09 Maths and Stats / Betfair Education / 31 July 2008 / Leave a comment

In a previous article we talked about regression to the mean as a way of looking at gambling in the long term. Here we will use the theory to talk about a significant and interesting point of gambling within a population. A better view is provided with a good example in the stock market industry.

Let us suppose that you we have just read a book about some of the greatest stock market brokers in the world. In fact, the book was written by the investor, who made the most money trading stocks last year. In the book, he tells all the secrets, the no-no's, and tips to become as successful as he was.

No doubt he made a lot of money by trading shares, and is making even more now that people are reading his book. However, we have to ask ourselves, how much of it is due to chance and how much is because he actually is a really good stock broker?

To find an answer we look at an example. Suppose there are one million traders in the stock market, and suppose that every single one of them was trading blindly. In other words, they were just buying and selling stocks irrespective of whether the company was doing well or not (sounds like some gamblers)! Now naturally, there would be some winners and some losers. Over the case of the year, if the stock market has increased (which it generally does), then there will probably be more winners than losers. However, there will always be one person at the top, the person who has gained the most amount of money.

Now obviously, in this case, we know that he has gained the most amount of money just through pure luck as he, along with everyone else, was trading blindly. There is, however, a good chance that this person is selling books and making a fortune giving seminars about how to make money in the stockmarket.

The same happens with gambling. There are countless 'systems' and 'tipsters' out there who have very good records. We shall assume that there are a thousand systems out there, and that all of them bet blindly as well. Because the odds are against you with gambling (although the real professionals turn it in their favour), let's say 60% of these tipsters and systems lose and the other 40% win.

Now, no one wants to follow a losing tipster do they? Therefore, these tipsters stop their services, perhaps change to other names, new web pages or something of the sort. We then find that there are 40% of the thousand tipsters, or 400 tipsters out there who are making money, but we know that it's just due to pure luck! All of a sudden everyone starts purchasing systems or following tips of successful tipsters, who are only there because of luck, and according to the theory of regression to the mean, will most likely lose in the long term.

Notice that previous sentence ended with the words 'long term'. It is the 'long term' with which we are really concerned. How do we know if tipsters, systems, or stock brokers are successful, because they have expertise in their fields and not just luck? We decide, based on the long term.

The stockbroker, who won heaps of money last year, might have done really well that year, but how did he perform the previous years? If he was still gaining in previous years, we could say that he probably is good at what he was doing. If he lost or only just started trading, then it is possibly due to chance that he is at the top.

If we are looking to follow a tipster, a system or even a friend who tells us he is doing well at gambling, then we do not just find out how well he has done in the last month, last six months or last year but find out how well he has been doing in the last three to four years. Luck doesn't follow someone for three to four years and if someone is still making profits over this time, consistently each year, then it is obvious that they know what they are talking about.

Similarly, with your own gambling, if you're doing really well over the past three to six months, don't put all your eggs in the same basket. The regression to the mean theory will come back and bite you. Expect future loses as well as wins, and only when you have a long term successful betting history behind you, can you then increase the ante a little because you know that it was not all luck that put you in the position that you now find yourself. It requires two or more years of successful betting history before one can tell if luck has, or has not, played a part.

By Matt Elliott - co-owner of Australian based leading sports and racing portal www.puntingace.com