the thing though, I can’t do it.  These high performing portfolios are accompanied by high volatility, steep drawdowns and a lot of transactions.  Conceptually, that shouldn’t be a problem for a long-term investor with high risk tolerance and low transaction fees.  In practice however, I’ve learned that my conceptual stomach for risk is stronger than my real one.  When faced with a large drawdown or long cold streak, the fact that the scope of this adversity falls within expectations is of little comfort.  At a certain point the pain of losing leads to doubt.  Doubt reveals our inner 14 yr. old baseball player and leads to an urge to fix rules that may not need fixing.  Rules that aren’t followed aren’t really rules at all.  For many people, including me, trading away some of the return expectation in return for less pain can end up being more profitable in the long run because we are more willing to stick with a bet that we have confidence in.

Simulations can be of tremendous use in developing investment concepts with a high probability of a favorable outcome, but investing is more complicated than a probability calculation.  If the characteristics of your chosen approach cause you to choose something different when confronted with an inevitable period of disappointment or frustration, it was a bad bet, regardless of the odds.  An investor is only set up for success when their approach is built on both a good bet and a realistic ability to carry it out.

Past performance is not indicative of future results.  All indices are unmanaged and investors cannot invest directly into an index. Ideas and opinions expressed in this article are the sole responsibility of Patrick Crook/PLC Asset Management and do not reflect any stated opinions of Commonwealth Financial Network, National Financial Services LLC, Member NYSE/SIPC or any other person or entity.