CPC's Horse Racing Systems Performance and Ideas
-
@lee-woodman
Sorry for the delay. I have the L88 running at micro-stakes on all races currently to see if a minor tweak can get it back on track. Added an additional filter that the fav had to be priced at 3.5 or less for a bet to fire. So far it is about on par after 300 odd trades. Will give more updates as the year progresses.
-
This post is deleted!
-
Hi Charles. You specified that your LS88 system is handicaps only.....have you looked at the results for non-handicap races aswell? Ive been running it for about 6 weeks at small flat stakes whilst accumulating some data for staking purposes but its inadvertently picked up non-handicaps also. Im going to start recording which are handicap/non-handicap races, but i had a quick check of the 'losers' and only 2 out of the 25 losers were non-handicap. Ill keep recording until i have enough data to draw any conclusions, i was just curious if you'd checked
-
@charles-cartwright
June results below:
As I discussed on the horse trading thread I discontinued the LaySys88 in the middle of the month after it dropped over 5 pts. No explanation for it but it was just heading downhill. I will run it at microstakes for the rest of the year just to see how that turns out and will share at the end of the year. Overall a solid month for everything else a little loss on the L2B but the B2L66 bounced back after a crappy May.
So at the end of Q2 this is what it looks like:
Another solid quarter and I am very pleased with how this is tracking right now. Right around 60% ROI per quarter (so up 30% over my forecast).
Won't be changing anything for the next few months since all systems are running right now (except for Lay88). On track for my goal which is to have 100K in my betfair account in 2 years time.
-
@charles-cartwright Thank you for sharing buddy, I've always been a big fa of the 1-4 approach.
-
Loss Recovery
Got asked about what approach I take to loss recovery which is a topic I could drone endlessly about (being a bit of a stats nerd). Anyway, after much trial and error with this my approach now is the following for any straight lay system:
-
All trades are to fixed liability.
-
The standard liability is 0.5% of your betting bank. The maximum liability is 2% of bank.
-
Betting bank is reset monthly.
-
Loss recovery uses the 1-4 approach as a base. This simply increases the liability by 1 unit at the first loss, remains at 2x until the original loss has been recovered or until there is another loss. At the second loss, liability increases to 3x the original stake again until loss has been recovered. If a third loss occurs before loss is recovered then liability goes up to 4x until recovery is complete.
-
I vary the above slightly by only requiring that loss recovery be continued until the overall P/L on the system is equal to or above the level at which the liability was last at 1x.
Confused? Here is an example:
We lose two in a row here, the 16:40 at Plumpton and then the 17:25 at Wolverhampton. So that results in us staking 3x (1.5 units) on the 17:55 at Wolverhampton. We continue staking at 3x until the P/L (rightmost column) returned above 5.72 which was the last level at which we made a successful 1x trade.
- Most importantly you have to have a stop loss. I use 10x base stake as a monthly stop loss. So for lay systems this is 5 pts. Doesn't happen very often but you have to be prepared to bite the bullet sometimes. Did it this month on Lay88. Will start it up again next month.
-
-
@charles-cartwright brilliant, I will give this a go in the morning. Much appreciated!
-
@lee-woodman said in CPC's Horse Racing Systems Performance and Ideas:
@charles-cartwright Hi Charles. What software do you use for the LS88 system? Ive got BF Bot Manager (only just, so im still figuring it out) and im looking to automate some of the stuff and i cant quite get my head how to implement the condition 'fave to be 1.5+ lower odds than 3rd fave'. Ive been running the strategy manually but its not viable to catch every race!
You can set that parameter in the markets tab under the rule; "Skip market if favourite price difference is not satisfied".
-
@charles-cartwright Hi Charles. What software do you use for the LS88 system? Ive got BF Bot Manager (only just, so im still figuring it out) and im looking to automate some of the stuff and i cant quite get my head how to implement the condition 'fave to be 1.5+ lower odds than 3rd fave'. Ive been running the strategy manually but its not viable to catch every race!
-
Here is my update of how my systems performed in May and what 2019 is looking like through the first 5 months.
Pretty solid month overall after a somewhat sluggish April. Pleased that even though a couple of my main systems (LS88 and B2L66) underperformed, the month as a whole was well above forecast. Had a couple of nice winners thrown up by the ProForm Turf Sprint system (TS19) and the lay-to-back (L2B) that I developed in the past couple of months continued to be productive.
Here is what the year to date looks like as we go into June.
For simplicities sake I combine all of the lay systems for the quarterly and YTD updates. My forecast ROI on a quarterly basis is 45% and so well on track to meet that assuming June doesn't bomb horribly.
Any questions please don't hesitate to ask. Again hope this provides some kind of framework for realistic expectations of what your returns can be when trading horses on BF.
-
@charles-cartwright
Here are April's final results for all of the racing systems that I currently use:
Overall a disappointing month. April can be a tough month for trading races with the transition between NH and Flat racing and the disappearance of competitive AW racing.
The system listed as TS19 is a ProForm derived approach for trading sprint races on turf and that will appear in the listings from now until November.
The main lesson here is that these kind of periods happen to all of us. Sometimes there are obvious explanations and sometimes not, but the important thing is to take downturns in context of long term strategy and returns. Given the really solid first quarter I am still ahead of forecast for the year (74% ROI v forecast of 60% through April) so with the gradual improvement expected as we move into the summer confident I can meet my goal for the year.
-
Going to try and post semi-regular updates on this thread with commentary on how things are going and any other thoughts I may have.
April is a really weird month in racing. Represents a big transition from NH and A/W racing over to the flat. NH opportunities tend to dwindle as the ground gets better and fields start to dwindle. A/W also dries up somewhat as horses are switched back to turf. For systems like my B2L66 this is a challenge to find anything worth making a trade on. Always need to resist the temptation to just 'force' trades to maintain activity.
Don't think I have forced anything this past week though the results on B2L have not been great so far this month. Just 6 trades this past week and 3/3 so -2.04 pts. None of the horses would have converted a true DOB, however, and none came even close to winning the race so somewhat fortunate to only be a bit below par. My guess is that there will not even be 30 trades on this system in April.
The L2B system that I have been working on has also been very quiet, with only 5 trades this week. All were successful, so that is good, so +1.39 pts on that one.
The Lay88 system has got off to a terrific start this month, and appears to be following the same pattern that happened last year.
27/27 as of today and + 2.2 pts. Since each pt here represents 2.5% of the 'at risk' bankroll that is a 5.5% increase in bankroll on just that system or about 17% yield.
Last year only lost 1 trade on this system in the whole of April and drove the system through for the whole summer. No idea if that will repeat but it's an interesting phenomenon.
Will be traveling at the end of the week and so may only have this system running next weekend. Might turn out to be a blessing in disguise.
Later
C
-
@charles-cartwright That's not a worry at all - I'll come up with some Qs
-
@ryan-carruthers said in CPC's Horse Racing Systems Performance and Ideas:
This is incredible!!!
Would you be up for us recording a members only video call delving into this a little and the mistakes you've made?
Ryan....flattered by the offer but for a number of reasons I prefer to stay somewhat in the background. I would be happy to answer questions in blog postings though.
-
The second post in this series will deal with my approach to Staking and Bankroll Management. This is pretty intuitive and you can see how it works by looking at the results table posted earlier.
One of the things that I think is universally helpful is to basically forget about the 'money' that you are 'winning' or 'losing' with each session, day or whatever of trades. When I started on this I would go crazy if I lost money and immediately try to 'win' it back. That meant scouring for any old betting market and lumping in with the predictably disastrous results. So everything for me now is related to points which simply represent fractions of what I call my 'effective' bankroll. This 'effective' bankroll is the fraction of my total account that I am committing to be used only for system trading and is usually around 85-90% of what is in my account. The remaining 'float' I use for regular punting, daft accas and other trivialities that for my are part of the fun of watching sport. Again I find that by segregating those two pots I am able to enjoy a small gamble without it affecting my trading bankroll.
Another part of my philosophy is that my trading is always agnostic with respect to the 'probability' of any individual trade being successful or not. Varying the level of risk according to your subjective opinion of that particular trades value is simply gambling. I identify an appropriate level of exposure for each strategy and use it without fail.
I start each quarter with a nominal bankroll of 100 points. Depending on what has happened the previous quarter, this may be the same as the effective bankroll or I may have harvested some profit and thus reset the bank. For the quarter just finished, for example, it was well ahead of forecast and so I did recoup some of at that extra profit.
For the LS88 system 1 staking point is equivalent to 2.5 bankroll points (2.5% of the bankroll on day 1). For the 1-4 loss recovery approach, my standard stake is 0.5 staking pts and this will go up on losses to 1, 1.5 or 2.0 staking pts. The maximum risk therefore is 5% of the effective bankroll. I will only go that high because of my absolute confidence in that system proven over analysis of thousands of races.
For any backing system, a simple 1 to 1 ratio of staking to banking points is used. For B2L66 I use a 2 pt level stake, for most other backing systems either 1 pt or 0.5 pts.
I used to use a % bankroll staking system and update it daily or weekly. I found that both a lot of work and stress inducing at the start of quarterly intervals (when a couple of bad days can make things look really bad). All I do now is update the results on a monthly basis and then adjust the staking points appropriately so that they match the bankroll (so that you can take advantage of a very positive month and keep your risk level if you have a bad month). You can see this in the results table. Everything relates back to the original 100 bankroll points that the quarter started with.
There is no one size fits all for managing risk and bankrolls but I have learned the hard way that for me discipline is the absolute key to success and all of the details of how I work are to ensure maintenance of that discipline.
Happy to answer any questions and I will add any additional thoughts I have as they come to me.
-
@dave-jessop said in CPC's Horse Racing Systems Performance and Ideas:
@charles-cartwright said in CPC's Horse Racing Systems Performance and Ideas:
How do you implement the 1-4 loss recovery on the LS88?
Basically this is a set and forget system using the bot so is your loss recovery determined on a daily P/L?The loss recovery is automatically updated after every race where a trade is placed. The only limitation on this (which is not really an issue) is that you have to make sure that you only have one open market at a time. If you don't then the risk is that it will overstake on the additional markets.
-
Very interesting read, I look forward to your next post
-
Q1 2019 closed today on a quiet Sunday for me on the racing front. Have tabulated the final results for the systems that I have used this year along with P/L statements:
Bottom line is this was a very good quarter with an increase of 71 base points from a nominal starting bank of 100. That represents 26 pts over my forecast of 45 pts based on past performance. The biggest difference is that I feel that I have really tightened up my B2L approach and that produced a 33pt return as a consequence (a 14% yield versus an expected 6% or so).
This table probably needs a fair bit of explanation for anyone trying to understand it, and I will try to cover all the pertinent points over a couple of posts.
Firstly......
Systems
There are 4 independent systems listed here. My nascent attempts at a Lay-to-Back (L2B) system is only listed for the month of March. This is still very much a work in progress but has some promise to be complimentary to my other lay system and so it will retain its place for at least the next quarter.
LS88
This is my bread and butter lay system that I have worked on, refined, torn my hair out over, and finally settled on for at least 3 years. It was always my intention when I started this to have a highly reliable, hands-off, laying approach to the horse markets that I could simply let run with basically zero effort or intervention and this is it.
System places lay bets on 3rd favourites in the exchange market at race time under the following situation: Handicap only; 10 or more runners; favourite price must be at least 1.5 pts shorter than 3rd fav; 3rd fav lay price must be greater than 5.5 and less than 7.5. This will very reliably produce about 100 bets per month at an average price of 6.9-7.1 and a S/R of 88% (hence LS88). This gives you an expected yield at level stakes of 2.7% (0.147-0.122).
Not exactly spectacular but very reliable. Most of my tinkering involved attempts to 'improve' on this return by using loss recovery mechanisms. I have to say that my early ones were laughably naive and incredibly scary and resulted in massive losses if say 2 losers in a row or 3 in 10 happened...and it does! In the end I found that simple is better, and so now I use a 1-4 loss recovery system and have my base liability set at 0.5 pts (which is 1.25% of my 'in play' bankroll).
This quarter's results on this system were pretty typical:
87.7% SR at an average price of 7.2 and 253 trades. A little down on total trades because of the loss of NH racing for a week in February. As you can see, by using the loss recovery approach I only wagered 1pt more in liability than I would have laying level stakes at 1pt but my yield was 3.8% garnering 9.57 base points profit. Level staking would have resulted in only a 2% yield for this quarter (due to marginally higher average price) and thus only 4.57 base points profit. Pretty dramatic.
The key point is though that you HAVE to know the performance of your system with a high degree of statistical confidence before trying to employ loss recovery. That means analysis of at least 500-1000 trades to obtain highly accurate assessments of variance, yield etc.
This approach will basically never run into any kind of liquidity issues at bankrolls of up to around 200,000 (because of the massive liquidity on even very low level racing in the UK) and it is my ultimate aim to use it as essentially my sole high stakes system. For people who don't have the time or are on the wrong time schedule to really take on the in-play driven markets (football, T20 cricket, tennis etc) then this IMHO is the way to go.
LOOF (Lay Odds On Fav)
This is one of 2 microsystems that I use; the other is a backing system for Turf Sprints that will kick in now the flat season is underway. LOOF lays horses that are odds on (price below 1.81) that have only run once and that race was in the last 45 days. Those horses win only about 57% of the time at an average price of 1.6. About 150 qualifiers a year and turns about a 10% yield in a typical season. Will pick up in numbers once the flat season gets rolling.
B2L66
This is my most 'interactive' system. Idea is to identify horses that are 'likely' to shorten in running such that a profit of 66% of the original stake can be obtained. Betfair starting price. Means the horse has to drop about 40% in the market (and the lay bet has to be matched) for a win. This requires some basic prep work (identifying horses that have good records of running will below SP) and then some hands on work to look at more detail (race quality, going, opposition etc) to make it successful. I have worked much more diligently at those latter aspects in the past few months and that has started to pay off. Takes me between 20 min and an hour depending on how busy the next racing day is to sift through the races and taking the extra time has proven well worth it. I remain concerned about whether this can be sustained up to the level of wagering I want to be at, because of the challenge of getting substantial in-play bets matched in running (especially in relatively short races) unless you are placing the wagers manually and can vary your strategy. Will see what the next three months bring.
Next post will discuss staking and bankroll management.
-
Look forward to seeing these Charles
