Once you’ve read this article you should check out this one I wrote about how often odds-on favourites lose. It was the great comments, and in particular two, that spurred me to write this post.
I was asked these questions:
- What about focusing on the very short losing runs?
- What about laying?
- What about using a progressive staking plan?
The truth is that I didn’t know the answer to these questions. Because I’ve never investigated them.
Well, today it’s time to investigate.
As in the previous article I’m going to use data from 2013 to investigate.
We start with with knowledge that odds-on horses win a lot of their races. 59% to be precise.
But…
In 2013 the longest losing run was 9. That’s pretty long considering how much profit you’re making back. This shows that the short losing runs aren’t as short as we may think.
Instantly that tells me that we’re very unlikely to find any kind of progressive staking plan that will allow you to make your profit back at odds of less than evens when you can hit nine losers in a row.
At least one that doesn’t require a huge bankroll and a bookie who’ll take your bets.
That means we need to narrow down the selections. And I’m going to do it differently to the way I did it last week.
This time I’m going to focus on the individual horses as opposed to the race conditions.
Amazingly there are horses which go off at odds-on but have never won a race. Would you want to bet a horse odds-on that has never won a race?
But maybe I’m being unfair. It could be the horses first time out.
Although, to be honest, I still wouldn’t want to bet it at odds-on.
However, I will remove any horse from the sample that hasn’t had at least ten previous races.
This significantly reduces the sample down to 589 runners. But… there are still horses that have never won a race being bet odds-on!
The average strike rate for these horses is 24%. So what happens if we look at runners who have won 30% or higher of their races AND had at least ten runs:
| -3.83 | Profit |
| 137 | Bets |
| 86 | Wins |
| 63% | SR |
| -3% | ROI |
| 6 | Max Losing Run |
The above results are to Betfair SP and take account of commission. We’re still making a loss, but the Max Losing Run is now just six instead of nine.
A big difference.
Of course, we’ve sacrificed a lot of bets and now down to just less than one bet every two days on average.
What happens if we decrease the minimum race threshold to three:
| -7.27 | Profit |
| 294 | Bets |
| 183 | Wins |
| 62% | SR |
| -2% | ROI |
| 4 | Max Losing Run |
Interestingly the maximum losing runs decreases to four. This could be a data anomaly but could also be due to the fact that there are now more bets and more chances for the losing streaks to be broken up.
So, is it possible to use a progressive staking plan now?
At this point it is important to say that I am very anti staking progression. It’s dangerous, and usually does more harm than good. However in the interests of a fair test I will go ahead with the investigation.
The average odds of these horses is 1.63, with the lowest being 1.07 and the highest being 1.99.
Let’s use the average odds of 1.63, which after commission is actually 1.60, and losing streak of four to look at progressive staking where we just want to make our profit back. If you were to:
Bet ÂŁ1 on first runner and lose you would be -ÂŁ1
Bet ÂŁ1.67 on second runner and lose you would be -ÂŁ2.67
Bet ÂŁ4.45 on third runner and lose you would be -ÂŁ7.12
Bet ÂŁ11.87 on fourth runner and lose you would be -ÂŁ18.99
Bet ÂŁ31.65 on fifth runner and win you would be ÂŁ0.00
You have risked ÂŁ50.64 to break even.
Imagine if you just wanted to come out with a ÂŁ0.50p profit. Then you would have risked ÂŁ73.69 to make ÂŁ0.50p.
Does that sound like a suitable risk?
It doesn’t to me. And so I can conclude that you would not want to bet on these runners using progressive staking.
So… what about laying…
If we lay every horse at odds-on then we get:
| -32.41 | Profit |
| 1760 | Bets |
| 690 | Wins |
| 39% | SR |
| -3% | ROI |
| 14 | Max Losing Run |
Not looking great.
What about if we look at those horses that have never won a race?
| 13.74 | Profit |
| 581 | Bets |
| 241 | Wins |
| 41% | SR |
| 4% | ROI |
| 17 | Max Losing Run |
We’ve made a profit for the first time. The ROI isn’t great but we have just over a selection a day on average and have come out with 13.74 units profit at the end of the year.
Although it’s not going to make us rich, considering it takes a few seconds to find these selections that’s not bad.
Let’s change our criteria and look to only lay those horses who are racing for the first time:
| 3.96 | Profit |
| 69 | Bets |
| 31 | Wins |
| 45% | SR |
| 8% | ROI |
| 5 | Max Losing Run |
A better ROI but a smaller profit due to the quantity of selections. If we add them both together it gives us:
| 17.71 | Profit |
| 650 | Bets |
| 272 | Wins |
| 42% | SR |
| 4% | ROI |
| 18 | Max Losing Run |
Last week Joe pointed out that we should be testing these results on out of sample data. This is data that hasn’t been used to build our theories on.
And he’s right. This could just be a lucky year, or things could have changed.
So what has happened in 2014 so far is:
| -25.06 | Profit |
| 600 | Bets |
| 226 | Wins |
| 38% | SR |
| -6% | ROI |
| 10 | Max Losing Run |
A very different story. And the key is in the different strike rates. The strike rate in 2013 for laying these runners was 4% higher than it has been in 2014.
Why has this happened?
Well there are a few possibilities:
- The market has adjusted to become more efficient on these selections
- 2013 was a lucky year for these lays
- It will even out over the last three months of the year
The most likely of this possibilities is the first one. The betting exchange markets adapt very quickly to inefficiencies. In 2013 the odds were too low on odds-on runners that had never won a race or were running for the first time. In 2014 the market has adapted and that is no longer the case.
Summary
In summary it’s very difficult to make long-term profits from odds-on runners.
Can it be done? Yes.
But you will spend a lot of time building your strategy for very little reward that could disappear very quickly.
So the question you should be asking is… Is it worth it?
For me the answer is definitely no. I prefer to focus on horses with higher odds and learn that I will have to cope with losing streaks. That’s a small price to make a lot more profit.
You have just confirmed all my conclusions on both short-priced favs and
staking systems reached the hard way -i.e. over 50 years of experimentation & homework starting with
Methodmaker in The Handicap Book…anyone remember him?! However in the “olden days”,with much less racing than today(no AW or covering of tracks, or interest in overseas events) losing runs of 8 to 10 were commonplace with the Deathstar 16+ putting in an occasional appearance to wipe you out. This was backing odds-on favs in fields of 7 runners or less. Hard to remember now but damage limitation strategy was restricting bets to certain tracks in certain months. Overall NH was more proftable & so were the smaller tracks where SPs often reflect the grapevine effect where small trainers know pretty much exactly where they stand in relation to one another and to any big stable intruder too. Betting SP was essential becos the clever local money piled on just before the off…getting early prices was no big deal. Anyway thanks Michael 4 crunching the numbers. never a dull moment eh?
Thanks Josephine, for the great info and comment. I do remember Methodmaker. 🙂
Thanks for the well-reasoned analysis. I recall Little Acorns system for laying odds on favourites with certain conditions – eg other reasonably well-backed horses in contention and excluding any selection that was BF last time out. This used a frightening fibonacci progressive betting plan which of course keeps you handcuffed to your digital link to Betfair, but the system – unlike most – seems to have stood the test of time and remains in use. Not for the faint-hearted.
I haven’t heard of Little Acorns for quite a while, but do remember it performing well. I will try and dig out a copy and run an analysis on it.
Hard to back-test it since backing decisions were made live when odds were fluid on Betfair Exchange, 10 or 15 minutes before the off. Second and third horses have to meet odds criteria before a bet is placed. I remember the details, but not fair to give them here. There were (?are) two recovery plans, the Fibonacci being the more aggressive.
Although I never used it, my notes from the time show that the longest losing run experienced by users in 2008 was 6. Punters also ran it through Adrian Massey’s database and found that over the 7 years previous to 2008 there would have been one losing run of 9 and one of 10.
Clearly you would need a stop-loss system to avoid having a heart attack. After, say, 6 losing wagers on a Fibonacci progression you would be wagering 13 pts on the seventh bet. Eight sequential losses would take you to a ninth bet of 34 points, by which time you would already have lost 54 points, albeit the liability per point staked might be about .75. But punters betting at low levels eg ÂŁ2 per point could cope with a losing run of 9 on a bank of a few hundred pounds. Some called it a day after 6, some after 8. And started again.
Thanks for the extra info. I do have a copy of it somewhere myself so will dig it out. We have some Betfair data in our databases so it’s possible we may be able to back test the results to odds 10-15 minutes before the off with second and third horse odds criteria. I will let you know if we can 🙂
I maintain data on favourites and over 11808 odds-on or evens races there have been 6738 winners (57%)
Handicap races accounted for 2125 of these races of which 1151 were winners (54%)
Non Handicap races numbered 9683 with 5587 winners (57%)
Percentage winners were:
Flat: 57%
All Weather: 55%
National Hunt: 59%
Field sizes 1 to 5: 60%
6 to 10: 57%
11 to 15: 53%
16 plus: 54%
From the above I drew the conclusion that I should turn my hat round and face the other way and look at laying. Laying is a mugs game on anything other than odds-on bets which have a losing value less than the stake. From this you can see I am a pessimist!
I did try laying and can tell you what the odds-on statistics are at every course plus the profit or loss figures for every price at the course.
This method didn’t excite me so I looked at other ways of gaining profit from odds-on races.
From June to the end of August this year I made a profit of 135.62 points net BSP by using my Daily Selections where I pick the best 3 rated horses and place bet only the selections which are not odds-on. As Michael is always saying – losing runs will come and you must give a system 3 months before determining whether it is profitable or not. Well I have had three profitable months followed by a loss of 35.27 points net BSP and currently am losiing 10.74 points net BSP. I do believe however that the method is robust and profitable and has proven decent profits can be made but not in the blinkered way some punters react.
Hi i have read both these blogs & am very keen to check out the 9 loss odds on you mention could you please give me a date in 2013 when this happened so i could look back at them please…
Regards
Andrew
Ps: keep up the outstanding work
Hi Andrew, thank you for your comment. The data sets I use are taken at the time to write the article and I don’t keep them which means I haven’t got that specific dataset anymore to give you the date unfortunately. However if you were to go through the results of 2013 odds-on runners you would find the losing run pretty quickly.