The BTC Horse Racing Thread
-
@Dan-MacKinnon said in The BTC Horse Racing Thread:
@Martin Yeah I'm always interested in forward testing ideas
here you go: 5fbacks.json
-
@Martin Yeah I'm always interested in forward testing ideas
-
@Dan-MacKinnon oh well if you want a copy let me know
-
@Martin No it wasn't. Not sure who did come up with that.
-
@Andrew-Gray thank you - will collect data and examine those SP's in detail.
I haven't got back text data simply because I'm downloading stats from the BTC Horse Racing Stats and other places to plug into my model, so its not one I've done on a filter to back test .. I'm still playing with versions for that, meanwhile this came along in my head and so I've dived into it - and so far so good. Will report back here when I have done another week or two of testing.
-
@Mark-Maguire You pretty much just want to find out the average odds of each selection on the place market then compare it to your strike rate on the place market. If you are getting 80% SR then you need av odds to be just above 1.25 to break even, accounting for a bit of commision. You can obviously do the same for the win SR and average odds, and for the 2nd rated horse at 70% place = 1.42ish odds needed to break even, and so on.
You might even find that say the 6th rated horse or something like that actually provides more value, as the odds might be better. Worth looking at everything not just the top 2?
I have all win and place Betfair SP data going back to around 2008, if you wanted to send me your back tested selections in CSV format I can probably knock up a back test pretty quickly for you if you want.
-
@Dan-MacKinnon that’s brilliant thanks
-
@Mark-Maguire From my research, you get a bit limited on options going below 3.0 other than straight bets. But if you wanted to mess about place market this might be useful https://www.aceodds.com/bet-calculator/each-way.html
-
@John-Folan thats the kind of thing I wanted to ask.. I'll admit to being a complete ignoramus when it comes to how to approach trading the horses.
I've been interested for a while, think it will work well long term to sit alongside football.
I've basically done a load of studying and, rightly or wrongly, decided to create a model which ranks the horses for every race I've made it easy enough in terms of time setting up a power query which then feeds into a pivot table and finally a probability model looking at about twenty different comparisons between runners and those split into positive and negative factors... ending in a ranked order .. I've been measuring paper trading/min stakes the results and am pretty happy so far on the basis that:
-
After the initial data download I can get the ranking done in 5 mins or so for each race.
-
Its finding winners and places with decent regularity - focussing on 1st and 2nd ranked in my model .. 1st ranked wins or is placed 80% (stress this is very early days) and 2nd ranked 70%.
So now as I continue to monitor and refine this its about how I look at trading it - I'm measuring odds and will try and get some stats of movements as the amount of data improves but in my ignorance its about understanding how to find the best way to trade - which is why suggestions like backing the place or dobbing will take me in a direction that I can learn .. and improve .
-
-
@Mark-Maguire so the stuff between 3.0 to 5.5 you might want to look at for dobbing?
-
@John-Folan the range has generally been between 2.02 and 5.5 with a couple of outliers (one at 55 which came in but I wont build anything hoping/expecting that happens every week
) -
@Mark-Maguire said in The BTC Horse Racing Thread:
@John-Folan thats great thanks will look at that and test for a little while
What sort of odd ranges are you playing with?
On the win side?
-
@John-Folan thats great thanks will look at that and test for a little while
-
@Mark-Maguire assume that's to win prices? may be worth looking at place ones
-
@Mark-Maguire said in The BTC Horse Racing Thread:
@John-Folan I'm looking for some advice please .. Its very early days but I've created a model for horse racing which early impressions are has something about it..
I'm plugging all runners into a stats model using BTC stats and a couple from elsewhere .. ranking the horses .
So far the strike rate of my 1st of 2nd favourite finishing in a place is really quite good...one of both of them finishing in place 80% plus .. I spoke with Ryan earlier but as I develop and test this want to consider the best way to trade it.. back to lay, or for example Ryan suggested dutching.. I've put a fair bit of work into the model but the clever bit will be how to trade the selections.. and help welcome
I’ve never dutched but would follow Martins comment. Backing the place is so under utilised and there is probably a lot of value there
-
@Martin brill will do thanks .. odds range so far from 2.8 to upper 4's with the occasional crazy outsider thrown in .. thanks as usual
-
@Mark-Maguire maybe have a look at odds ranges and how they perform, place backing could be big value if the odds are good!
-
@John-Folan I'm looking for some advice please .. Its very early days but I've created a model for horse racing which early impressions are has something about it..
I'm plugging all runners into a stats model using BTC stats and a couple from elsewhere .. ranking the horses .
So far the strike rate of my 1st of 2nd favourite finishing in a place is really quite good...one of both of them finishing in place 80% plus .. I spoke with Ryan earlier but as I develop and test this want to consider the best way to trade it.. back to lay, or for example Ryan suggested dutching.. I've put a fair bit of work into the model but the clever bit will be how to trade the selections.. and help welcome
