The BTC Horse Racing Thread
-
@nick-allan it's funny how things differ - I'm struggling to get anything very interesting going on the football side and I'm thinking about focusing purely on horses. The lay the beaten fave strategy is my best performing, with back fave/c&d winner/pace lays bumping along just above breakeven but have been positive a fair bit before some negative variance.
-
@joshua-varley said in The BTC Horse Racing Thread:
Variance: A Short Story
So I've spent the last 3 or 4 months (even as far back as 7 months for some) testing a range of 8 strategies that provide what I believe is a comfortable range of risk/reward. I've adjusted all of the strategies slightly, usually to remove variables, rather than include more. It's a nice mix of back and lay, with a few of the members presets, plus a few of some of the early ones that were added in June/July last year that I've tested since. There's even a couple that I've made and tested too, thrown in the mix.
I have done a lot of work around analysing the last two years of data, using it to create the ideal stake size, based on doubling the largest recorded drawdowns and then applying it to the individual bank of each strategy. All a very easy and rewarding process with this software. ChatGPT has also been a game changer and allowed me to create some absolutely killer spreadsheets.
I wanted to wait until the end of March to go live with the new stakes just to make sure I had a full 2 years of data behind the project. The equity curve below, with all strategies combined shows a lovely profit of about 4000 points on flat stakes (all lays to a stake rather than liability), and looks even healthier when my new, adjusted stakes are applied.
Anyways, all loveliness aside, I finally went live with this new plan, two and a half weeks ago. As can be seen in the first pic, the first 14 days were stellar and it lulled me into thinking I'd cracked the code... I was 50% of the starting bank up and loving life. A few of the strategies were slightly over-performing, so I knew a correction was coming.... but wow....The last 3 days have definitely had me crashing back to earth...(Those damned Pace Lays!)
I just wanted to share for a laugh - variance and probability distribution are cruel mistresses and have planted my feet now firmly on the ground!
How is everyone getting on with their set and forget horse racing strats?
Poor if im honest, im trying 3 strategies now and im not changing them, giving it 4 months and if things dont improve im going to drop horses and focus on football
-
@kevin-pepper Yeah just myself really. For example (and bear in mind I had no real knowledge of excel before), I figured it might be useful to create a table with each date listed over the two year period and it's corresponding daily P/L listed beside it (to analyse longest daily losing/winning streaks etc.) So rather than go through all that data and do it manually, I just asked chatGPT to create it based on column references etc that I provided. After a few little tweaks, it did it perfectly. There are lots of other things like that I have asked it to do and now my life is easier for it.
One of the things I liked the best was saving a tonne of time creating a monthly breakdown of all the strategies in the pic below. That's been super useful
-
@joshua-varley
Hi Josh, the CHATGPT piece sounds fascinating. Did you just work out yourself what you wanted to ask it to do or did you find a source for guidance?
Thanks
Kevin -
@david-milligan not so much in terms of optimising strategies, more in the organisation of the data. My excel knowledge was super basic before undertaking this, but chatGPT has allowed me to explain what I want excel to do and it creates the formulas for me. There was over 23000 bets across all 8 strategies for the last 2 years and my spreadsheets have now be able to break all of those down (both collectively and individually) to make analysing much more friendly. Being able to plug in different stake sizes to the whole data set and then redetermine max drawdown etc has been brilliant.
Just hoping the next 2 years of data broadly mimics the last!
-
@joshua-varley great post, I've been playing about with ChatGPT too trying to get it to work to my advantage with limited success. How did you use it to help with your strategies?
-
@joshua-varley I think this is the time when they start changing some of the courses over? I remember John said something about April and October being awkward months
-
@dan-mackinnon Yeah that's the question isn't it? I think we are definitely coming into the period each year where all strategies that I use typically perform quite well, so I think we'll be ok!
-
@andy-donnelly Absolutely grim mate - they were definitely over-performing in the lead up though. Seasonality would suggest that the next few months are good for them, so we'll have to keep fingers crossed!
-
@joshua-varley I've been doing alright but have experienced some big drops. The main question is how quickly does it recover?
-
Yeh the pace lays have had a shocking week so far, think it broke its record for most consecutive losses at one point. Had been doing pretty well for the previous couple of weeks tho so hoping it will correct itself in the weeks ahead
-
Variance: A Short Story
So I've spent the last 3 or 4 months (even as far back as 7 months for some) testing a range of 8 strategies that provide what I believe is a comfortable range of risk/reward. I've adjusted all of the strategies slightly, usually to remove variables, rather than include more. It's a nice mix of back and lay, with a few of the members presets, plus a few of some of the early ones that were added in June/July last year that I've tested since. There's even a couple that I've made and tested too, thrown in the mix.
I have done a lot of work around analysing the last two years of data, using it to create the ideal stake size, based on doubling the largest recorded drawdowns and then applying it to the individual bank of each strategy. All a very easy and rewarding process with this software. ChatGPT has also been a game changer and allowed me to create some absolutely killer spreadsheets.
I wanted to wait until the end of March to go live with the new stakes just to make sure I had a full 2 years of data behind the project. The equity curve below, with all strategies combined shows a lovely profit of about 4000 points on flat stakes (all lays to a stake rather than liability), and looks even healthier when my new, adjusted stakes are applied.
Anyways, all loveliness aside, I finally went live with this new plan, two and a half weeks ago. As can be seen in the first pic, the first 14 days were stellar and it lulled me into thinking I'd cracked the code... I was 50% of the starting bank up and loving life. A few of the strategies were slightly over-performing, so I knew a correction was coming.... but wow....The last 3 days have definitely had me crashing back to earth...(Those damned Pace Lays!)
I just wanted to share for a laugh - variance and probability distribution are cruel mistresses and have planted my feet now firmly on the ground!
How is everyone getting on with their set and forget horse racing strats?
-
@greg-mitchell thanks I thought that may be the case especially with such a high strike rate also. Will try some different angles see what I can come up with.
-
@craig-vaulks Don't think this will work. It is fine for backtesting, but it is reliant on post race and in play historical values. So the BSP, this isn't known until after the race, and the In-Running price obviously wont happen until the race is in-play. If you ran this strategy as it is, you would never get any daily selections as both of these factors are unknown until the race is completed.
-
@john-folan so I've had a play around with laying at short odds in play I think I've come across a possible strategy. I've only been using this software for a short while. So I don't know if this is something that could be used.
If you or someone else could take look at this and let me know if it's a good idea to start testing this one below, it looks like it has good potential on paper. I'm sure it could be improved by someone also. -
@john-folan im also looking into this, i've created a simple profitable strategy so far but it's only profitable if you exit the trade at around 15/20 seconds, adding a stop loss also improves it slightly. So it's more for manual trading I guess.
May look into a HDOB or TROB strats to see if there are any better alternatives.
I am also going to look into laying the pace horse as a alternative. My idea is looking for horses that typically lead near the end of the race but usually weaken so you could set the odds at say evens to lay in-play. A bit like laying the field but being more selective finding horses that usually run well but struggle to win.
Just an idea so far anyway. -
Trying to nail down a decent dobbing strategy.
Going for the old favourite of running 50% lower than bsp over a career which on the face of it shows a loss. Quite interesting as all the usual suspects (Caan Berry, Patternform IRTT, etc) push this as the go to starting point.
I’m investigating further but finding something this way is not really providing enough selections to make it worthwhile
-
Hi Martin has the horse racing email been sent this morning, I've received the football, but not the horse racing. Thanks
-
Hi folks, does anyone know how to pull the average odds from the software without doing it all manually?
-
@emily-knowles Brill, thanks!!
