The BTC Horse Racing Thread
-
@Toyin-Butler It's usually closer to 08:30-09:00 tbh, just depends on when the results appear for the previous day's races.
-
The joys of ETL pipelines and eventual consistency. Is 10am the earliest you can run the jobs?
-
@Jonathan-Jones Generally the results are all in by about 10am, but they do get delayed sometimes. The import is a pretty intensive insert operation on the DB and any strategy queries that people may be running at the time are prioritised, so if the software is busy at the time then the import could take longer. More of a deadlock-avoiding issue than a performance one really.
Also the pace letters and ratings are updated as separate job just after all the results currently, so it could be possible to run a strategy that doesn't use pace and see results from the previous day, but then run a strategy that does use pace and not see results.
We're going to tighten all this up but sounds like we could do with adding some indicators to the software in the meantime to show when the last updates happened.
-
@Joe-Beeston I had a good look at this today- It seems quite profitable even where there is only one selection out of the two. If you scroll through the individual results of the preset, you’ll notice lots of occasions where it’s just the one. In forward testing, I’m going to be happy to take either one or two.
-
Thanks,
I will try that -
@Adam
Sorry, been out all day.I have just re-run the strats and the missing ones are there. Guess it must of just updated a bit later than normal today (or i checked a bit earlier). Seen this Last week too. I think stuff is wrong because some strats on some races are not right, re-run them an hour or so later and its all fixed.
-
@Jonathan-Jones said in The BTC Horse Racing Thread:
@Adam
Do we have missing data fromyesterday? I have a shed load of bets taken in reality (by both my live Fairbot strats and my BFBM test) that are not in the software. Some of these are not complicated ones either. If its between price a and price b place the bet sort of simple.I hadn't spotted any but will check in more detail, what sort of time-frame did they happen in?
Alternatively which horses and races (and strategy name)?
-
On the Lay 1st & 2nd filter, just in the initial phases of forward testing. What happens if 1 selection is within the odds range but the other isn't. For example, Fav is at 2 and the 2nd Fav is at 4. Would you lay the 1 selection or would you leave that race alone as you need to lay both the Fav and 2nd Fav for the strat to profit long term?
Thanks
-
@Jonathan-Jones I'd use the sections filter (selections conditions>selections filter) to do this... Also, for what it's worth, I have 5 starts running, with about 40ish bets yesterday, and every single one is in the software as it should be. So I'm not sure there is any data missing.
-
A question for @John-Folan (or anyone else who is good with BFBM).
I am using the tipping service method of getting my selections into BFBM but dont know how to set the criteria so it only bets if that imported selection is 1st,2nd,3rd,4th favourite.
I have this currently set up but not sure if its correct:
The strategy is set to look at the correct provider and that is working but i seem to of had a few more bets yesterday than the software. In fact i had a lkot more bets accross the board than the software so i think we may have the missing data issue again but since i wasnt sure about this setup anyway its not helped.
-
@Adam
Do we have missing data fromyesterday? I have a shed load of bets taken in reality (by both my live Fairbot strats and my BFBM test) that are not in the software. Some of these are not complicated ones either. If its between price a and price b place the bet sort of simple. -
@Jonathan-Jones looking forward to hearing how you get on
-
Going to be an interesting day today. First day of testing a 100% hands off automation workflow with BFBM (sim mode obviously). Surprisingly didnt take me long to create the strategies so i am sure there will be a million bugs to work through in the morning. Really wasnt sure about how tipster selections and favourite positions 1 to 3 etc work so only way it to try and and see what it does.
If it works then it should make my results much more consistent with the backtest results in terms of non runners for sure (and more profitable since that little problem is killing me at the moment). Then my last consistency bugbear is the favourites swapping in the last few seconds.
#Edit
Well, its running and so far all strategies appear to be actually doing stuff. Bets are being placed on selections at least. Will see in the morning if they are the right bets though.
-
@David-Hodgson
Try this:
Handicap and AW
Fave BSP > 3
Placing bets from race start actual to 10 seconds inplay.
Bet type is an offer @ odds of 1.5
rules_export(2).jsonYou get pretty much no trades with those criteria though. By guess is that fav odds over 3 in a handicap is pretty rare.
-
@Greg-Mitchell said in The BTC Horse Racing Thread:
@David-Hodgson You couldn't backtest this strategy in the software, because as far as I am aware, it doesn't facilitate "in-play" only betting
He hasn't mentioned in play in his post, but it's definitely possible to restrict your bets to in-play using the entry time slider.
-
@David-Hodgson you can certainly filter on all weather (race type), handicap races and favourite price (currently this will be based on the BSP).
You'd also want to set your bet type to "Offer Price" and then probably group your results by "Specific race".
You'll then see how many were matched in each race, along with overall profitability.
As you say, you can then play around with different price points.
-
@David-Hodgson You couldn't backtest this strategy in the software, because as far as I am aware, it doesn't facilitate "in-play" only betting. But you could use your three criteria to get all of the data out of the software and then analyse it off line in Excell (other spreadsheet products are available)
