Always Experiment

Speak to any trader and they will always emphasise the need to experiment; to continually develop new or tweak existing strategies. Stand still and you will ultimately fail, or at least, not be as effective. So many things can have an influence on your trading that failure to adapt will have a detrimental effect on your success.

I am always experimenting. Indeed, I probably spend more time playing around with crazy ideas (and some of them are downright bonkers) than I do on actually trading. Monitoring of a strategy can take such a long time, especially if it has to be done manually and cannot be assessed from historical data. This can easily lead to stopping the process way too early before enough data has been collected to make it statistically significant. One has to commit to the process to make it worthwhile.

Whether that monitoring process leads to a strategy going live is in many ways not the point. It can be just as important to eliminate a particular idea as being a long term loser. In the meantime however, that monitoring can indicate other areas that warrant further examination. Indeed, expanding the data collection to cover as many (often unimagined) bases as you can is, in my view, a vital aspect of any strategy monitoring process. The generation of other ideas even when those ideas are unrelated to the original strategy being assessed is often the beneficial consequence of doing so.

By way of example, the attached graph (blue line) illustrates the performance of a lay system I've been looking at since early January. Covering some 700 races of all types, going and field sizes on both turf and AW, the performance, while somewhat erratic, has produced a tidy profit over 6 weeks or so of some £2000 based on maximum liabilities of £50. Note that the graph represents a race by race account which explains the rather roller-coaster nature of the graph. If the same data was plotted as a daily P&L, the line would be a lot smoother.

I'm not going to detail what the strategy entails but rather draw your attention to the red line. This represents the performance if I had traded out of those lay picks pre-off. It has been scaled by a factor of 5. In other words, using a maximum liability of £250 which is a figure that is easily supported by the markets.

First thing to note is how incredibly consistent it is but I draw your attention to it because it was never my intention to monitor this as a strategy. It simply fell out of the data I was already monitoring. As I alluded to above, I made sure when investigating the original strategy that I collected as much data as possible.

This 'discovery' has proved quite exciting. Long time regular readers will be aware that pre-race trading has always presented me with difficulties. I've been so impressed with the results over 700ish races that I've now gone 'live' - initially to lower stakes - in order to confirm what I found on paper.

This wouldn't have been possible if I didn't ensure that the data I was recording wasn't limited to the original idea that led to the monitoring process in the first place. Moreover, that data has also given me other ideas for a dutching strategy which will need further investigation.

So, in conclusion, keep thinking of new ideas and strategies. Investigate your own ideas but make sure you collect enough data to be meaningful. Keep adapting, keep testing and keep positive.

No comments: