Do the Last 3 Form Figures Really Help Find Winners?
I was asked by a reader if I could investigate the strike rate of runners based on their last three form figures and restricted to the first six in the betting forecast. As I always try to write a requested article I have done so today. There is just one change and this is that I am not restricting it to the first six in the betting forecast.
The reason for this is that all betting forecasts are different and so it would be unfair to test it on a betting forecast as yours may be different to the one I am going to use it on.
There are many ways to analyse this data and I have been asked do it by adding together the results of the last 3 finish positions and then analysing the total score. The lower the score then the better the horse has finished in its last 3 runs. Please note that I am removing any runners who haven’t had 3 runs in the current season or who have fallen, pulled up etc… In other words, if there is a letter in the last 3 form figures then I am not using this runner in my sample.
My first figures are going to be from those runners that have a total of <=9. This means that they have an equivalent minimum of 333. Of course they could also have 711 or 117 or any number of formations but they will need to have had a decent performance in their last three races to achieve this figure.
Profit |
Wins |
Losses |
Bets |
SR |
ROI |
-9040.57 |
2162 |
21612 |
23774 |
9% |
-38% |
As you can see the results are not good at all with just a 9% strike rate. We are betting an average of 1.35 runners per race, which means that the winner would be one of the selections in around 12.33% of the races but even so, with a ROI of -38% the performance is not good.
I am going to go much lower now and look for runners with a total of <=3 only, which means that if they have only been in one race they must have placed and if they have been in 3 races they must have won all of them.
Profit |
Wins |
Losses |
Bets |
SR |
ROI |
-4992.78 |
803 |
10227 |
11030 |
7% |
-45% |
This performance is even worse with fewer bets, a lower strike rate and a lower ROI. This indicates that we are not looking at this data in the right way. Let’s change our tactic slightly and be a little more focussed by looking at runners that have achieved a win in all 3 of their last races.
Profit |
Wins |
Losses |
Bets |
SR |
ROI |
-19.14 |
22 |
93 |
115 |
19% |
-17% |
We have now only got 93 bets from our 17551 race sample. As you may expect the strike has increased significantly to 19% and while we are getting a good strike rate this information is well known amongst race bettors because it is the first thing they see. This means that there is no value in the odds on offer and we are still losing 17% of every bet.
It is already clear that if we want to find a profit then we are going to have to look at scenarios that are less obvious to the majority of bettors. The most obvious example of this is to use the SP betting position. If a horse has a solid performance but isn’t in the top 3 betting positions a lot of people will automatically assume that the horse has something wrong with it.
You and I both know that this may be the case but it also may well not be the case. This situation could well provide a profitable scenario because we are finding value in the market.
Profit |
Wins |
Losses |
Bets |
SR |
ROI |
5.5 |
4 |
30 |
34 |
12% |
16% |
Value was there! Admittedly it wasn’t a huge amount but it was still there nonetheless. In 9 months we made 34 bets and gained 5.5 points in profit selecting these runners.
This shows the main problem with using the last three form figures in a standard way is that there is no value, but at the same time proves that it is possible to make a profit with nothing more than this basic information.
I would like to provide what I was originally asked for and that is a break down of the strike rates of sum of the horses last 3 finishing positions. I have also included the IV values for this breakdown so that you can see where the value lies. Please note that I calculate IV slightly differently to normal and use a winners/expected winners formula which I believe is a better way of calculating impact values.
Total |
Profit |
Wins |
Losses |
Bets |
SR |
ROI |
IV |
1 |
-149.6 |
57 |
377 |
434 |
13% |
-34% |
1.16 |
2 |
-261.82 |
109 |
583 |
692 |
16% |
-38% |
1.62 |
3 |
-419.46 |
133 |
867 |
1000 |
13% |
-42% |
1.32 |
4 |
-456.16 |
174 |
1200 |
1374 |
13% |
-33% |
1.26 |
5 – 6 |
-1455.76 |
431 |
3484 |
3915 |
11% |
-37% |
1.07 |
7 – 8 |
-1250.4 |
499 |
4292 |
4791 |
10% |
-26% |
1.01 |
9 – 10 |
-1271.75 |
497 |
4317 |
4814 |
10% |
-26% |
1.01 |
11 – 15 |
-2727.2 |
820 |
8294 |
9114 |
9% |
-30% |
0.89 |
16 – 20 |
-1705.79 |
263 |
3788 |
4051 |
6% |
-42% |
0.65 |
> 20 |
-479.21 |
46 |
847 |
893 |
5% |
-54% |
0.52 |
The above table shows that runners with a total, calculated by adding up the last three form figures, of between 2 and 4 win significantly more races than they should do.
Can you make a profit from this information? Definitely but you need to consider how you are now going to find the runners which have value.
A quick look at some of these figures when looking at the top 6 in the betting forecast only (as originally requested) indicates that they will be very similar to the above figures.
Did you know that forecast favourites (in Racing Post) win around 31% of the time. However, if the forecast favourite starts favourite at SP this win percentage increases to 37%.
I didn’t know that Stephen, interesting. I would guess that then forecast favourite starting as a live favourite will have a worse return though?
Sounds like your reader is looking at VDW.
VDW stated that if a horse had less than 3 runs than you should add on the most recent run. So If a horse had two runs and finished second in its first race and third in its last race the total would be 2+3+3 = 8. If a horse has had only one run then simply multiply its finish by 3. All finishing places over 9th count as 10 as do most F, P UR etc.
Why look at the first 6 in the forcast, well VDW claimed that 85% of the winners could be found there.
AR
It has been a long time since I read VDW thoroughly, maybe it is time to take another look. Yup that is right around 85% do come from the first 6 in the betting.