From Racecourse to Off-Course: How the ISP Changed Everything
In June 2020, as racing returned after COVID, calculating the Starting Price, SP, of UK horse racing was changed from using on-course bookmakers to 10 large online firms. This SP became known as the Industry Starting Price ISP.
I looked at the bookmakers for races from 2016 covering both the on-course and Industry starting prices.
The bookmakers’ overround is how betting companies help ensure they make a profit. They do this by adjusting the odds so the chances of all the horses winning increase to more than 100%. This extra amount is their built-in advantage.
Let’s look at a horse race with four horses:
• Horse A: Odds are 6/4 (means £1 bet wins £1.50 if it wins).
• Horse B: Odds are 5/2 (means £1 bet wins £2.50 if it wins).
• Horse C: Odds are 7/2 (means £1 bet wins £3.50 if it wins).
• Horse D: Odds are 4/1 (means £1 bet wins £4.00 if it wins).
Now we work out the “implied probability” for each horse:
• Horse A: 1 ÷ (1.5 + 1) = 40%
• Horse B: 1 ÷ (2.5 + 1) = 28.6%
• Horse C: 1 ÷ (3.5 + 1) = 22.2%
• Horse D: 1 ÷ (4.0 + 1) = 20%
The total is 40% + 28.6% + 22.2%+20% = 110.8%.
This extra 10.8% is the overround. It means the bookmaker has built in a margin to help them earn. The higher the overround, the worse the odds are for the bettors.
So as the overround goes up, the more bettors tend to lose.
The overround also varies with the number of runners in the race, so to help deal with that, we calculate the overround per horse, OPH, which is the overround for the race divided by the number of runners in that race. Other factors that affect the OPH are related to the number of runners and, for example, an odds-on favourite. However, as we are averaging the results over a month, these should not significantly affect the OPH we calculate.
The overrounds were extracted from Proform software and averaged each month, and the OPH was calculated. Races with Rule 4 deductions were not included. Here is a graph of the UK OPH monthly
This graph shows several interesting features. The on-course OPH is on the left of the graph, and the green vertical line denotes the introduction of the ISP in June 2020. The values to the right of this line are the OPH from the ISP.
Looking at the on-course OPH, it hovered around 1.85% a horse until mid-2017, when it moved upwards, peaking at 2.1% a horse in the summer of 2019. It then fell back to about 1.9% before COVID-19 arrived. There is quite a marked month-to-month variation.
The introduction of the ISP showed very low values for the OPH at its inception before stabilising at 1.7% in the summer of 2020. From then on, it steadily increased until February 2024, reaching about 1.9 %. In March 2024, it fell sharply, going back to about 1.7% during the summer of 2024 before steeply ascending back to about 1.9% in December 2024.
We can look at the same data but add the results from Ireland to those of the UK.
Interestingly, this is mirrored in the Irish OPHs, shown together in the following graph.
The on-course OPH for Ireland was higher than the UK at around 2.2% a horse; apart from a drop in 2019, it kept over 2.1% until it ended in 2020. The ISP gave an OPH for Irish racing, which was consistently higher than the UK’s and went almost parallel to it. The Irish OPH had the same sharp drop and rise through 2024 as the UK.
The OPH does this across the number of runners in the race, with smaller fields having higher OPH values than larger fields. We can see how much effect this has by making the same graph as above but only including races with 10 runners. This also avoids issues with “bad each-way” races, which can happen, for example, if you have an odds-on favourite in an eight-runner race, which inflates the OPH.
On the next page, we put the two graphs together for comparison.
The graph of 10-runner fields is more jagged as we have fewer races. However, the overall pattern of the UK and Irish results is very similar and has the same features described in the graph for all field sizes.
This means that any changes in field sizes over time haven’t caused the graph shapes we found earlier, and it allows us to generalise the results and use all field sizes. If we were to do a more detailed comparison, we would need to allow for field sizes.
Here, we can see the precipitous drop between weeks 10-14 and a bottom between weeks 22-27. These correspond to March 11th to April 8th and June 3rd to July 8th. The start of the first period is the Cheltenham Festival. The second period closely corresponds to the Euros football tournament, June 14th to July 14th.
This drop was mirrored in Ireland.
The vertical bars around each point show the variation in the OPH that month. With on-course SPs, the OPH for the UK has smaller vertical bars than does the Irish, showing there was less variation. Some of that is due to there being less racing in Ireland. However, the vertical bars became much smaller when the ISP was introduced, indicating that the OPH was less variable between races that month.
The variation in the on-course OPH is expected to be higher as they were the result of sampling different sets of bookmakers at each meeting, whereas the ISP has a much smaller group of betting firms to draw from.
The Starting Price aims to give a price derived from a range of bookmakers competing with each other. The median price is obtained from this competitive arena and used to settle bets.
The graphs above are interesting to bettors for several reasons. The gradual increase in the OPH from 2020 to 2024 showed that bookmakers were making it more difficult for bettors to win. However, in 2024, this all changed with a sharp drop in the OPH. This drop would have made it easier for bettors to win in the summer of 2024. However, the concern is that the most likely way this steep fall could occur, followed by a rebound, is if the firms that were part of the SP cooperated. There is no proof that this happened, but I can see no other explanation. No UK or Irish racing change would account for this, as field sizes and other factors are pretty stable, not dramatically changing over a few months. The fact that it affected Irish racing, too, suggests this was possibly a commercial decision between the firms concerned.
Bettors did well with this, but it does raise a red flag about the competitive element of the ISP. The drop and subsequent rise in 2024 is deserving of explanation,
The difference between the OPH for Ireland and the UK is of concern to Irish bettors, and the difference and its consistency again suggest commercial decisions may be behind this difference.
The SPRC, in their most recent report for 2023, note the rise in OPH and suggests that this could be due to many factors, including affordability checks, a drop in betting turnover on horse racing, recent media rights deals and the decline in the ability of bookmakers to cross-sell from horse racing. These factors do not explain the changes in the OPH during 2024.
One interesting comparison is whether bettors were better off regarding OPH when derived from the on-course markets. The answer to this in terms of OPH is no. The ISP over the last four years has generally produced a lower OPH, but this disappeared over 2023. However, this is at the cost of having the firms be able to adjust the OPH at will. This can be shown by the steady rise from 2020 to 2023 and the drop and rise in 2024. I can see no reason for this pattern to occur without the consent of the betting firms.
I will leave it to readers to make their deductions and ask their own questions of the various authorities involved.
I am sure the betting companies and the SPRC will be able to explain why these graphs are as they are. I’m sure bettors on Irish racing will not be happy that their OPHs are higher than those in the UK, and they would like to know why.
The OPH data is clear: something changed dramatically in 2024, and the betting public should know why. Whether it was coordination, coincidence, or commercial convenience, the patterns point to forces behind the scenes. Transparency in betting markets isn’t a luxury—it’s essential. If trust is lost, the very foundation of the SP system is at risk.