Every now and again you hear of someone who managed to win the lottery with their magic scheme, you might even hear that some mathematician has found out what numbers are most likely to be drawn. When I hear of such things I am always amazed that said mathematician seems to confuse lucky with random and not understand the difference. But then statistics is the branch of mathematics most likely to make a fool out of people. What is being demonstrated here is survivor bias – that is get enough people doing enough random things and one of them will appear to be a genius. Lotteries are particularly prone to this idiocy because people not only dont understand how numbers work but also because the sheer volume of people playing them give rise to all sorts of phenomena that ordinary people will think are somehow special or divine yet which when looked at through the lens of the law of large numbers are perfectly reasonable. As a simply example if you get enough people playing enough lotteries sooner or later someone will win one multiple times. It has nothing to do with their ability to pick special numbers or some psychic power they possess but rather it is simple a function of massive numbers of participants. Get enough people doing enough things and the improbable becomes the inevitable.
Survivor bis raises its ugly head in everything from the gullible playing games of chance to medical studies to trading. In trading is a problem when it comes to systems testing – systems testers go to great length to try and deal with any form of bias that creeps into their testing. However, whilst systems tester go to effort to remove bias the same cannot be said for the wider investment community. Let me give you two recent examples that were thrown up at me.
Every now and again I go onto to LinkedIn which is essentially Facebook for people with a job. In my feed the following image appeared.
This is apparently a reproduction of an ad from 1960 for an apartment that sold in Neutral Bay for £5800 which apparently sold for $875,000 which when adjusted generates an annual rate of return on investment of 7.7%. There are a few problems with this sort of historical long bow style of comparison such as failure to take into account inflation and the real cost of money. What does appear as a problem is the simple notion of survivor bias. As such you cannot really generate any form of conclusion or recommendation from it other than to point out that on rare occasions someone will hold an instrument for almost 60 years. This unfortunately is not the reality of investing. The same is true for stories you see of how someone bought a stock way back in the day and refused to sell it an now they are worth more money than God. It does not take into account all the people who bought said investment and then sold it. In terms of financial advice it is no better than telling someone they should have picked the right lottery numbers last weekend.