Is anyone actually making a profit from the BTC Horse racing software???
-
I am new to BTC and I have created a couple of Horse Racing Strategies, I am still testing them however they don't seem to be doing very well. At what point would you give up on a strategy? I like to automation to Betfair Bot Manager. Is anyone successfully using the software?
-
@Tony-Hastie It's easily done. There are some tell-tale signs, like weird multiple number ranges. For example if your strategy "works" on races with 5-8 runners, but not for 9 runners, then it works again for 10-18 runners, then not for 19 or 20 runners, etc.
I guess it's possible there could be some kind of race that only ever has 9, 19 or 20 runners which doesn't fit the model, but it's unlikely.
Where runners are concerned, it's more likely to work/not work at the extremes of the range, so a strategy might either work for small or large fields. If you plotted it on a graph you'd get a nice bell-curve shape
-
@Adam subtractive backfitting.... OK, guilty as charged
-
Just some general pointers here in response to all the comments...
The first thing is a bit of a disclaimer which I hope doesn't sound like a copout, but the software won't "make" you profitable. It's a tool which takes the legwork out of testing your knowledge and ideas. It doesn't come up with the ideas for you (yet!) and most of your ideas probably won't be good ones. That's OK, because you didn't have to spend a year gathering the data required to tell you it wasn't a good one!
I've obviously helped people with quite a few strategies now and I do think one of the biggest problems is backfitting. It's actually really easy to come up with a strategy that happens to have made an overall profit over the last few years, but sometimes what people then do is either:
- Add a load of weird and wonderful rules like betting on mares wearing a cheekpiece for the 3rd time if they're running on a Tuesday, just because that seems to add to the profits. We could call this "additive backfitting"
- Systematically go through each result grouping and filter out certain factors based on nothing more than the fact that they resulted in a loss so far. We could call this "subtractive backfitting"
The problem is, if that gain or loss was just due to natural variance, then logically it will be due a recovery at some point, but by then you've filtered it out, so you don't get the benefit at that future time to balance out some other factor that's experiencing a downturn due to variance and suddenly your strategy goes to pot.
There should be a reason for every single rule in your strategy. Don't be put off if you feel you don't have the knowledge of horse racing that you think you need. Just approach it with curiosity - if you find an angle that seems relevant but you can't explain it, dig into it, follow the breadcrumbs and do the research. ChatGPT is actually great for this kind of thing, especially the new "1o" model.
Use the stats we provide in the software as well. The A/E score will give you an idea of whether your strategy produces more winners / losers than expected and the Archie score is a good measure of statistical significance, but it works best on a large sample size and the more categories there are for a particular rule, the lower the score per category there will inherently be. For this reason, I always try and start with rules that...err.. rule out as many categories as possible.
For example, I see a lot of people adding course filters to their strategies. Why do you think the strategy does or does not work for that specific course? Maybe you should look at something like course characteristics first, which gives you far fewer categories to work with and might expose the real influencing factor.
"Handicap Race" is often the first rule I add, because every race is either a handicap or not and often, the strategies that work on a handicap race are different to what might work on a non-handicap. Same with race type...quite obviously different things are going to work on the all-weather vs jumps races.
Also, if you're constantly changing your strategy rules so it improves the most recent results, you might just end up in a losing battle with variance as you're making what would have just happened better, which probably means something will take a natural downturn soon.
The other thing with variance is that people tend to start trading or betting a certain strategy when things are at their rosiest. You create a nice-looking strategy and it's doing well at the moment so you want to get involved, so of course it starts losing as soon as you start betting it.
This is definitely always annoying, but if you're confident in the long-term results of the strategy then it shouldn't concern you too much as you can reasonably assume a recovery (unless something else is going on!). Just use appropriate staking and bank management. It's just one of those flaws in human psychology and you see the same thing in the stock market all the time. COVID hit the economy hard and a bunch of share prices tanked, but unless you think capitalism itself is going to collapse, you should be happy that you can buy shares of big companies with the capacity to recover well for dirt cheap! I piled into my index fund and SIPP hugely at the start of COVID and it's currently sitting at a frankly ridiculous return of about 38%. This will of course even out over time but hopefully not any time soon
Remember that you're trying to create a model for what could happen in the future by looking at the past. It's probably going to be very difficult to come up with a catch-all strategy that applies in every race going forward, because some variables will undoubtedly be different about the upcoming runs compared to past performance. You're probably going to find yourself naturally using the same base rules again and again with slight variations, which is fine.
For example, you might focus on horses of a certain minimum age with a certain running-style (pace-wise, etc) who are running in all-weather handicaps. This might reduce the number of variables to a level where you can actually manage to create a good model.
Remember that seasonality is a thing, the most obvious being the flat vs jumps seasons. There could be changes within this on any given year as well though. This year (last time I checked), something like 70% of jumps races were run on soft or heavy ground compared to 39% last year. 300 fewer jumps races were even planned for this year too. I'm not sure of the effect that had on prize money, whether there was more on the races that were planned or not.
Regarding price, value, staking etc, if you're currently using flat stakes in your strategy then consider using variable staking. The software makes it very easy to see the difference between laying at flat stakes vs to liability, or to vary your stakes based on different price ranges. We're going to add variable back staking based on the payout of the bet as well. This ensures that whatever edge you do have is delivered in proportion to the odds.
For what it's worth, don't shy away from higher odds, lower strike rate strategies. This is a bit of a "do as I say, not as I do" because I'm really bad for this. But the fact is, it's easier to find value at those higher prices than it is on the favourites which are more efficiently-priced.
Actually, a question for anyone who's managed to make it to the end of my early morning ramble...
If you've got a strategy that's a bit close to the bone, i.e. your strike rate is only just above or below the break-even strike rate, see if you can identify what kind of percentage odds uplift you would need to make it profitable. If you're decent with spreadsheets, maybe you could export your results and add an "adjusted price" column where you improve your entry price by x% until you're happier with the result. Or just calculate it from the overall average odds of the strategy. Let me know the percentage you increased it by.
I have an idea for a betting rule we could add to ensure everyone obtains the best price possible in certain situations and want to test it out
-
@Joseph-Henderson I keep fundamentally the same strategy, Ill just tweak whats within it. Thats generally odds range, courses and race distance. Ill look at race classification, days of the week and class too sometimes.
Having had a look at Martins now he has the horses age as well. Id like to add that(as have in the past) but its too hard to automate(for me) so I go without it.
I dont know if Im backfitting or not. I believe that Im working with the latest stats in a market that doesnt stay the same for long, so Ill keep telling myself that! Time will tell if it works or not...
-
@Tony-Hastie that's interesting that you are making changes monthly. Do you just create a new version and adjust based on what did/didn't work the previous month? Could you give me some examples of the kind of tweaks you have been making month to month? My only concern with that approach is, are you not just back fitting to improve past results?
I agree, automation doesn't seem to be working for me long term with horse racing. It seems a very complex sport with so many variables at play.
The only strategy I am really still going with is a version of the short odds lay strategy, which has returned me 4.42pts since July. After 451 bets, ROI of 0.87% and an estimated edge of 0.22% I think its days are numbered really, doesn't seem worth pursuing.
-
@Joseph-Henderson Ive been DOBing for over a year now. What have I learned? No one thing works for very long! The same criteria that had the best 2 months(may and june) Ive ever had has bombed in july and august. Ive become firm in my belief that horse racing in general requires a more fluid approach so Ive taken to updating my strategies monthly.
The lay the third fave that Martin shared would probably be the goto racing strategy to look at. I run something similar, and as mentioned above I shot to close to 40 points profit in 2 months....it then dropped to +30 points...and has annoyingly hovered there ever since! Ill take a look at seeing what I can change at the end of the month, and Ill have a gander at Martins version too
-
I've not been having much success recently on the horses! The Dobbing strategy that John shared earlier in the year appears to be on a downward trajectory for me. Since 1 June 2024: 254 trades, 48% strike rate, -4.10% ROI and -10.42 units. Think I will call it day with this strategy, but keen to try any new strategies people are having success with!
-
@Alex-Weston did you download the dob strategy?
I've got a video coming out today at 11am, make sure you are subscribed so you get updates: https://www.youtube.com/@BetfairTradingCommunity
-
@Martin Thanks I've been doing your lay the 3rd and its going okay so far. I think my Bot was was not using the correct odds previously. Have you got any content on Dobbing?
-
@Alex-Weston I'll create a new database for them and tag you.
Dobbing is double or bust back to lay strategy, you back for £10 for example and trade out when price reduces 50% for a £10 profit on each outcome.
-
@Martin So its adding recovery? That would be much appreciated thanks.
-
@Alex-Weston dobbing is double or bust. If you reply to this message so I get notified I will upload my filters when I'm on my PC
-
@Martin I have the lay the 3rd Fav too but its down a bit, shall I send you it? Whats dobbing?
-
@Alex-Weston yep been doing well with lay 3rd fav and dobbing recently, what are your strategies? I can give you some help if you like!
