Racing Systems – Overview

In our last article we introduced the concept of racing systems, how they can be used as a terrific starting point for your overall betting strategy and the steps to developing your own systems.

The term racing system or selection system is loosely used to describe a set of rules a punter might follow to determine their bets. When the word ‘system’ is involved it typically refers to a fixed set of rules applied, where every horse that meets all of the rules is considered a bet and failure to meet even one of the rules makes the horse a ‘no bet.’

The concept of selection systems has been around for as long as racing itself and the idea of being able to solve the betting game through a set of mechanical selection rules continues to fascinate and consume punters all over the world.

While the idea that all of your betting dreams can come true by finding a magical set of fixed rules is both naïve and detrimental to your long-term success, the concept of using racing systems as the starting point in your selection process has good merit.

All across the world in business, sport and other competitive pursuits, people are using technology and new ideas to become better informed, more data driven and applying that insight to create more efficient and effective winning strategies. Racing and betting is no different.

Racing systems are about using historical research to find weaknesses in the betting market and exploit them. Weaknesses are basically scenarios where the betting market has historically undervalued horses with certain characteristics, offering a slightly longer price than they should, which helps create the opportunity to profit.

Once you have identified some of those factors, I certainly don’t endorse the traditional idea of racing systems which involves taking them as fixed rules and the only factors that determine whether you bet or not.

Regardless of the technical effectiveness of those rules, success as a punter relies just as heavily on your own confidence and overall psychology. For that reason you must take control of your own final betting decisions… it’s absolutely essential to achieving long-term success.

• Increase your profit by weeding out those that will under-perform relative to the overall group (e.g. horses that will settle back in the field or look clearly unsuited by other conditions of the race.)
• Pass on those horses that you simply don’t feel comfortable backing for whatever reason. The benefit of that to your overall state of mind as a punter and therefore ability to achieve long-term success should not be underestimated.

I call this approach Targeted Bet Selection because as a punter, you know the type of horses you are looking to back (or lay) each day and your effort is focused on finding those horses. You are using a combination of well researched fixed principles to create a list of potential bets and then value added decision making filters to maximise your result.

That’s notably different to the more traditional approach we have been conditioned to follow where we study a meeting and then back the horses we “like” the most. The reasons that many punters like a horse enough to bet make perfect sense at the time, but the problem is that those reasons are not always consistent and more importantly, they are often not related to factors consistent with the market offering a value price.  For example, you may think and inside barrier makes the horse a more appealing betting prospect, but the reality is the market tends to overestimate inside barriers and they are typically less profitable.

Remember, successful betting isn’t just about finding horses that win or lose, it’s about finding horses that are offered at a price sufficiently greater than their real winning chance (for win betting) or sufficiently less than their real winning chance for lay betting… that’s the only way you can make a profit.

Using racing systems in an intelligent way as I’ve outlined above can help you to

• Focus on a group of horses that history says have positive betting traits in their favour, increasing your chances of making a profit (more on this later).
• Have clarity about the type of horses you want to back and create consistency in your selection process.
• Save you hours of manual form analysis, which in the end doesn’t guarantee that you will find a better group of horses.

The downside of course is that you may look past horses that are otherwise good value bets, because they don’t fit your principles. However it’s important to recognise that no single punter finds all of the value bets that exist in racing. Every winning punter I know either deliberately or instinctively understands the types of horses and betting scenarios they can make a profit from and structures their work to find them. If a race or horse doesn’t fit, they move on to the next race.

If your principles are broad enough there will always be ample opportunities each week to sustain a successful betting career. Don’t become obsessed with the need to be the smartest on every race and concerned that whatever your chosen approach, you might be missing out on profitable bets elsewhere.

There are four fundamental principles to creating a genuinely useful selection system: A good selection system should

1. Be based on an overall theme or topic that has logic as the basis of a profitable advantage.

2. Have as few rules as possible (the more rules it has the more likely it is back-fitted to past results and won’t continue to produce the same results in the future.)

3. Provide a strike rate that is practical for betting purposes. E.g. it doesn’t matter how profitable your rules, a strike rate of 15% will result in massive variance and losing runs that even the most persistent punter could not endure. I would suggest a 25% SR for win betting is the minimum.

4. Be developed with a rigorous and credible testing process to ensure the integrity of results and confidence that they can continue.

The key principle is obviously the testing process. Whatever ideas you come up with, you need to be able to test them to see if they offer a list of horses that can actually give you a head start in beating the market and making a profit.

In decades gone by punters would achieve this by manually keeping records of horses that met certain rules and then checking the results. It was an arduous task that could stretch on for months and most times would end in disappointment.

Organisations that sell racing systems with the implied promise of making big betting profits have misguided punters about what is realistic when it comes to the achievable profit from racing.  Profit on turnover figures of 50% + are not realistic for any system or betting method. A system may be constructed that shows this level of profit in the past, but the rules have simply been back fitted to suit the results and will not continue in the future over a meaningful number of bets.

Aim for a system that achieves anywhere from around a break even result (i.e. -1% to +1%) to a modest profit of approximately +5% POT from a good volume of bets. This is much more realistic and will greatly increase your chance of developing principles that can reproduce the results into the future. If you develop something that is genuinely better, then that’s great, but starting with more modest expectations will keep you on the right track.

Don’t forget that a system with a small but very reliable profit can be turned into a big profit with your own personal decision making on the contenders each race day.

If you are testing over historical data that already exists, separate it into two groups:

1. “Seen” data: this is the data that you will test your selection system over as you develop the rules.
2. “Unseen” data: This data won’t be included in the analysis when you create your system. Instead, once your system is finalised, you will run it over the “unseen” data to get an unbiased measure of how effective it could be.

For example, if you have 2 years historical data you can make 1 year’s worth your “seen” data and the other year the “unseen” data. If you have 3 or more years you can choose to develop your system over 2 years and create a separate 1 year of unseen data.

Consider a theme or angle on which to base your system. Successful systems are born from clever thinking about profitable angles that can be exploited in the betting market.

Here are a few examples

Once you have developed a theme, you need to consider a ‘prime’ rule consistent with that theme.

For example, if you wanted to explore last start winners, then the prime rule would naturally be horses with a last start finishing position of 1.

With any selection system there is a prime rule and then other factors that clearly relate or don’t relate to the theme as a whole. For example, horses having their first start could be considered as irrelevant to nearly all system themes, except those that focus on first starters.

If your system theme is focused on last start winners (prime rule) then it is most likely that horses first up from a spell don’t relate to that and should be excluded.

You should also think about the RACE rules that are relevant to your theme. Consider that a 3200m Open Hcp race may be suitable for one theme, where as a 1000m 2YO Set Weight race may not. You may also determine that races with too many first starters or too many horses resuming from a spell should not be included.

What is the strike rate and profitability of the horses that have qualified under your rules? Does it show a betting return in the small loss to small profit range (or better)? Or do the results show that there really appears to be no advantage at all?

Sample sizes are certainly important when it comes to testing ideas. The larger the sample size the better. Very small samples of 50,100 or 200 can be very misleading due to the nature of statistical variance.

If your results show some promise then you can consider if there are any sensible and logical refinements that could add further value.

This is where most people get off track with the process of creating a useful selection system. They start to look at horses that have won / lost in the past and specifically adjust the rules to fit them. The desire to make your results look as good as possible results in illogical rules being added. This will become evident when you test it over “unseen” data.

Any rules that refine your primary theme should have some logic.

For example, if your system focuses on last start winners then a logical refinement to consider is whether horses that won their last start in a similar class of race perform better or worse than the average? If you have ratings intelligence in your data then you could consider whether those rated high up in today’s race have a better record than last start winners that rate lower.

This is the big test for any selection system. You have developed your rules by looking at the results of your analysis over a period of time and then used logical principles to refine them and the performance.

With the same set of rules, now test your system over the period of “unseen” data you allocated in step 1. How do the results look? Check both the strike rate and profit performance.

If you’re system has merit then the results should be reasonably comparable in both sets at data or at least still prove it offers an advantage you can work with. Hard and fast rules are a little hard to apply, you need to use a bit of your own judgement as to how useful the second set of results are.

If your testing showed an initial positive result and your testing over “unseen” data produces a big negative result, then it almost always means you need to find a new idea (assuming both sets of data have a good sample size.)

If you determine that your second set of tests have produced a suitable result, then you can give serious consideration to putting that system into play and using it as one way to come with your potential bets on a given day.

In our next article we will give you a head start on the above steps by providing some specific examples of systems you can actually use.