Speed Ratings

The idea behind speed ratings is to assign objective performance figures to every horse that competes in every race that you assess. Because they are objective, they are immune to the psychological pressures that can see us over- or under-rate certain performances (e.g. when a hyped horse wins compared to one that hasn’t been rated highly in conversation by others).

You can view speed ratings I have produced for all horses that have run in all races at Sha Tin and Happy Valley in the 2019/20 Hong Kong season by clicking here and downloading a fully searchable database.

Speed ratings, in their simplest form, compare the time posted by each horse with a standard time (the time you would expect each horse to run based on the class and distance of the race). These standard times should be compiled from as much recent data as can be found. It is imperative to calculate a going allowance based on all races run at each meeting, which provides a measure of the extent to which the going was speeding up (firmer) or slowing horses down (softer) on the day. Once you have that information, it is quite straightforward to assign a rating to each horse following some simple calculations.

Another great thing about speed ratings is that you can make direct comparisons between the figures produced by horses running over completely different distances, at different racks, on different going, and in different classes. In other words, you can be confident that a horse that achieves a speed rating of, say, 85 under one set of conditions has run to the same level as another one that returns an 85 over totally different conditions.

There is one major problem with using unadulterated speed ratings: they rely entirely on race times being accurate in reflecting each horse’s performance in a race. But we know this isn’t true. Horses don’t always distribute their energy efficiently throughout a race. Some will run too fast too early and run out of gas late on, while others will do the opposite. The solution is to apply a correction to each speed rating based on how efficiently the horse in question ran its race. This is easier done than you might think.

How do you run a “fast” time? Par finishing speeds, sectional-adjusted speed ratings, and visually adjusted ratings

Firstly, you need to calculate “par” sectionals for different segments of a race. As before, you need to analyse as much recent data as can be found for each course and distance. The basic premise requires you to calculate a finishing speed over a set segment of the race (I find the final 400m / 2f is the best option most consistently). This finishing speed is expressed as a percentage when compared to the rest of the race. Imagine a 1600m / 1m race. You would need to compare the time taken over the final 400m / 2f with the time taken over the whole race and then see whether the horse in question was finishing faster (above 100%) or slower (below 100%) than was average for the whole race.

If you have access to enough recent data it is relatively easy to calculate what a “par” finishing speed is for each course and distance, and if you home in on only the races in your sample that were run faster than expected and take an average of the par finishing speeds that led to those fast times, you can be fairly confident that this par finishing speed is close to the ideal way in which to finish a race to run a fast time.

With that information in hand, you can compare the actual finishing speeds with the par finishing speeds needed to run a fast time, and then perform a few additional calculations to upgrade each horse’s speed ratings based on how far away from the ideal par finishing speed it ran. Of course, a horse that was close to the ideal par finishing speed would receive a much smaller upgrade than one that finished its race a long way away from the ideal par finishing speed.

I call these adjusted ratings Sectional Adjusted Speed Ratings (SASRs). They take already very useful information conveyed by my speed ratings to the next level. It is not uncommon to rate horses finishing behind the winner or placed horses in a given race higher than their conquerors based on these data, and that can be a golden piece of information from a betting point of view, as they often pinpoint horses that have achieved far more than is obvious to a casual form book reader or even someone who has produced his/her own speed ratings.

I use a very similar method to that explained in great detail here in an excellent resource produced by Timeform. I highly encourage interested people to read that guide.

To increase the accuracy of my ratings still further, I upgrade certain runners that suffered significant interference. I rewatch each race multiple times (from side-on and head views) and decide, as closely as possible, how much time a horse has lost due to interference, before upgrading its SASR the appropriate amount to take that into account. This final rating I abbreviate to VAR (visually adjusted rating).

This isn’t an exact science, but it does improve my ratings a little bit. For absolute clarity, I only upgrade horses if I am certain that any interference they have suffered affected their finishing times, and try to make as few adjustments as possible; there is always going to be some interference in every race and I am not going to worry about every little bump — only the major ones.

Limitations of SASRs and VARs, and BESTCAP4 ratings

We all know that most horses have preferences when it comes to going, distance, running style, jockey etc. So, when conditions change from one race to the next, it would be illogical to expect a horse to run to exactly the same level as it did last time. In most cases though, traditional form book study should help you decide roughly what each horse’s preferred conditions are. As an example, if a horse that has previously shown it performs very well on soft ground has been producing SASRs or VARs just 2 or 3 points below others in a race in its last few runs on quick ground, I wouldn’t be at all surprised to see it run a higher figure if it now gets deep ground.

Additionally, another key limitation is in the upgrading of horses beaten a long way. I am always wary of very high SASRs for horses that were very well beaten. These horses typically ran much too fast early on and staggered over the final 400m / 2f, returning finishing speeds that deviate by a long way from the ideal par. The objective upgrades they receive can be inflated as they may not take into proper account that a horse running so fast early on would not possibly be able to sustain the pace. In other words, these horses are getting upgraded despite them never having a hope of running close to a par finishing speed based on their early exertions. Many of these horses are habitually headstrong types and may repeat the dose next time.

With this in mind, I set a cap of 4 points as the maximum upgrade I give any horse I assess, and call these “final” ratingsΒ BESTCAP4Β ratings. These are the ratings I typically use when comparing horses in upcoming races. So, if a horse staggered home in last and achieved a speed rating of 80, its SASR upgrade may suggest it be rated 95. That is plainly ridiculous and it would only receive a BESTCAP4 (final) rating of 84 from me.

Using the ratings to select value bets

Finally, as always, I am interested in seeking value in my bets. As such, the odds are all-important when compared to the recent BESTCAP4 ratings I have for each horse in any given race.

In an ideal scenario, I am looking for a horse that has the highest or near-highest figures in recent runs and which is a big price. That’s asking a lot but it happens occasionally. I will also find plenty of bets through other means though. For example, I might find a horse at a big price that has been producing BESTCAP4 ratings quite close to those much shorter in the market and it now gets to run under its preferred conditions, or one that is fairly short in the market but that has produced figures that stand out in comparison to its opposition today.

I hope you find my speed ratings useful and interesting. For case-study examples of how I use them to select value bets and eye-catchers to follow, see here.