Today we have a regular weekly catch up with Malcolm Pett of http://greyhorsebot.com
In the first two articles we looked at why I like systems and how I came up with the ideas.
In this final part I want to share with you some other things I have learned along the way which may help you when designing your own.
First of all it easy to get caught up in an idea…
Often you pursue an idea and start looking for ways to make it more profitable by filtering out more and more races until you end up with something that resembles the system you want.
I call this “Stacking and Racking” filters.
The problem is, everytime you add one filter you affect another and the more you “stack and Rack” the more likely it is that you have missed a better more profitable angle.
When you add a filter you should step back and see how that one filter works on its own against the original idea.
Secondly sample size is really important.
If you do a search initially on hurdle races you could have a massive sample size which will give you a good indication of the statistics.
Sometime during your research you may see that filtering out a certain element like an age group or last time out result may appear to improve the system.
But you have to be aware that a small sample size could be misleading and may not give you a very accurate statistic.
I don’t normally use a filter unless I have s sample size of at least 100 and really prefer much higher.
Finally you may want to stick to one race type. Hurdle, Chase, AW, NHF and Flat are all different types of racing.
It isn’t very often that you find a system that works well on one race type will also work just as well on another.
Even when you do…you will often see different win/loss trends across the race types.
If you analyse them individually you will probably find filters that work better on one than they do on others.
To finish off this series of articles here are some other things you may want to consider if you decide to develop your own systems.
Be aware of trends. All weather systems are a classic when it comes to trends.
Something that appears to have worked for the last 6 months or maybe even a year suddenly doesn’t.
Many archives don’t include Non Runners. This is not a big problem until you realise that on the day you may have selected that NR and not have been aware of it not running.
I find this really important when looking at runners ranked by a Rating Systems.
If you run the ratings leaving out NR’s you will of course get different selections than if you left them in. Don’t forget you don’t always here about NR’s until after the off.
Don’t forget when you look at a system you are seeing the results based on SP or BSP and without you adding money into the market. You will only see how well it performs when you start using it.
Paper Test first always…Then small stakes. Any money into a market effects the price no matter how little. I use what I call “Price Pressure” to see what effect lowering the Average Winning Odds will have on a system.
Always be aware of your Strike Rate and Average winning odds. Look out for trends of either falling. But be aware short term trends can give false impressions.
You need to look at your system month by month to get an idea of how it performs and always create a graph. There is nothing like a visual display to show you how a system has performed.
Watch out for the “One big winner trap”. You often find a collection of data looks really profitable only to find it is actually just one or two big winners that actually created all the profit.
Finally please remember this…
I am a big believer in betting systems but I also realise that the biggest flaw with them is that you are using the past to predict the future.
The big advantage of creating your own systems is that you can tailor them to suit your own style of betting and instead of creating one. You can create many “nano” type systems that look at one particular type of racing.
Thank you as always for reading I really do appreciate it.
Find out more about the Greyhorse Bot at http://greyhorsebot.com
Today's Selection courtesy of http://bookiesenemyno1.com
7.20 Hamilton Bi Focal 3/1 generally