Horse Racing Strategies Megathread
-
Thanks @joshua-varley. It's always good to have a fresh pair of eyes look at these strategies.
TBH I'm not sure. I was just looking at the software whilst the better half was watching Strictly.
If the filter is looking at the full race then it;s not as good as it first looks. Maybe @Adam can look to add traded volume and rank to the start of the race in a future build.
P
-
@paul-foster HI Paul - looks bloody good, well done! My question is to do with the Traded volume rank. When looking for horses to qualify, is the software looking for horses that ranked first for traded volume in their previous race, who have gone on to be favourites and won their next race? I know that presently, the traded rank takes into account the whole race, including in-play, but I am just unsure whether it only applies to the race before the win. Brain melting stuff for a Friday night haha - perhaps @Adam might have an answer?
-
This is my latest attempt at a strategy. I'm looking for favorites which have the highest traded volume from qualifying races.
Next task is to get this to work in BFBM
-
Do you have the files for bet angel?
-
@nick-riley said in ***Horse Racing Shared Strategies Megathread + Downloads***:
Can't you just look at the monthly breakdown to see how it's performed over the last 4 months? I'm not sure what extra information this is telling you?
For the base filter you make up yes. But when your filtering more like removing courses and horse classes your using the known results to make those decisions so this way lets you test if the extra filtering worked with unknown races.
-
@nick-allan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@john-folan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@robert-h Sound logic there.
so the logic is to look at back testing the last 4 months not 18?
Nah its to back test using 14 months so exclude the most recent 4 months. Once happy with the filter you test it only using the most recent 4 months. As that data for the last 4 months was unavailable when you made your strategy you get to see how it works with unknown results just like u were using it to place a bet on a race tonight you don't no that results.
So the last 18 months the first 14 months used to backtest to make strategy then the most recent 4 months used to forward test and verify the strategy
-
@nick-allan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@john-folan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@robert-h Sound logic there.
so the logic is to look at back testing the last 4 months not 18?
No. It's find a strategy. Then keep an eye on it for 3 months going forward. So what he has done is looked at the date posted, then looked at results since then.
-
@robert-h said in ***Horse Racing Shared Strategies Megathread + Downloads***:
So I just picked a random preset strategy Back | Maidens Long Odds and added the specific date filter of 1 jan21 to 18 jul 22
Gives this result between 1 jan21 to 18 jul 22So the strat looks good backtested but now to forward test it change the specific date to the last 4 months
We didn't have any of this data when making the strategy and running it originally so this will see how it would have run the last 4 months. So run it and this is the results of last 4 months of forward testing
So forward testing shows its a big change to the worse and makes it look like it was just back fitted with the data we had but with unknown data it wasn't a profitable one.
Hope that makes sense I'm not really good at explaining things in detail and making it make sense
so the logic is to look at back testing the last 4 months not 18?
-
@john-folan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@robert-h Sound logic there.
so the logic is to look at back testing the last 4 months not 18?
-
@martin Can you pull my Slow Horses and Pace strategies please. I'm having a bit of a rethink on both. Slow horses because it has just gone static and Flat Pace, because I've just noticed all of the profits is in the 8's and i want to know why. Plus the variance on those is a bit meaty for most people.
Short Odds Lays please leave. They have returned a reasonable amount over 500 trades as per below.
42 points from 522 trades. ROI 0f 8% and a max drawdown of 15% of it's bank. This despite a 20 point drawdown during August's extreme heat which it has now recovered from.
Thanks
-
@nick-allan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
Is there anywhere where you can ask direct questions about the software perhaps with Adam directly? I noticed there is nothing on the QandA section. I’ve done a deep dive looking at software results and my results based on the selections that the software provides each morning. Now to be honest some strategies showed completely different results, which I can't understand why? I need to get some advice as i don’t understand if the filters provides selections and they meet the trade criteria at race start why they are not in the software results? does anyone know if you can message Adam directly?
Yep use the Asana form in the opening post: https://form.asana.com/?k=V4sICWddOgH3CMQ96g0XSA&d=1120854233877885
-
So I just picked a random preset strategy Back | Maidens Long Odds and added the specific date filter of 1 jan21 to 18 jul 22
Gives this result between 1 jan21 to 18 jul 22So the strat looks good backtested but now to forward test it change the specific date to the last 4 months
We didn't have any of this data when making the strategy and running it originally so this will see how it would have run the last 4 months. So run it and this is the results of last 4 months of forward testing
So forward testing shows its a big change to the worse and makes it look like it was just back fitted with the data we had but with unknown data it wasn't a profitable one.
Hope that makes sense I'm not really good at explaining things in detail and making it make sense
-
@dan-mackinnon said in ***Horse Racing Shared Strategies Megathread + Downloads***:
Sorry, I'm a bit confused by this. Are you saying you forward test 4 months on the software? Or you actually forward test and compare?
I use the software to forward test for 4 months. As using the specific date filter removing the last 4 months then creating your strategy you can filter your strategy using all the data before the last 4 months then when happy run it again but only using the last 4 months in the specific date to see how your strategy works with data you didn't have access use to create the filter.
If the last 4 months results are bad I normally scrap the strat as it prob was only profitable due to back fitting or I modify it if its not to bad.
I'll try give an example
-
Is there anywhere where you can ask direct questions about the software perhaps with Adam directly? I noticed there is nothing on the QandA section. I’ve done a deep dive looking at software results and my results based on the selections that the software provides each morning. Now to be honest some strategies showed completely different results, which I can't understand why? I need to get some advice as i don’t understand if the filters provides selections and they meet the trade criteria at race start why they are not in the software results? does anyone know if you can message Adam directly?
-
@robert-h said in ***Horse Racing Shared Strategies Megathread + Downloads***:
@nick-allan said in ***Horse Racing Shared Strategies Megathread + Downloads***:
This is a genuine question for everyone who's got the horse trading software? Is anyone out there making actual money on it? I've been testing for nearly four months now using about four or five strategies and hasn't been any that have been consistently profitable and most would have lost quite a bit of money (don’t worry I’m on low stakes). I see there is a lot of information in the chat about strategies that people have been tweaking and showing the backtesting which is great, but are the actually working in a live testing environment? I understand with any new strategy you need to test it for three to four months I totally get that, but sooner or later the software has to show some form of value would you not disagree? And if none of the strategies are profitable then where is the value? Even the pace lays which had a good month last month are having an absolutely dreadful month this month.
So I'm interested to see if anyone is actually making money and on what strategy they are because maybe I'm just looking at the wrong strategies? Another example will be this recent back the favourite on Saturdays, Mondays and Tuesdays, now granted I've only been testing it for two weeks now but it's on like minus 24 points which is some drawdown.
I hope this isn't perceived as negative because it's not meant to be I'm just genuinely interested if anyone is making money in a live environment and it doesn't matter even if their one pound stakes or 50p stakes have they got any strategies that are making some form of profit?
For me I found at the start most the strategies didn't work when forward testing them. I was using strats created by users on here but I now think that most of them were over over filtered mostly with courses being back fitted and some other filter being back fitted. I find alot of the improved ones of the originals posted here just been back fitted with results since the original has been made so of course going to look impressive.
I now try to filter less and only filters that make sense why its happening not just chance. I still might filter courses but have to be losings lots there get removed (ie certain course layouts). I use BF bot manager so now trying to make strategies that don't require importing selections to make it easier John short odd lay strategy is a cracker and he gave me some good tips with the bot when getting started. Running a modified pace letter lay which is good on the few qualifiers i get have to manualy put those into the bot. And run johns slow pace dob but that hasnt been doing that great but still sticking with it for now.
My main thing thats helped me was I found a way to forward test using the software. The way I do it is to forward test for 4 months so when making a new strategy the first filter to add is Specific Date then put a date from before the records start so greater then 18 months then finish 4 months before todays date. Then create you strategy run the result and do even more filtering if needed. Then once your happy save the current then click save new to make another copy then go to the new copy and change the specific date and change that from the end date of the original filter(so 4 months ago) to today's date. Then hit run which would give you the results of the 4 months you hadn't filtered.
Hope that made sense using this I have found alot of filters that looked good weren't actually that good forward testing so straight away you know and don't have to wait 4 months
I agree regarding the potential backfitting when specific courses are removed when there might not be any logic behind their poor results on a filter.
There may be times when logic could be applied, say when you're looking to lay short price horses in running when there's a long uphill run in, a poor TV camera angle etc. I remember some discussion a while ago regarding a strategy on the software for Chester run-ins...
That got me thinking whether there's a resource anywhere that captures this type of info across the different racecourses. Anyone know of any sites that might help? I'd be happy to collate something if there are a few sources?
Cheers
Kevin