What is this site about?

Our basic philosophy is that different horses are suited by different distance and surfaces, and that there are a number of reasons for this. There is the simple biological fact that horses have different sizes, shapes, conformations. They are also trained differently. Some horses respond better to a jockey's restraint than others. Many horses prefer to run in the middle of the pack because evolution has taught the species that it increases the chance of survival. However, some horses still prefer to run in front of or behind their equine friends. There is also the issue of competitiveness: some horses will only try their best when they are engaged in battle, others give up the moment they are eyeballed.

Many of these factors can be distilled into one concept: a horse's running style. Whether you believe that this running style is a consequence of genetics, evolution, human handling or even chance, the evidence reveals that a large majority of horses show a predictable pattern in how they run a race over a particular distance or surface. Some will run very quickly at the start, but fade later on, especially if they are challenged. Others only get interested in the last few furlongs and do their best running then. Others are one-paced and just keep plugging on mid-pack, race after race.

We firmly believe that most horses run a majority of their races according to a recurring pattern, and the way we model this pattern is by means of predictive ratings. These predictive ratings predict how fast a horse can run over the full distance of the race, but also how fast it run in each section of the race (early/middle/late). On latekick.com, the predictive ratings are displayed in the top table of the page. If you are using the free Predictive Ratings page, this is the only table on the page that you see. If you are a registered user or a subscriber, and you're using the past performances page, you will see the predictive ratings in the top table. (The table below that contains each runner's past performances.)

Each column in the Predictive Ratings table contains a predictive rating that represents a different part of the race. This is a short summary of these predictive ratings.

• Early – predicts the horse's speed over the first 2 furlongs.
• Middle – predicts the horse's speed over the middle 2 furlongs in a sprint or over the middle 4 furlongs in a route.
• Late – predicts the horse's speed over the last 2 furlongs
• Speed – is a weighted average of the first three ratings and predicts the speed of the horse for the full distance of the race.
• Form – similarly to Speed, this measures the speed for the full race, but recent performances count for more in the calculations.
• Pace – measures the horse's tendency to lead, and is based on the early running style that the horse has displayed in its previous five races. Technically, the calculation of this number has nothing to do with speed, although horses with higher Pace will usually have higher Early.

• If horse X has a higher Speed rating than horse Y, X is likely to finish the race in front of the Y.
• If X has a higher Late rating than Y, horse X can be expected to run a stronger last two furlongs than Y.
• If X has a significantly higher Speed than Recent, it means that X has run one or more poor races recently.
• If X has a higher Pace than Y, then X can be expected to break quicker than the Y. And if X has Early which is not lower than Y's Early, it is likely that X will be in front of the Y after 2 furlongs.

If you want to convert these numbers into lengths, we suggest the following guidelines. The optimal conversion rates differ from track to track, but the numbers in the table below provide an acceptable rule-of-thumb across tracks.

It is important to stress that these are predictive ratings. They predict how quick a horse will go today or (in the case of Pace) where the horse will be placed during the race, and these predictions are based on the horse's previous races. Suppose that DRF provided the average Beyer Speed Figure that a horse usually runs. This would (in concept) be comparable to what we call Speed. The difference is that Beyer Speed Ratings are based on finishing times, whereas we break up the races into three pieces, because a lot can be learned from analyzing a race this way. We will expand on this below.

How are these numbers calculated?

All ratings are computer-generated. The scale of American racing is huge, and calculating predictive ratings for each horse in each race requires a lot of automation. For this purpose, we have built a reliable framework that deals with all calculations.

Each predictive rating is calculated according to a proprietary formula, which takes many factors into account. The formula is based on many factors, including the traditional ones such as raw running time, weight, age, daily track variant. The pace rating formula mainly relies on a horse's position during its previous races.

How good are these numbers?

They are very good. Just blindly picking the top rated horse (using Speed) will give you 3 winners for every 10 bets on average, in other words the success rate is 30%. This is not too far off the performance of the betting public as a whole, who have a success rate of 33% with their top selection (= the favourite).

So they are very good, but they are not magic. They can be used as a stand-alone handicapping tool, but we would advise against this. This may not sound like a good salemanship, but it is important to get this message across. Take the following example.

A horse entered today has only run once before, against a group of untalented maidens. The horse got easy fractions and ran on to a seemingly big win. The predictive ratings will be based on this race alone. But how accurate can any prediction be, given that it will be based on the easiest race that the horse will ever run? The honest answer is: not very. The circumstances of the first race will not occur again today, and while there is no doubt that you can still learn a lot about how strong the horse is, you will need to know that the predicted ratings can never be as reliable as they are for more experienced horses.

The only way to address this issue is by digging a little deeper in the past performances of each race, which are available on this site as a subscription-based service. These Latekick.com past performances contain vital data about each horse's previous races such as our proprietary speed ratings, and are recommended for the users who prefer to study their races in a bit more detail. A separate user guide is available here. There are free daily samples available on this page (you need to be a registered user, but registration is free and simple). Pricing plans are available here.

How do you use these numbers?

Providing these numbers is only half the job. The other half consists of using and interpreting them. The handicapper still needs to show creativity and good judgement and apply sound betting strategies. He or she still needs to identify the value in a race in order to achieve long-term profitability. But these predictive ratings will help you achieve that goal. They allow the handicapper to analyze a race in a more detailed manner because knowing when a horse does its best running, and how good that running actually is, is a crucial piece of the great handicapping puzzle. To illustrate the point, consider the following straightforward applications of the predictive ratings:

• Late speed in turf routes: We're not revealing a big secret when we say that good late speed can be of great benefit in a turf route. The Late number is great for identifying late speed in record time, often at surprisingly high odds.
• Early speed in sprints: the numbers Early and Pace can help you in more than one way here. They will point out horses who are a few lengths quicker than the rest, they can expose a short favourite as a pace horse who is unlikely to get the lead here, etc.
• Late speed in dirt routes: at the start of each calendar year, you will find many top 3yo's going two turns for the first time in their career. Highly regarded horses with a weak Late number are unlikely to retain their lofty status. They did not see out their sprint races very well and are not likely to run better at the business end when extra distance is added.

Why are we offering this for free?

We have invested a lot of time and effort in researching, programming and collecting data from different sources. We are currently at a point where we can produce these ratings reliably and are convinced that they can be a very valuable tool for both seasoned and novice handicappers.

As mentioned above, we have also added a subscriber-only 'Past Performances' page recently. Past performances allow you to look beyond the predictive ratings which, powerful as they are, do not always reflect subtle details, such as a return to a surface/trainer where the horse was previously succesful; a bad last race which was perfectly excusable; or maybe a more obvious clue, such as a suspiciously long lay-off combined with a big drop in class. You can only subscribe after you have registered. Registration is very simple and completely free. Registered users have free access to the first two races of the day at all the different tracks. The registration page can be found here. If you are already registered, you can access the free daily races here.

How can you get more information?

If you have questions, suggestions or any other feedback, do send us an email!

Happy Handicapping!!!