Chocolate and Nobel Prizes
I’ve taken this headline from a tongue in cheek article on the BBC web site.
This article reminded me of conversation I had with somebody recently about their ‘success’ in gambling markets. They came up with a thought recently, put it to use in the market and were walking away with hundreds a day. They came to me to ask if this was good. It was an outright gambling strategy, so I guess they also wanted to show me that you could make money without necessarily trading. Which of course you can, just not as effectively.
I asked them what they were doing and what their selection criteria was. Necessarily they didn’t confirm this, probably nervous at giving away a strategy, but they did mention is was a back only win strategy. This is where the chocolate comes in.
Seeing a relationship very often doesn’t mean that one actually exists, such as this previous and rather tenuous link to Chocolate in the BBC article. You could do a survey on your own and conclude something like ‘Eating banana’s dramatically increases your risk of cancer’. But if you accidentally included a high proportion of smokers in your study, you may have overlooked some less obvious, but critical information, on the headline data.
My advice to the backer was to (a) Get a much bigger sample, this evens out any oddities and (b) To have a control sample in place. I.e. know what would have happened at random.
The reason for this advice was experience, but also because his strategy resonated with me. At the time of contact I was actually testing a back only strategy over the same period. I had noted some important aberrations in the control sample. When you looked at a sample, the market was going in waves. So in one period, below the central line, laying something at front of the field would produce generally good results and above the line backing would do the same. So it turns out you could actually deploy any strategy based on any selection criteria and quite easily be fooled into thinking you were doing well, if you caught the market in a good mood. Catch a period above or below the line or in one of the trends and it really would feel you had cracked it. Compare to the control sample though and you would realise you hadn’t. In fact, all it meant for our backer was that his banana eaters just happened to be heavy smokers, or the chocolate eaters geniuses.
When you have been trying for ages to crack something it’s very tempting to think you have, but true success comes from proving that you have. I spend a lot of the last decade doing exactly that. I think it’s a good policy; I recommend you follow it.
Category: Trading strategies
