Next to every horse’s name on a racecard sits a short string of numbers and letters, something like 2-11P0 or 1/1/13/7/1/5. That string is the horse racing form: a compact record of how it finished in its recent runs. Once you can read it, a racecard stops looking like code.
Below, how to read it, and then how the same information looks once it is stored in a database instead of printed in a newspaper.
What the figures mean
In British and Irish racing, each digit is a finishing position. A 1 is a win, a 2 a second place, and so on up to nine. A 0 stands for a finish in tenth place or worse, because the string only has room for one character per run.
Letters replace a number when the horse did not finish:
| Letter | Meaning |
|---|---|
| P | Pulled up |
| F | Fell |
| U | Unseated rider |
| B | Brought down |
| R | Refused |
| S | Slipped up |
| D | Disqualified |
Most publications put the most recent run on the right, so 2-11P0 reads, from left to right, as an older second place, two wins, a pulled up and then a run finishing tenth or worse. Dashes or slashes usually mark a break between seasons, which tells you the horse had a rest. Conventions differ slightly between publishers, so check the key on the card you are using.
Form in Hong Kong
Hong Kong lists the last six finishing positions as numbers separated by slashes, and does not squeeze them into one digit, so you see 13 or 11 for a thirteenth or eleventh place. A real example from a recent card is 1/1/13/7/1/5. We store it as published, so check the order against the source before you rely on it, in case your tool reads the string in the opposite direction.
The Hong Kong racecard has several other fields that describe a horse’s recent history:
- the rating, and how much it changed since the last run (for example 80, up 6)
- the number of days since the last run (81 in that example)
- the horse’s best time
- the gear it will wear, including whether any of it is new
What a database stores instead
A database does not keep the string. It keeps the individual results, one row for each run, and the form string is built from them when needed. That has advantages. You can ask for the last three runs at this distance, or this course, or on this going, and get an answer a single string cannot give.
In our UK records the finishing position is held in the place column, and a non-finishing run carries a code instead of a letter. The commonest are PU (78,682 runs), F (18,364) and U (12,205), which are the same events as the P, F and U in a form string. Our result codes guide lists them all.
UK runners also carry a rating, shown as a number such as 70, 65 or 80, and a dash where the horse is unrated. A dash appears on about a quarter of runner records, so unrated is not a rare case.
Other things people read as form
Beyond the recent finishing positions, readers look at a few more.
How long since the last run: a long break can mean an injury or a planned rest, and the data tells you the gap but not the reason.
How the horse has run at this distance, on this surface and on this going. All three are fields on the race, so each is a query.
Class: whether the horse is stepping up or down against the rating it won off.
What the horse did in each race, not only where it finished. Margins and, where available, sectional times show whether a fourth place was a close run or a long way back. See margins in racing data and sectional times.
Building it yourself
Form is the oldest summary in racing and a lossy one, because a string of digits throws away margins, going and class. If you build it from the raw results instead, a rolling window that stops before the race is the safe way, and the feature engineering guide has the SQL. The results behind every form string are in the datasets and the API.
Comments are closed