The Horseracing Bettors Forum has undertaken a study into Starting Prices in Britain in 2017, measuring the “over-rounds” compared to what might be expected given a race’s field size.
“Over-rounds” are the bookmakers’ theoretical mark-ups and are derived from the sum of the amount that would need to be staked to return a fixed figure at each of the prevailing prices.
For instance, you would need to stake £20 at 4/1 to return £100 (your £80 winnings plus your £20 stake) but just £10 at 9/1. If you add up all of these stakes it indicates the amount which would need to be invested in proportion to the odds of all of the runners in a race to return £100 regardless of the outcome.
In nearly all instances, this figure will be larger than the £100 returned. You might need to stake £110 to guarantee a return of £100 come what may, or you might need to stake £120 to do the same, depending on how much the prices are in the bookmakers’ favour.
The difference between the amount staked and the amount returned – 10% or 20% perhaps – is the theoretical “over-round” which works in the bookmakers’ favour (though at this stage it must be acknowledged that a bookmaker is unlikely to lay every horse in strict proportion to that horse’s odds).
Over-round is affected by field-size, but not in a simple, linear, fashion. That is why a measure other than over-round-per-runner needs to be used.
HBF used an equation derived from a study into Starting Prices in 2014 by HBF member Simon Rowlands (https://www.timeform.com/horse-racing/features/rowley/Rowleyfile_Investigates_The_Industry_Starting_Price,_part_one) and then established the degree to which actual Starting Price over-rounds differed from that simple prediction.
The following are some of the findings.
- Over-rounds were generally higher than those predicted by the 2014 model, suggesting that %s have crept up in real terms
- The WORST courses for punters in terms of over-rounds compared to the model’s expectation are (in order): Ffos Las; Chelmsford City; and Towcester
- The BEST courses for punters in terms of over-rounds are (in order): Nottingham; Catterick Bridge; and Thirsk
- The smallest over-rounds (compared to the model) come in Chases, the biggest, perhaps unsurprisingly, in National Hunt Flat Races
- Again, perhaps unsurprisingly (given that more is known of the horses and that odds can be set with greater confidence), handicaps have smaller over-rounds compared to non-handicaps
Summary tables can be found at the bottom of this piece, but it is worth just dwelling on the figures by individual course. In particular, the over-rounds at Ffos Las in 2017 were frequently out of kilter with what might be expected and merit investigation.
These include a card at Ffos Las on 25th August, 2017, when the following betting percentages were returned: 122% for a 6-runner handicap; 128% for an 8-runner handicap; 142% for a 10-runner non-handicap; 130% for a 7-runner handicap; 128% for a 7-runner handicap; 136% for a 7-runner handicap; and 131% for a 5-runner handicap, all of which figures greatly exceed what might usually be expected.
The returns for the card on 17th July, 2017, were not quite so eye-watering, but they included: 135% for a 13-runner handicap; 133% for a 12-runner handicap; and 132% for an 11-runner handicap.
Something odd is going on at the South-Wales course, with nearly all races disadvantageous to punters betting at Starting Prices compared to expectation, and sometimes by large amounts.
Ffos Las may be a course to steer clear of until an explanation – not to mention a solution – has been provided.
Summary of methodology: races in which the same number of declared horses as runners were used to avoid skewing by late withdrawals; the average difference from “par” was taken for all qualifying races at a course; “par” was taken to be [100+(field size*((-0.0245*field size)+1.9286)]
In the below table, the figure given is the over round compared to par, with a positive value an excess (i.e. poorer value betting track than par) and a negative value a better value betting track than par.