Understanding Speed Performance Ratings

Breaking down SPR

Whilst the following individual factors aren’t in isolation anything new, combined they hold the key towards a better understanding of the value of knowing what a SPR is, so that you can more accurately peg a horse’s ability.

SPR in isolation are:

• not designed to reliably predict the likely outcome of a race; and are
• not designed as an indicator of the likely market prices/rank.

It is important to note that in some races, the predictive SPR will, for the most part be inaccurate when measured against market expectations and/or actual results.

These inaccuracies can be either positive or negative and it is for this reason, that SPR should not be taken at face value or considered in isolation as an indicator of the likely outcome of a race.

Rather, with a thorough understanding of the components of SPR, they can be considered as a unique method of comparing race horses, many of which have not previously competed against each other or raced at the same tracks and/or distances.

SPR fall into two categories:

• Historical (assigned to each historical run); or
• Predictive (an SPR derived from a historical run using a unique algorithm that is then used to compare against other competitors)

Historical SPR

The race winner’s time is adjusted using a complex algorithm for a range of factors including track condition and other proprietary factors before being compared with a ‘par’ time for that track and distance. At R2W we maintain PAR times using a dynamic approach whereas most speed rating analysts/ratings providers have a set of PAR times that are fixed and are only updated periodically, this can often extend to long periods of time.

The SPR calculated for each historical run is a numerical value that is firstly assigned to the winner of each race. For the purposes of this article and for Axis software users, consider 2 rating points as one length.

For those individuals that create, or for organizations that supply ratings, that are derived from the standard publicly available form (other than R2W), a Rating of 106 for a race winner could indicate any one of the following:

• the winner ran a time that (when adjusted for track speed) was equal to or faster than the par time;
• the official time was incorrect;
• the distance of the race was shorter than the official distance (this can often occur when the moveable rail is adjusted); or
• the prevailing conditions caused faster times (e.g. wind factors, track surface, lane bias etc.)

Conversely, a Rating of 94 for a race winner could indicate any one of the following:

• the winner ran only as fast as it needed to (to win the race)
• the winner ran a slow time (relative to ‘par’ and/or other races on the day);
• the official time was incorrect;
• the distance of the race was longer than the official distance (again this can often occur when the moveable rail is adjusted); or
• the prevailing conditions caused slow times (e.g. wind factors, track surface, lane bias etc.)

In summary the R2W Axis SPR assigned to each historical run can be considered an accurate and comparative measurement of how fast a horse ran (relative to par) on that occasion after track speed and other necessary adjustments have been made.

SPR Scale

Fast performances = high SPR

Slow performances = low SPR

The SPR for all other runners in a race (other than the winner) is derived from the rating assigned to the race winner before being downwardly adjusted for beaten margins.

Logically because horses run at a faster speed over 1000m than they do at 2400m, the incremental adjustment is less as the distance increases. Again I suggest, for the purposes of explanation, that you can assume an adjustment of 2 rating points as being equal to 1 length.

E.g. If a race winner gets an SPR of 100, a horse beaten 2 lengths would get an SPR of ~96

Automated Predictive SPR

Within Axis, a complex algorithm selects a historical SPR that is then assigned as a run that is representative of each horse’s current ability and likely fitness level. That run is called the qualifying run (QR).

Having selected an appropriate QR, the SPR algorithm then assigns that numerical value to each runner as a Base SPR. This is then interrogated by the R2W computer before determining whether any adjustment is warranted. Once this process is complete the predictive SPR is displayed in the field view of Axis. Importantly, to our knowledge R2W SPR have a higher strike rate than any other publicly available speed based rating.

Caution needs to be applied to races that are run in fast times

As has been explained above, races run in fast times (relative to par times) produce high ratings. However, it does not always follow that all horses that run a fast time, and therefore generating a high SPR will subsequently ‘run up’ to that performance at their next race start. However, it is true that horses that can run genuinely fast times, particularly in good class races, have more ability than other horses they may race against in future.

Why R2W SPR are more accurate than other commercially available ratings?

©ratings2win Pty Ltd 2016. Printed under licence to Betfair Pty Ltd. This document is for personal use only and must not be reproduced. This information is supplied “as is” and is to be used for general guidance only. Use of this information is entirely at your own risk. Neither Betfair nor ratings2win make any representations or warranties of any kind, including but not limited to, the accuracy, quality or completeness of the information contained herein. All intellectual property is owned by ratings2win Pty Ltd. Please gamble responsibly. For more information, visit Betfair’s Responsibility Gambling Policy.