There’s a lot of information out there about how favourites in horse racing finish, but what about the second favourites?
We know that the majority of winners come from the first four or five in the betting. But when people are looking for betting systems they always start with the favourite.
Of course, there’s a few reasons for that. One of those reasons being that favourites win the most often and generally punters like to have more winners.
But more winners generally means less profit as you sacrifice your return for a higher strike rate.
So what about these second favourites… can we make a profit from them?
In the UK and IRE in 2015 there were 13,003 second favourite horses.
Of these 2583 won for a strike rate of 19.86%.
Not a bad starting point. The only problem is that they returned 8969 to SP and 9869 to Betfair SP. Which once you’ve taken off the 10420 losing bets means a loss of -1451 units to SP and -551 units to Betfair SP.
At a first glance that looks pretty grim.
But… that’s -11% ROI to SP and -4.23% to Betfair SP, which in fact is not a bad starting point. Particularly considering there’s no kind of analysis at all and you would have been betting an average of 36 selections a day.
Let’s look at the breakdown of second favourites by race type:
| Race Type | Wins | Bets | Profit | ROI |
| Chase Turf | 389 | 1852 | -148.10 | -8.00% |
| Flat AW | 505 | 2668 | -451.45 | -16.92% |
| Flat Turf | 989 | 4972 | -362.28 | -7.29% |
| Hunter Chase | 23 | 117 | -26.29 | -22.47% |
| Hurdle Turf | 575 | 2911 | -393.97 | -13.53% |
| NH Flat | 102 | 483 | -68.50 | -14.18% |
That provides us with some very interesting information. We can immediately see that on Flat Turf and Chase Turf second favourites perform significantly better and both have a decent quantity of bets.
Let’s focus our attentions on just those race types going forwards. Starting with Flat Turf races…
| Wins | Bets | Profit | ROI | |
| <= 7 Furlongs | 437 | 2230 | -162.05 | -7.27% |
| > 7 Furlongs and < 2 Miles | 388 | 1914 | -98.88 | -5.17% |
| >= 2 Miles | 23 | 129 | -14.83 | -11.50% |
And then looking at Chase Turf races…
| Wins | Bets | Profit | ROI | |
| >= 2 Miles and < 3 Miles | 282 | 1324 | -117.38 | -8.87% |
| >=3 Miles | 101 | 473 | -2.59 | -0.55% |
Looking at the original ROI’s and comparing, we can see that we should be focusing on Flat Turf races that are between seven furlongs and two miles and Chase Turf races that are three miles or longer.
But what happens if we look at these to Betfair SP instead of SP:
FLAT TURF
| Wins | Bets | Profit | ROI | |
| > 7 Furlongs and < 2 Miles | 388 | 1914 | 53.89 | 2.82% |
CHASE TURF
| Wins | Bets | Profit | ROI | |
| >=3 Miles | 101 | 473 | 38.48 | 8.14% |
If you’d simply bet second favourites in Flat Turf races between seven furlongs and two miles, and Chase Turf races three miles or longer at Betfair SP… you’d have made +92.37 units profit.
Oh yes, that’s after a 5% commission, and with more than 100 wins in both that is also statistically significant.
If you’re looking to find contenders then this can be a great way to start and then use these horses as your shortlist to form read further.
Of course, if you have access to other ratings then you can use these to refine your shortlist further. For example, I’ve looked at a horse’s average speed rating over the same race type and then looked at which horses had the highest (ranked 1) down to the lowest.
Here is what I found…
FLAT TURF
| Wins | Bets | Profit | ROI | |
| 1 | 82 | 391 | -67.22 | -17.19% |
| 2 | 85 | 381 | 32.31 | 8.48% |
| 3 | 62 | 343 | -57.16 | -16.66% |
| 4 | 65 | 300 | 19.72 | 6.57% |
| 5 | 54 | 246 | 26.73 | 10.87% |
| >5 | 150 | 812 | 47.77 | 5.88% |
CHASE TURF
| Wins | Bets | Profit | ROI | |
| 1 | 20 | 79 | 10.33 | 13.08% |
| 2 | 13 | 80 | -21.62 | -27.03% |
| 3 | 13 | 61 | 1.67 | 2.74% |
| 4 | 15 | 61 | 11.99 | 19.66% |
| 5 | 12 | 42 | 14.20 | 33.81% |
| >5 | 19 | 112 | 15.49 | 13.83% |
This allows us to reduce our shortlist of contenders further. Those selections we find with Flat Turf we can further reduce by only considering those ranked four or worse for their average speed figure over the same race type.
You may be wondering why the profit is coming from the horses that haven’t performed as well on this race type and why we’re skipping those horses ranked two and who have made a profit.
The reason the horses ranked one and three don’t make a profit is because the market has already taken account of the fact that they have been the fastest runners on this race type in the race. Because this has been taken into account by the market the odds are now too low for us to make a profit on them.
The reason horses ranked two are making a profit is most likely due to not having enough data collected yet. I would expect to see that profit to get eroded the more bets we place, which is why I have ignored it.
Horses ranked four or worse are being under-estimated by the market which is why they made a profit in 2015.
Looking at our contenders on Chase Turf races we see a similar issue with horses ranked one for average speed over the same race type, apart from this we see a steady increase in Profit as a horse’s rank gets worse.
This kind of linear improvement in Profit is a good indication and again we would want to be concentrating on horses ranked four or worse where the market is underestimating their ability.
You can use these rules to find contenders. Once you’ve found your contenders then investigate them further to determine whether you should place a bet on them or not. You can also use the same process I’ve done here to discover the best way to use different factors to reduce your shortlist further.
Interesting. Can you do a stakes/handicap split on these figures? Also with favourites there can be a wide difference on the various courses. Does this apply to 2cd. favourites?
Hi Roy, I will look at getting the stakes/handicap split figures, just for seconds favs or with the other data in as well? There is a comment already posted very kindly by formstats with the course info 🙂
Where would I get the data of ranked four or worse for their average speed figure over the same race type, please?
DB
We have a rating that would allow you to do this in the Racing Dossier
Hi Michael
I record statistics on favourites a by-product of which is the number of 2nd favourite winners at each course. Overall my statistics are within 1 per cent of yours. The percentages for todays meetings of 2nd favourite winners are:
Wolverhampton: 17% Plumpton: 18% and Catterick: 19%
How those figures equate to prices and race type I don’t record as the statistics do relate to favourites only.
Favourites at Wolverhampton won 35% of races, Plumpton 38% and Catterick 33%. One statistic noted was that out of 200 days racing at Wolverhampton there were 58 days when a favourite didn’t win! There were 3 days with no winner at Plumpton and Catterick from 143 days at Plumpton and 69 days at Catterick.
It’s obvious Wolverhampton is a place to avoid when backing favourites.
If you were to blindly back all favourites at Plumpton you would have placed 976 bets. In order to reduce the number of bets I looked at the prices of the winners and those with a winning percentage higher than the favourite percentage wins for the course dropped the number of bets to 228 a reduction of 67%.
These SP prices of the favourite in my view should give a profit. They are 11/10, 6/5, 5/4, 11/8, 6/4, 7/4, 15/8.
However, whilst checking my statistics I noticed a statistic where the strike rate was 31% at Wolverhampton,
32% at Plumpton and 36% at Catterick.
These horses will be either joint-favourites (very few), second favourite or even higher.
I check the actual favourite in the race and if my expected favourite is replaced I back my expected favourite. No complicated procedures to go through .
Formstats
Very interesting, thank you for the information. Pleased to hear our stats are very close to each other 🙂
Hi Michael
Have you compared these 2015 figures to 2014 or 2013? Does history really repeat itself?
I ask because I was comparing the correct score frequency in different football leagues in different countries and found quite big discrepancies. Thinking that I had found the golden egg, I then compared the figures to earlier years’ results and came back down to earth: the statistics were totally different !
Allan
I haven’t done that specifically but would be a good thing to do. I use quantity of data rather than years. I like a minimum of 100 winners in a sample before being happy to bet to
small stakes, statistically that is incredibly low but I’ve found it to be a good point in sports betting for quantity of selections versus beginning to be statistically significant.
Bookmakers on the High Street and on their web sites take bets on the un-named fav and 2nd fav. Using betfair exchange I thought you could only back named horses. Am I wrong?
You are right, on Betfair you have to watch the market to see the second favourite
Would backing second favourites in races where the favourite is over backed and vulnerable work?
Yes, that could work well 🙂
I find that horses having their first start after a layoff are worth opposing because most winners are those which raced recently therefore to accept short odds about horses who are statistically less likely to win is madness.
What is your evidence for this Bob? And what do you define as a ‘layoff’? It’s a meaninglessly vague term which suggests to me that your theory is not based on statistical research, but mere speculation and the misconception that the fit horses raced ‘recently’. That’s another undefined term which is useless in the context of research. It is a physiological fact however, that the more frequently they run, the more the benefits of fitness begin to diminish. A horse can win run in races in a week, even four in two weeks, but at some point they become more tired than fit and need a ‘layoff’. Or a rest as we humans call it when we’re knackered! But whether the rest required is a week, a month or more is subjective and will vary enormously accordingly. Another thing you’ve not even considered is the class of horse and ability of trainer. The classier the animal and the more skill and fitness/training facilities the trainer has at his disposal, the more likely the horse will be hard fit and very competitive even if it hasn’t set foot on a racecourse for a year or more. Trainers will often say their horse will need the run, but that tends to be those who either haven’t got much in the way of true racecourse type ground and decent horses to ‘race’ against at home. Or horses with the type of physique or attitude which makes them hard build muscle definition in a water walker or in the limited grounds at the trainer’s disposal. There are loads of trainers who have no problem getting their horse to the race ready and able to win after months off, even though running against horses with one, two three runs in the past week, month etc. Bob’s theory will apply to SOME horses in SOME races, but to oppose horses that have been off the course longer than any rival on that basis alone is as hopeless a strategy as randomly backing 35% of favourites at Wolverhampton solely because 35% of favourites win at Wolverhampton! I’m sure Bob is a nice chap, but when it comes to horse racing advice, he is only going to be befriended by bookmakers.
Similar comments apply to Michael regarding the stats. he sent in the email. He’s ‘assumed’ that a ‘layoff’ means 90 days or more. WHY? It could mean 70 days or more or a month or more. He’s just guessing in order to cobble some ‘evidence’ together. Along with more vague terms like ‘short odds’….. Statistics can easily be manipulated to produce whatever result suits the agenda. And it’s much easier if you define your own parameters subjectively or don’t define them at all…the greyer the area the easier it is to disguise the legitimacy of the case being made.
Thanks for the comment Craig. As for the why choosing 90 days, this is simple to answer. In order to write an email, or an article, I need to choose a point to pull data from. Yes, I could have chosen 70 days, 700 days or even 1 day, all are a layoff. I opted for 90 days because, in my experience, the majority of punters consider a 90 day or longer layoff to be a long layoff. There was no ‘cobbling’ of evidence, I was using Bob’s message as a basis, chose a starting point of 90 days, and showed what happened in 2016 based on that information. ‘Short odds’ meant odds-on, I’m sorry that wasn’t clear. I have not claimed that this is a total analysis of lay-off’s, it isn’t, it’s a starting point, you can take it further, not use it at all or do with it what you will. It’s the statistics from 2016 based on the information I outlined in my email. As I’m sure you know, as a reader of my emails, I always recommend that you use tools, advice, stats or whatever you choose to eliminate horses and then find contenders. From that point you should use your own judgement.
I think it would be interesting to view the PLACE record of 2nd favs, especially in races of 8 runners and more. Do you have such data by any chance? Thanks.
Thanks for the suggestion George. We shall look at investigating this in the future.
WHERE WOULD I FIND THE AVERAGE B.S.P’S OF 1ST, 2ND, 3RD, 4TH, 5TH, 6TH FAVOURITES IN THE BETTING. CLASS SIX AND ABOVE RACES.
Hi Brian, the average BSP is 26.96 over 58561 horses from Jan 1st 2017 to current date. Please note that we have had some issues getting BSP figures recently, something we are in the process of fixing, so some may be missing from our sample. However the sample is large enough that the missing ones should not make a difference.
Hi I wonder where I could find the race reults last 5/6 years as 120032 or 13200. as I like to see races 2nd favs do well also select perhaps 6 races where 2 2nd favs have won. and design my strategy from there.
I’m not sure where you could be this specific, as we have our own database we can pull this type of information out but none of our current software would allow you to do this and I do not know of any publicly available that will unfortunately. We are designing a system builder that will allow non-techies to query a racing database to ask these types of questions, but we are still some way off it at the moment.
Hi Michael been looking at which race cloth number won by how many runners ran in the race 1 2 3 4 numbers come up the most is there a way to run the stats on this through a full season Flat AW Jumps not sure if type of race or distance or ground type would alter things got the most stats from 8 to14 runners in the race thanks
Hey Rob, you’d need a database to be able to run that query on. Generally speaking anything relating to cloth number doesn’t hold up long-term, that’s because there’s no reasoning behind which cloth get a number.
Hi Michael, Greetings from across the pond.. I am new to your site.as i just registered yesterday. I hope to become better informed by reading what you and other members have to share. I do have a question. During the flat season, how often is the race won by a horse from the fist 4 in the betting line and how often do you think they finish in the fist 3.
Hi Jim, great to have you as a reader and member 🙂 I’m not sure about the first four without investigating it further, but the winner comes from the first three, on flat racing in the UK, about 67% of the time. Hope that helps.
Thank You Michael