As a horse racing systems tool developer, this is a question I often get through email and see posted on forums and social media.
It is a bit of an open-ended question because horse racing systems come in lots of different shapes and sizes and it’s a little unfair to put them all under one banner.
Many are no doubt based on information that has not been properly thought out and researched while others are based on information that would not be relevant in today’s race.
I like to break horse racing systems into 4 types.
These are normally based on a casual observation or passed down (or around) by people that have not even studied the information themselves.
A good example of a horse racing myth is the saying
“Bet the Gray on a rainy day”.
There is no data to back this up and anyone who trains horses also know that Gray horses do not do well in the wet conditions (Or is that another myth!).
These systems are normally based on racing as a whole, without any thought for the type of racing, course or even surface type.
They may just be based on the runners rank or a whole set of rules which won’t include specific information about the type of racing.
These systems generally fail because the differences in racing from day-to-day can be enormous and so the chances of being able to predict the winner becomes difficult.
3. Race type.
These systems are built around a specific race type like AW (All Weather) or chase or flat racing but they don’t take into account individual courses.
If you were asking for my opinion this is where you should start when it comes to creating systems.
It’s not perfect but it does mean that you are basing your research on like for like racing.
There is no doubt a successful system will probably be more likely to come from this category or the next just simply because the research data is more relevant and could easily contain information that is relevant to today’s racing.
These systems take advantage of the specific statistics of a course and allow you to uncover information that could be missed by the normal punter.
The only downside of being this specific is that the amount of useful data that you have to base your findings on could be limited.
These are my own 4 categories and after developing systems for over 12 years (on an almost daily basis) I tend to only use 3 and 4.
Users of PR Ratings are often adding shared systems and if I see one in the first two categories I immediately look at what happens when you break them down to 3 or 4.
People also mix together Handicap and Non-handicap races which I think can sometimes be another big mistake and a cause for failure.
Here are some other reasons why systems fail.
Price (Betfair or BOG or Industry standard SP)
There is no doubt the market picks up on patterns and often if you look at the price trend of a system over several months you can easily see the market is compensating.
So you only have to draw a line into the future and you can see when the system is likely to stop being profitable.
Strike rate. (Number of wins / The number of selections.)
Horse racing systems by their very nature tend to have strike rates that go up and down but again if you look at trends you can often see that the strike rate is getting worse over time.
This often happens to systems based on Jockeys or Trainers.
We live in a country of changing conditions which seems to be getting more unpredictable every year.
Although some systems may take into consideration the “Going”, often they do not. So, when we have weather that is really unsettled, very cold or very hot these can have an effect on individual runners or racing in general.
Maybe even some Jockeys!
Draw (Flat and AW racing).
Many courses have what’s known as a “Draw Bias” which will be more or less substantial depending on the distance of the race.
There is no doubt that some horses can overcome a draw bias that is not in their favour but not all runners can.
It could be argued that is would be taken into consideration in the statistics but that is not necessarily true.
Some statistical software available use after race information for its data.
This means that it may use actual runners in the race rather than those declared on the race card.
So, you could create a system that selects 8 runner races only using this software which you would have to do using the race card because some runners can be declared Non-Runners after the race starts.
The same can be said of systems that use SP or BSP to select the runners. Both of these are not known until either just after the off of a race or at the finish…
…Which means there is no way you can accurately select a runner based on these prices.
Many statistical software solutions also remove non-runners from their database.
This may seem like a good idea but if you are only selecting one runner a day that could have been your selection…
…But you would now see an alternative which means you are not getting accurate results from the software.
How can you create a horse racing system that wins?
1. Be as specific as you can.
This means starting with the same race type (AW, Chase, Flat etc) to begin with. Then if you can, look at individual courses.
2. Constantly re-check filters to make sure they are still needed.
I see this happen a lot. A filter is selected to try to improve the profit or strike rate and then another is added and then another without checking how the new filter has affected the ones already being used.
Every time you add a new filter you should re-check the others to make sure they are still needed.
3. Use the information that you can replicate.
As already mentioned using Declared runners instead of actual, being careful of when you take your price.
PR Ratings was created specially to allow you to see how the racing looked before the start of the race. All Non-runners are shown and there are various price points throughout the day to use instead of BSP or SP.
4. Test your systems.
I know it sounds like a “no brainer” but you should check your system over weeks and even months to see if the strike rate and price hold true to what you were expecting.
5. Check prices at different times of the day.
Generally, BSP is regarded as the best price to take (Some argue BOG) but that doesn’t necessarily always hold true.
My own system “Mals Shorties” shows more profit when you take a price nearer to 1 pm.
The real question is are you going to create your own systems that work or are you going to pick from someone elses?
There is no doubt creating your own takes time, patience and testing and generally a lot of failures as you learn what and what does not work.
The alternative is to use one someone else has already created.
That's why I came up with the "Shared System" idea as part of PR Ratings.
This allows people who haven't got the time to create their own to look at systems created by members who do like to create them.
They even give you a basis for creating your own.
Can horse racing systems work?
Yes and there are plenty making profit in PR Ratings to prove it!