Betting System: Simplifying winning patterns

Social racing punters who set-out to develop and maintain a winning betting system face a common challenge. They struggle to produce and maintain the quality data which is necessary to achieve their goal. We've simplified the complex multi-factor betting system which many professional punters use, transforming a labor intensive approach which requires an army of form analysts, to a a few simple steps that take only a few hours each week.

Multi-variable form analysis

The majority of professional punters use some sort of multi-variable form analysis technique to drive their betting activity. Put simply this approach involves scoring a horse on a range of variables (e.g. weight, distance performance etc) for a given race and then using simultaneous algebra to identify what weightings should have been applied to each score for each runner, to achieve the eventual outcome of the race. As each race result is updated in the system over time, the system learns which are the optimal weightings to apply to the scores, giving the punter a formula of weightings to apply for future race assessments.

As score weightings develop over time punters apply them in the following way.

1. Score each horse for each form variable e.g. weight carried, recent form, distance form, track condition etc.
2. Apply the optimal score weightings that have proven to be successful for the given race type and conditions.
3. Obtain the raw scores and order the runners accordingly.

Professional punters will normalise the scores to produce their own betting market for the race, compare those prices to the market and then bet according to value.

William Benter is an American punter who made his fortune in Hong Kong by using a similar multi-variable form analysis approach. Benter hired a team of analysts to produce the scores and to perform the analysis required for his success.

Other punters using multi-variable analysis include Australian born Alan Woods (now deceased), Patrick Veitch from the UK and Australia's most successful punter Zeljko Ranogajec.

The success of these punters and others that use this approach is achieved through countless hours of number crunching and analysis with much of it occurring within 24 hours of race fields being published. To meet this challenge betting professionals hire teams of analysts to get the work done in a timely manner. So why would a social punter, with only a few hours to spare each week, even dream of producing meaningful multi-varied analysis?

Simplifying multi-variable form analysis

We outline an approach that reduces the social punter's workload by intelligently simplifying the professional punter's approach.

To start with the social punter reduces the scope of work by focusing on only a few selectively chosen racing types. This means less races to score, cull and analyse to arrive at a successful betting formula.

The selection process is key and requires careful consideration about which races to collect data for. You might consider tracks which have certain peculiarities which are more likely to influence the outcome of a race. For example, I've always been interested in the 1100 - 1200m distances at Rosehill on account of the challenge it poses for outside barriers with such an acute hairpin turn so close to the respective starting positions. On this basis I believe it is a good candidate for patterns to arise for certain form variables.

To ensure that I don't put all my eggs in one basket I might pick another one or two track-distance combinations or alternatively I might pick a track-condition combination using similar logic. With two or three combinations I am increasing the chance that I will find at least one scenario where a pattern between form factors emerges over time giving me a betting formula to use.

Which racing form data to collect?

Having restricted the volume of races that we have to assess, the punter needs to consider which form variables they will include into calculations. William Benter scored each horse in each race using no less than 130 different variables. That's not feasible for the social punter however it doesn't mean we have to compromise the analysis. Thinking about the race-type chosen, the social punter can be a bit more selective by considering which variables are more relevant. Taking the 'Rosehill over short distances' scenario, I would definitely include:

• Barrier position
• Early speed (first 400m sectional times)
• Jockey strike rate over shorter distance ranges
• Preferred running style (front, mid or back-marker).

I would then add the following standard variables:

• Weight (above minimum)
• Recent form
• Best form
• Distance range performance
• Track condition performance

Then consider what else I should add from the large list of remaining variables.

Once I have arrived at my list I have to score each variable for each horse in each race that I assess.

Racing form variable scores

Most of the form variables are already in a metric format on your form guide. For example weight carried above the minimum is a simple calculation. Other measures may be more subjective like recent form. Scoring recent form by using the finishing position in lengths away from the winner is a good way to ensure that you don't discount good runs finishing outside the placegetters.

Some variables might require other information sources. For example you might use speed map images to rate a runner's early speed relative to the field, or you can go a step further and acquire sectional data to perform a more accurate assessment. Whatever you do, keep it consistent and use a simple scoring approach to make it easier to identify patterns amongst successful runs. A 1 to 5 scoring scale or even a 1 to 3 scoring approach applied to each form variable is best. Use the smaller scale if you are more time poor.

Post race analysis

After each race has resulted, you need to eliminate the unsuccessful horses leaving those runners that finished close enough to the winner to be considered a success in the race. For some punters that's the first three place getters, for others it's any runner finishing within say one or two lengths of the winner. As your data grows you want to look for patterns that exist between the variables for successful runners. Combining your simple scoring method with a colour scheme will aid pattern identification between form variables. Once you've identified these patterns you've created your simple betting system formula.

Identifying the winning formula

The example below shows a clear pattern in a relatively small sample with a correlation between barrier, early speed and jockey's performance over shorter distance races at Rosehill.

* This example is for demonstration purposes only and does not necessarily reflect real scores

Punters need to look out for sample size error which is more likely to occur earlier on when the data size is small. If the trend persists as the data grows, then the punter is well on their way towards developing a betting formula. The example above highlights a formula for short distances at Rosehill where horses have very good early speed, steered by jockeys that perform well over shorter distances from barriers that have an average strike rate over the distance.

In this case there might be an opportunity to further develop the formula by thinking further. If the trend relates to barriers with an average score then intuitively it would apply to better barriers too. Extending the barrier score range to be 3 - 5 might highlight an even more significant pattern to follow.

Using the analysis above, the punter can now apply a simply formula where horses with scores of 5 for jockey performance over short distances on horses with outstanding early speed from at worst an average performing barrier at Rosehill over 1100 to 1200m are singled out for betting.

I have applied the simplified multi-variable betting system approach for the past three years now. I spend no more than three hours a week on my betting system and I have two track-distance scenarios that are working well for me. Hope it helps.

James Cleary (edited by Mike Steward)