The BTC Horse Racing Thread
-
@john-folan I appreciate that. As silly as it is, I was relieved that the software was still correct and my strategy hadn't taken a sudden unexpected downturn.
It felt like when you see people looking for their glasses but they're on their head the whole time! I was going through loads of different possibilities with the software but the issue was with the user all along. What do they say about a tradesman blaming his tools?!
-
@dan-mackinnon said in The BTC Horse Racing Thread:
I was banging my head against the wall and not understanding why one version of a bot in Fairbot was working fine on one VPS but I created another version on my laptop and getting different results.
Over the weekend I ran both together and got different results. So I imported both versions and looked at them side by side and the answer was obvious.
I forgot to put a rule in when I created the new version to cancel any trades from going in play. I've had such a rubbish month and it's been my own fault for missing something basic. What a pillock!
I’ve made so many silly mistakes over the years. Don’t beat yourself up too much.
-
I was banging my head against the wall and not understanding why one version of a bot in Fairbot was working fine on one VPS but I created another version on my laptop and getting different results.
Over the weekend I ran both together and got different results. So I imported both versions and looked at them side by side and the answer was obvious.
I forgot to put a rule in when I created the new version to cancel any trades from going in play. I've had such a rubbish month and it's been my own fault for missing something basic. What a pillock!
-
Thats pretty much the case. I dont have enough data to make a judgment either way but i was findong using LTP for the execution would result in some unmatched bets. Dont see any now i use best available but i do get some quite far off (tends tyo be the bigger lays on the top prices too). I added a second filter to fairbot so its LTP and best available but i think this might result in a few missed trades (not noticed any so far but logically there will be some).
-
@jonathan-jones said in The BTC Horse Racing Thread:
using the trigger on LTP and bet placement on best lay price i often get several ticks off the price.
Out of interest, are you getting better prices by using best available? I guess it's hard to judge, as you're probably getting matched every time using best available but not always with LTP...
-
@dan-mackinnon said in The BTC Horse Racing Thread:
I seem to have some conflicting results with Fairbot again. I'm not sure if it's something I'm doing that's changed but I'm going to pause it for the time being.
Looks like I might be joining the BFBM club when my subscription runs out.
-
I seem to have some conflicting results with Fairbot again. I'm not sure if it's something I'm doing that's changed but I'm going to pause it for the time being.
Looks like I might be joining the BFBM club when my subscription runs out.
-
@jonathan-jones they do but it’s not a bot. It’s more based on manual trading with ladders, dutching, etc.
-
Not really getting any differences in trades any more on fairbot. Getting a few odds differences and thats always because the software triggers in LTP and always places and gets the bet on LTP. I have found if i use LTP for bet placement i miss trades (perhaps this is part of your problem) but using the trigger on LTP and bet placement on best lay price i often get several ticks off the price.
Very much coming round the the idea its a swings and roundabouts thing. Seem to benefit as much as lose out from any differences i do get.
You do know that Bintenko also do a Daqbot to go with Fairbot right? Might try it myself one day.
-
@adam yeah. Also, the likelihood of that happening I assume is quite rare so it's not really a factor.
Overall it shows that both Fairbot and BFBM are pretty much bang on with the software. BFBM is probably slightly more accurate, whereas Fairbot picks up more trades around the 3 and 15 odds crossover.
-
@dan-mackinnon said in The BTC Horse Racing Thread:
Adam has told me that some of the data from Betfair takes a couple of weeks to show up on the software so there is a chance it has been updated, this is taken usually the day after).
The only reason there should be that much of a delay is if the Betfair API went down or something and we need to plug the data gap with their historical data, which is about a week behind.
Otherwise, the results for the previous day should be in about 36 hours max after the race.
-
I've gone back to the start of February when I was tracking my results vs the software.
Software: 167 trades
Fairbot: 194 tradesSoftware: 8 losers, 95% SR
Fairbot: 13 losers, 93% SR8 trades were on the software and not Fairbot, 19 were on Fairbot but not the software (Adam has told me that some of the data from Betfair takes a couple of weeks to show up on the software so there is a chance it has been updated, this is taken usually the day after).
I haven't worked out the P&L but it's roughly the same. It's starting to make sense why this month is so rubbish if I was going at 95% SR last month from 124 trades!
I know it's said here so often but I love the software because it gives us an opportunity to mess about with stuff like this without having to wait years to build up data.
-
@dan-mackinnon said in The BTC Horse Racing Thread:
This is my early doors analysis of the software results vs Betfair (Fairbot) vs Betdaq (BFBM) using LTBF.
Trades
Total 57 horses as per the upcoming qualifiers from 5th to 12th April- Software: 32 trades
- Betfair: 47 trades
- Betdaq: 43 trades
Losers
- Software: 4
- Betfair: 9
- Betdaq: 6
Strike Rate
- Software: 88%
- Betfair: 81%
- Betdaq: 86%
I know this short example makes it seem that the software is being misleading by avoiding a few losers, but last month it was the opposite and left out a few winners so overall it works out probably +/1 5% overall.
Overall, Fairbot seems to pick up more of the trades, especially around the 3 and 15 marks. Annoyingly it got matched at 2.98 and went on to lose. I would put some of the differences down to whether or not it got matched at 15.
I will keep records and when I get a decent size I'm happy to share more. I think that BFBM is probably more in line with the software results. It's just a shame I started recording this data during a losing month!
Interesting, I was just checking mine:
From the 4th March:
Software - 157 Selections
BFBM on Betfair(uploaded from Software URL) - 152 SelectionsSoftware - 145 "Winners", 12 "Losers"
BFBM - 140 "Winners", 12 "Losers"Both at roughly 92% Strike Rate
P & L to a £1 Stake
Software + £43
BFBM +£74Interestingly there were 20 on the Software that weren't on BFBM and 15 on BFBM that weren't on the Software
-
This is my early doors analysis of the software results vs Betfair (Fairbot) vs Betdaq (BFBM) using LTBF.
Trades
Total 57 horses as per the upcoming qualifiers from 5th to 12th April- Software: 32 trades
- Betfair: 47 trades
- Betdaq: 43 trades
Losers
- Software: 4
- Betfair: 9
- Betdaq: 6
Strike Rate
- Software: 88%
- Betfair: 81%
- Betdaq: 86%
I know this short example makes it seem that the software is being misleading by avoiding a few losers, but last month it was the opposite and left out a few winners so overall it works out probably +/1 5% overall.
Overall, Fairbot seems to pick up more of the trades, especially around the 3 and 15 marks. Annoyingly it got matched at 2.98 and went on to lose. I would put some of the differences down to whether or not it got matched at 15.
I will keep records and when I get a decent size I'm happy to share more. I think that BFBM is probably more in line with the software results. It's just a shame I started recording this data during a losing month!
-
I believe tomorrow will be 1 year since the horse racing software was released.
-
@nathan-cooke said in The BTC Horse Racing Thread:
Hey team, bit of a random one. I was digging online and found an article that suggested where there was a short priced fav, backing the 5th fav has a positive ROI.
We'll add this to the software. Probably makes sense to include nth fave with a max of 6 (6 being the max number of official "places" in any race).
-
Unfortunately the fav criteria only goes to 4th fav, otherwise it would of been an easy one to test.
-
@emily-knowles
Not seen the thread but you can easilty test this out yourself with the software if i get what you are asking. You can test laying to stake, liability by flat stake or by % bank.
-
@david-milligan I can't find the thread that's what got me thinking if I really did read it or not!
-
@emily-knowles I have read this somewhere too, can you post a link? Would love to see some data
