Despite having already mentioned Leonard Mlodinow’s A Drunkard’s walk in a less than stellar light, we should specify that we agree wholeheartedly with his ideas.
In order to redeem ourselves, we feature an excellent article he has written in the Wall Street Journal. This one deal specifically with Joe DiMaggio’s hitting streak of 56 games, although he shows how it ties into many other factors.
The Hot Hand bias is a the logical fallacy which states that if a basketball player has hit many baskets in a row he is “on fire” and is more likely to score his next throws. This has been disproven, but it is still brought up time and time again. Mlodinow goes one step further in showing why it occurs.
The premise is simple: randomness creates clusters. If you were to ask a group of humans to spread canapes randomly on different tables, they would find a way to more or less spread these snacks evenly apart. However, if nature had decided to do so, there would be one or more clusters of hors-d’oeuvres in certain areas, while they would be missing in other areas. By the same token, if you take a group of very talented baseball players, who had a better than average chance of scoring a hit, one of them will, at some point, go on a winning streak. Mlodinow shows how this could have been created by randomness (in fact, mathematically speaking it was more likely Ty Cobb would have achieved such a record, but that just underscores the randomness).
Why do we care? Simply because this same analysis is performed for fund managers and investors. The so-called alpha is a huge factor for these funds, but if we give each manager a 50% chance of over or underperforming the market, certain managers will better the market several years in a row. These managers will then be able to command high premiums to investors in their funds, although we all know that “past performance is no indication of future results”.
We therefore propose an experiment, which is completely unethical and most likely illegal, so we propose it in an academic setting: Someone, let’s call this person Fundmaster, could find a way to open many different funds with different names. These will be high risk/high reward funds and they could operate for two or three years. Once these have passed, most funds will have performed quite poorly, but one or more will have over-performed.
At this point Fundmaster can close the funds that have not performed well, and just advertise the performance of the others, commanding a high commission from investors. This will work for one or two years, until these firms also underperform. Since, however, profits are gained through commission, Fundmaster should have been able to do quite well for himself in the meantime. Considering the fees that can be commanded, versus the expenses of openings several fund accounts, this seems entirely plausible. In fact, who is to say a version of it is not happening right now? Maybe it’s even a fund you are invested in.
