There are two inherent truths in betting systems and these are:
- They produce a low ROI
- They have generic rules
In fact these are only true for betting systems built using the standard approach, which I like to call the IF AND OR approach.
IF number of runners is less than twelve AND the race type is all weather AND (the horse is the top weighted OR the second top weighted) then it becomes a selection
I’ve just made the above up as an example so please don’t use it as an approach, but as you can see the rules are made up of IF AND OR statements to narrow down the runners to the final selections.
I’m sure that this approach is very familiar to you as 99% of all available betting systems and system builders use it.
There’s actually a much better way. But we’ll look at that in another post.
Today we want to focus on this system and the inherent rule number 2…
They have generic rules
We need to start by looking into this statement. Using the IF AND OR approach of system building there is a continual danger of over fitting your betting systems to the data.
This means that we make the data set so specific and small that we find a way to make a profit from that specific set of historical data. When we put it into practice in real life we find it doesn’t work because the rules only work for the data we built it on.
In order to prevent this from happening we need to keep the rules quite broad and generic.
Race type must be All Weather
Number of runners must be twelve or less
Look for the favourite
Find the top three weighted horses
Now I’m not saying that using this approach can’t make you systems that are profitable. It absolutely can.
But… what you end up with is a very quick process to find selections where there is actually very little information about the selections chance in the race you are analyzing.
You know that overall it will make a profit but what are the chances of your selection being right in “todays” race.
Of course there may be a few rules that compare horses to each other but there are not often very many of these because it makes the over fitting we were talking about earlier much more likely
We need to look at the issue:
- We want to have a very quick process of finding selections
- We want to be make sure that if we’re betting on a horse it has a good chance “today” and isn’t just the type of horse that will make a profit on average over a year
Well, we already have the first part… a quick selection process. It’s only the second part that we’re struggling with at the moment.
But what about if you didn’t just blindly bet the selections found by your system?
What if you used them as a shortlist instead?
If you do this then you are not only going to see your strike rate jump, but you’ll also find that your profits do to.
You know that your system rules have to be inherently slightly generic to prevent over fitting. But you also know that in the long-term they find enough of the right type of horse to make a profit.
Now imagine if you took a shortlist of horses that you knew contained enough of the right type to make a profit and used your own independent analysis to trim the list further. To remove the runners that you can be pretty confident aren’t the right type.
What will happen is…
…you’ll make more profit.
So, here’s what you need to do:
- Each day use your betting system to find the selections
- Take each horse found and look to see if it looks strong enough to contender (not win) in todays race
- Any horse that doesn’t remove from the list
You will now be left with a shorter list of selections for today. Just doing this stage is going to make big difference to your results.
But if you have a bit more time then you can also…
- Take each horse on your new list and compare it to the others in the race and give it a strength rating from 1-10
- Check the odds and determine if the odds are high enough for the strength of the horse and if so how much to bet
To start with try using this approach with just one of your betting systems and from there you can expand out.