WFA Analysis
Since the horse racing weight for age (WFA) scale was introduced in Britain back in the 1860's it has undergone very few changes - until recently. Modifications to the Flat scale in 2017/18 were prompted when it was realised that 3yos were distinctly advantaged at longer distances.
5f | 6f | 7f | 8-9f | 10-11f | 12-13f | 14f+ | ALL | |
3yos | 1.02 | 0.96 | 0.98 | 1.10 | 1.21 | 1.29 | 1.57 | 1.11 |
4yos | 1.10 | 1.12 | 1.07 | 1.02 | 0.92 | 0.99 | 0.87 | 1.03 |
5yos | 1.07 | 1.09 | 1.14 | 0.92 | 0.90 | 0.85 | 0.89 | 1.00 |
6yos | 0.95 | 0.93 | 0.99 | 0.95 | 0.79 | 0.70 | 0.69 | 0.89 |
7yos | 0.88 | 0.87 | 0.71 | 0.87 | 0.77 | 0.54 | 0.47 | 0.77 |
8yos | 0.76 | 0.82 | 0.75 | 0.74 | 0.63 | 0.44 | 0.67 | 0.72 |
9yos+ | 0.66 | 0.72 | 0.66 | 0.45 | 0.28 | 0.54 | 0.52 | 0.58 |
They still are. The table above shows significance* by age and distance in all-aged Flat handicaps from 01/01/18 to 22/11/24 (ie: since the modifications). It is clear that 3yos are still advantaged at longer distances and that older horses should receive weight allowances, especially at longer distances.
The recent unification of the British and Irish Jumps scales (May 2021) could be a step backwards. Here's an analysis of the current WFA scales..
This is a graphic representation of the official WFA scale for 5f Flat (above). The first observation is that there is a kink between the 2yo and 3yo scales. This has existed since the Jockey Club's genius, Admiral Rous, devised the scale in the times when there was no Flat racing between mid-November and March and horses were expected to mature at a slower rate during that period. There's Flat racing all year round now (since the advent of All-Weather racing at Lingfield in 1989). The kink exists at all Flat distances, as you will see later. The second observation is that the scale suggests horses suddenly mature at age three and three quarters. This 'sudden' maturity also exists at all Flat distances.
Here's the official Hurdles scale for 3m (above). The first observation here is that the scale suggests 4yos progress at a faster rate than 3yos. The second observation is the sudden maturity at age 5 years. Its the same at all Hurdles distances.
..and here's the official Chase scale for 3m (above). Again, older horses apparently progressing faster than the younger and suddenly maturing at age 5 and a quarter years. Same at all Chase distances.
SO WHAT SHOULD THE SCALES LOOK LIKE?
Here's DataForm's scale for 5f Flat (above). The DataForm scale is derived from a formula based on the results of a comparison between every horse's official ratings at different moments in their careers since 2002. That is; every horse's rating was compared to its best 'mature' rating. The differences were averaged and plotted, as you will see in SOURCE DATA below.
This is DataForm's scale for 3m Hurdles (above).
This is DataForm's scale for 3m Chase (above).
SOURCE DATA
The video above steps through all race-codes and distances and shows the official scale (blue), the DataForm scale (red) and the source data (black) from which the DataForm scale was derived. There was no source data for NHF. The DataForm NHF scale is estimated from Flat and Hurdles stats.
SUMMARY
In summary, horses appear to mature much later than the official scale suggests and progress at a more gradually diminishing rate.
REGRESSION
Data for older horses demonstrates how horses regress following maturity - further evidence that there should be allowances for older horses. This data was also helpful in developing the formula for plotting the curvature of both progressive DataForm WFA values and regressive.
PUDDING
Dataset | Official WFA | DataForm WFA |
GB Flat/AW | 1.44 | 1.45 |
GB Hurdles/NHF | 1.61 | 1.63 |
GB Chases | 1.27 | 1.30 |
IRE Flat/AW | 1.58 | 1.59 |
IRE Hurdles/NHF | 1.93 | 2.02 |
IRE Chases | 1.61 | 1.64 |
Above are the results from an analysis of all British and Irish races from 01/01/2012 to 31/05/2021 using the variable, LDH, which is the best of a runner's last 2 DataForm achievement ratings adjusted with WFA and weight-carried (not including jockey's allowances). The table shows the significance* of the runner with the highest LDH in each race. These results show that applying DataForm WFA to ratings increases the significance of the LDH variable in all race-codes in both Britain and Ireland. Hover over datasets for sample sizes.
*Significance (Sign) is expressed as how many times as likely a runner will win compared to the average runner (a significance of 1.00 means equally as likely to win as the average runner). This is similar to win strike-rate but accounts for field-sizes.
*Significance (Sign) is expressed as how many times as likely a runner will win compared to the average runner (a significance of 1.00 means equally as likely to win as the average runner). This is similar to win strike-rate but accounts for field-sizes.