The BTC Horse Racing Thread
-
@lee-colvin said in The BTC Horse Racing Thread:
Brilliant Feb for John’s SOL strategy folks eh? I ended the month just over 14.25 points up laying to a one point liability, and I missed out on a few winners as well, phenomenal return indeed.
Thanks John for a cracking strategy
Thanks also to the three BTC boys for bringing it all together, what a brilliant community this has turned out to be
jesus i made 2.5 points, its incredible how different people get different results, have to say its getting a bit frustrating, always seem to be 2 steps behind
-
Brilliant Feb for John’s SOL strategy folks eh? I ended the month just over 14.25 points up laying to a one point liability, and I missed out on a few winners as well, phenomenal return indeed.
Thanks John for a cracking strategy
Thanks also to the three BTC boys for bringing it all together, what a brilliant community this has turned out to be
-
@nick-allan
If you remove that single selection condition the backtest software will treat it like you bet on every horse that matches the filter as it falls into the odds range in that 10 min period. You cant trade it like the backtest does with that condition because until the race has started its not known how many horses will fully meet those criteria (including the odds). My guess is other people are using all horses or the first one that meets the criteria or if more than one is a match at that exact time (so on the ten mins) then dont bet, Or tagging every horse that meets the criteria and actually placing the bet or otherwise closer to the start. There are so many possible ways you can actually try to implement that rule without actually managing to implement it like the backtest.Pick one and see how it goes or do what i did and remove the condition so i know the rules i am actually placing bets by are the exact rules the backtest is evaluating its past performance on.
-
@jonathan-jones yeah if you are giving 3 selections how do you decide which horse to trade? of course if 2 are outside the odds range then thats fine but what if all 3 are in the odds range? do you place a back bet on all 3. Just trying to figure out how i set up the bot on BA Guardian
-
Pretty much yep.
The rules in the preset have a condition (under the odds) that it will only trade if only one horse out of the selections matches the criteria and odds. Obviously that not something you can replicate so not sure why its there. Be ok if the time was a shorter range but not over 10 mins. The 3 fav ones is similar. 10 mins is just too long to be making that judgement.If you remove both those you will see its still very good but the performance has dropped off in recent months. From there its up to you. Removing AW from the race types helps, re working the course selection (now you have removed AW) will help more etc.
You dont 'need' to do any of this, its a profitable strategy. I was getting geared up for a portfolio experiment so for me it was important i know i can trade the same rules as a backtest (so get the same trades/results) as the backtest. You can trade it as it is but you you wont get the same results as the backtest (or probably anyone else trading it) because of these little ambiguities.
-
@david-milligan said in The BTC Horse Racing Thread:
@karl-pick yep up about 25% for me since the start of Feb roughly
How are you trading it david? Are you only placing a bet if the horse selection is the 1st, 2nd or 3rd fav? Or just a straight odds selection (2-7) criteria? And are you trading all selections that the daily selections indicate as long as they between the specified odds range?
-
@karl-pick said in The BTC Horse Racing Thread:
@nick-allan said in The BTC Horse Racing Thread:
@jonathan-jones Tweak it in what way? Take the 1st, 2nd and 3rd fav out of the automation and just a straight odds selection (2-7) criteria based on the selections?
I use price only....2 to 6.8 as my guide not on whether its Favourite or Outsider.
thats probably why the backtest software isnt the same live would you not say? as you could have a selction that 3rd fav in the morning but 5th fav by the time the race starts and therefore the sortware wouldnt place the bet?
-
@karl-pick said in The BTC Horse Racing Thread:
@nick-allan said in The BTC Horse Racing Thread:
@david-milligan yeah the back the fav had a good finish, -3 for the month but it seems to be a good strat, not sure about the c and d winner ive just downloaded it for testing but i dont know the rules for the strategy? is it just backing the selection you get on on the daily selections? not like back the fav where you just back the fav
in say the 14:00 at kemptonSince January 1st ---- C+D £37.13...per £1 stake for me anyway.
How are you trading it karl? Are you only placing a bet if the horse selection is the 1st, 2nd or 3rd fav? Or just a straight odds selection (2-7) criteria? And are you trading all selections that the daily selections indicate as long as they between the specified odds range?
-
@nick-allan said in The BTC Horse Racing Thread:
@jonathan-jones Tweak it in what way? Take the 1st, 2nd and 3rd fav out of the automation and just a straight odds selection (2-7) criteria based on the selections?
I use price only....2 to 6.8 as my guide not on whether its Favourite or Outsider.
-
@karl-pick yep up about 25% for me since the start of Feb roughly
-
@jonathan-jones Tweak it in what way? Take the 1st, 2nd and 3rd fav out of the automation and just a straight odds selection (2-7) criteria based on the selections?
-
@nick-allan said in The BTC Horse Racing Thread:
@david-milligan yeah the back the fav had a good finish, -3 for the month but it seems to be a good strat, not sure about the c and d winner ive just downloaded it for testing but i dont know the rules for the strategy? is it just backing the selection you get on on the daily selections? not like back the fav where you just back the fav
in say the 14:00 at kemptonSince January 1st ---- C+D £37.13...per £1 stake for me anyway.
-
Today is the first day of going live with my experiment of using all the strats as a portfolio. Spent the last month (almost) in sim mode getting it to a point where the automation is working and the backtests match the previous days results (minor differences on back the fav but they are now very minor). So obviously had to tweak a few of the strategies. Going to give it 50 quid and a 1% compounding. Analysis suggests its moderatley risky but not terrifying (so the small bank) but should give a steady yet noticable growth.
So the strats taking forward>
Back C&D winner - I have removed the 1 selection and the fav conditions (needed so i can backtest and compare) but i have also dropped the AW and split the strategy. So i now have 2 strats with courses optimised for Turf and Chase. AW just has too much variance and adds too much risk to the portfolio.
Back Fav v3 - I have moved my bet time to 10 seconds after the official time to minimise inconsistency.
Back Pace - again removed AW and re-optimised the courses based on that.
Lay Beaten Fav - Removed the current fav condition and re-optimised courses based on ommiting the first few months. If you notice on the backtest for this its insanely profitable at the start of 2021. I beleive this skews the overal result so i adjusted the date for my optimisation. The final result is still very profitable on those out of sample months so i think its valid.
Low odds lays - removed the runners and classifications.
Lay 3-7s - My own based on laying horses that have not won on the course or distance at odds of 3-7.The backtest (fixed stakes) gives this chart:
An interesting thing about it (you can see from the trend lines i have added) is a distinct pattern of higher returns in the summer and lower (but still profitable) in the autumn/winter. This is expected since my strategies are more biased towards the flat.
It will be an interesting experiment. My results in live trading now match the backtests done the day after (in terms of trades taken, odds can be a tick or 2 off) with very few differneces using the amended strats so at least i can be confident had i started in March 21 and used these strategies that would be my result. Of course, things can change
-
Feb seems to of been a pretty rubbish month all round with only the low lays showing any real positivity. Expecting the pace backs to start pulling their weight this month so March shjould be better.
Be careful of Back C&D winner. Its not tradable in reality 'exactly' according to the rules so actual results will be very different from backtest results. 10 mins is miles and miles too long to even guess whats going to be top 3 at BSP and the backtest will only record a bet if there is only 1 horse that meets the selection and odds criteria (again, not something its possible to do in reality with a 10 min minimum timespan out from BSP).
Still shows profitable without those conditions though so fundementally its a good one i think, you just might need to tweak it a bit.
-
@matthew-sheehan sorry I meant the updated version, could you link it and I will post it to the OP of that thread, cheers
-
Not sure I'd done the same as others did, but the one I was using was the one I took from the main post on the bfbotmanager thread. https://forum.betfairtradingcommunity.com/topic/2915/bfbotmanager-automation-discussion
It's no problem, I just went in the strategy in bfbotmanager and added ireland and the courses - still learning with it but getting there
-
@david-milligan ah yes it would appear so
-
https://betfairtradingcommunity.com/en/trading-strategies/back-course-distance-winner
I think we back on the selection from the software as long as it's 1st/2nd/3rd favourite
-
@david-milligan yeah the back the fav had a good finish, -3 for the month but it seems to be a good strat, not sure about the c and d winner ive just downloaded it for testing but i dont know the rules for the strategy? is it just backing the selection you get on on the daily selections? not like back the fav where you just back the fav
in say the 14:00 at kempton -
Cracking day for the backing strategies today!
