My Racing System Stats

[account_deactivated]

Is this system based on form data or price data?

[B-Rex]

Is this system based on form data or price data?

No / No Let's just say it is related to 'volume' data and that's about all I can reveal.

Currently my live test is only actually doing the Back bets though, as the Lays only add about 10% extra profit, but means when I have a losing race i am losing a lot more, so it's a more swingy experience. I will look to bring those in, from month 3+ probably when I can consider my data a little bit more reliable.

Although, I am going up to the dizzy heights of £1.50 stakes today

I'll post the live results against turnover hopefully in the next few days

[Dublin_Flyer]

By the way, if anyone happens to know:
I'd be very interested to know how the last 12 months have faired in terms of number of 3-digit odds winners, versus other years.

I can look into it myself at some point, but if anyone happens to already know then it might save me a couple of hours poking around! Thanks

All Weather seems to have an increase in recent years and a higher variance.

[B-Rex]

By the way, if anyone happens to know:
I'd be very interested to know how the last 12 months have faired in terms of number of 3-digit odds winners, versus other years.

I can look into it myself at some point, but if anyone happens to already know then it might save me a couple of hours poking around! Thanks

All Weather seems to have an increase in recent years and a higher variance.

Thanks! That's really interesting

Also I just had a quick look at the live test results, and did the Profit / Turnover chart

[firlandsfarm]

A few years back, maybe 4 or 5, maybe more, as part of a review of BSP market efficiency I did a review of BSP rounded to tick points for all ticks and it surprised me how accurate they were to outcome. I looked at both laying as well as backing returns but couldn't see a pattern for profitability ... maybe I missed it. If you have found a way to profit from the higher odds runners well done.

[B-Rex]

A few years back, maybe 4 or 5, maybe more, as part of a review of BSP market efficiency I did a review of BSP rounded to tick points for all ticks and it surprised me how accurate they were to outcome. I looked at both laying as well as backing returns but couldn't see a pattern for profitability ... maybe I missed it. If you have found a way to profit from the higher odds runners well done.

Well just to be clear, i'm not looking at odds too specifically, as mentioned it is more related to where the money is/isn't going, using a very particular lense. As I say, I would highly doubt anyone is really looking at the data I am collecting, especially in the specific way I am looking at it.

It is just the way the system seems to play out, that results will go sideways and slightly upwards, until a big winner comes in, which will jump the profit up, which then maintains / more sidways until the next one, etc.

I am not taking EVERY high odds pick, as seem to average about 3 picks per race. But seems when one does come in, most of the time I have backed it. The exception being when we had a winner around 550 a couple of months ago, I had layed it it was only a small loss but the missed profit was a big shame!

I also try to filter out the ones that are very very high odds anyway because no way i can get enough data for those to consider any results reliable, probably for a few years to come, and largely just increases the bleeding on a day to day basis. I don't want to have to wait months for something epic to undo all of that pain, so I do have a cut-off. But the cut-off is based on odds close to the off, given that BSP odds cannot be read prior. So there are occasions where i've had very big winners that are above my limits, because the price drifted a lot in the last minutes. Those are the best!

[tico]

Hi Brex,
Well if you have twins you have two great reasons to succeed.... .
Kind Regards
Tico

[B-Rex]

A little update....So far, so good!

It's been going OK, making reasonable profit each month since April - and then had an absolute banger last weekend, pulled in over £1200 from £1-2 bets. In total my balance has gone from £250 -> £3.5K since April... all from £1-2 bets.

I would like to get some ideas for what to do now, stakes-wise.

If my system performed similarly all year round, I would simply be increasing stakes by £1 every time my balance goes up by £1k

However, whilst I am optimistic it will do OK over winter this time (now my system has been much improved from the iteration i was running last winter) - I would not be surprised to see my results become a lot more volatile, see bigger drawdowns etc.

So as we're getting to the back end of summer, I don't want to give this nice pot i've built back to the market during the winter. My plan is to go up now to £3 stakes and try to slowly grind my balance upwards from here to build as big a pot as possible over winter, and then come March/April when things should improve again, bump the stakes up to the maximum @ £1 per £1k balance, for whatever I have managed to build to over winter.

Do you think this is a bit too cautious though, given that you kind of have to make hay in this game, whilst you can, and maybe it won't even be any good next year, so maybe i should crank them up now to £5-6 for another month or 2 at least?

What would you do?

[jimibt]

A little update....So far, so good!

It's been going OK, making reasonable profit each month since April - and then had an absolute banger last weekend, pulled in over £1200 from £1-2 bets. In total my balance has gone from £250 -> £3.5K since April... all from £1-2 bets.

I would like to get some ideas for what to do now, stakes-wise.

If my system performed similarly all year round, I would simply be increasing stakes by £1 every time my balance goes up by £1k

However, whilst I am optimistic it will do OK over winter this time (now my system has been much improved from the iteration i was running last winter) - I would not be surprised to see my results become a lot more volatile, see bigger drawdowns etc.

So as we're getting to the back end of summer, I don't want to give this nice pot i've built back to the market during the winter. My plan is to go up now to £3 stakes and try to slowly grind my balance upwards from here to build as big a pot as possible over winter, and then come March/April when things should improve again, bump the stakes up to the maximum @ £1 per £1k balance, for whatever I have managed to build to over winter.

Do you think this is a bit too cautious though, given that you kind of have to make hay in this game, whilst you can, and maybe it won't even be any good next year, so maybe i should crank them up now to £5-6 for another month or 2 at least?

What would you do?

well done on that, genuinely impressive stuff. £250 to £3.5k off £1-2 stakes, and holding up live after all that data graft - that's the kind of result worth taking seriously rather than writing off as a hot streak. good on you for keeping it public too.

on the staking question - I'd say your instinct to bank thro winter is the right one. the thing i'd anchor to is your drawdown numbers rather than the balance.. size it so your worst historical drawdown (then doubled, cos the future always finds a deeper hole) is something you can sit thro without flinching. if you know your worst peak-to-trough in points, safe stakes kinda fall out of that on their own. out of interest, what's the deepest drawdown you've actually clocked live so far vs what the full dataset was suggesting? that'll tell you way more than the balance ever will.

one wee tip that saved me a lot of grief over the years - plot your equity curve with your best handful of results stripped out.. top 10, top 20, top 30. if it still climbs nicely then you've got something broad and robust. if it flattens, you know the engine is really just a small number of monsters, and you'd want to size a lot more cautiously for the dry spells in between. what does yours look like with the top 20 or so pulled out - still trending up, or does it go flat?

last one, worth doing as you scale up: keep a running log of the price you actually got matched at vs the BSP you'd have expected, esp on the bigger-priced runners. at £1 it won't show, but somewhere north of that your own money starts nudging the very prices you're chasing - and that's the silent killer for high-odds systems. seen any of that slippage creep in yet going from £1 to £2-3, or is it still filling clean?

anyway, keep the updates coming - one of the more interesting threads on here in a while. rooting for it

[B-Rex]

one wee tip that saved me a lot of grief over the years - plot your equity curve with your best handful of results stripped out.. top 10, top 20, top 30. if it still climbs nicely then you've got something broad and robust. if it flattens, you know the engine is really just a small number of monsters, and you'd want to size a lot more cautiously for the dry spells in between. what does yours look like with the top 20 or so pulled out - still trending up, or does it go flat?

last one, worth doing as you scale up: keep a running log of the price you actually got matched at vs the BSP you'd have expected, esp on the bigger-priced runners. at £1 it won't show, but somewhere north of that your own money starts nudging the very prices you're chasing - and that's the silent killer for high-odds systems. seen any of that slippage creep in yet going from £1 to £2-3, or is it still filling clean?

anyway, keep the updates coming - one of the more interesting threads on here in a while. rooting for it

Thanks, I appreciate the vote of confidence!

I've not done that specific test about removing top winners etc. However, my top 15 winners average odds are ~150 (and i do block bets above a certain point, don't want to be chasing 1,000 winners!)

But my main system figures in my data are currently totaling ~£480K (based on £50 back stakes, similar Lay equivalent) without even adding in July/Aug data yet - so those winners are only contributing ~£7.5K each, or 112K collectively, which amounts to less than 1/4 of the total. And obviously those stakes are silly, i just didn't want to use small numbers, to avoid rounding issues.

For your other suggestion... Difficult to measure BSP taken versus expectation - I'm not modelling anything to be able to get any kind of expectation, so can't really tell on that front. But like you said, at the stakes I am doing so far it's unlikely to make much difference thus far, but as I increase, it's possible it will. That is one of the reasons I am working on other variations on my system, such as one that focusses only on the favourites, and one just on low odds in general etc. so that there won't be much slippage.

One thing I just discovered today though, which is pretty disappointing - I didn't realise how extortionate 'Rest of World' markets are for commission. I had the same 2% comm applied to my ROW data which had similarly impressive results, but now i've updated it to be 9% average for AU and 6% for everything else, it's nuked those figures by 50%. On the flip side though, because there's so many more races, it's still looking likely to be returning similar to my UK main system. Currently showing £150K but just from 6 months data - so very early days - but the drawdown is decent, only 8k so far, versus 21k on the main UK system.

I have 7 in total on the go now, bust most of them only running for a couple of weeks. About £400 of that £1200 weekend was from the ROW system. Hopefully as 1 system has a bad run it will be made up for by some of the others (and not compounded!) too early to tell.

It is pretty funny though now, watching my software's race screen. When it places the bets, i get a mixture of about 15 bets on a 10 runner race, usually about 20% as Back bets. It's chaos.
Absolutely delicious last Saturday when it picked a winner @ 212 and layed every other horse

[jimibt]

one wee tip that saved me a lot of grief over the years - plot your equity curve with your best handful of results stripped out.. top 10, top 20, top 30. if it still climbs nicely then you've got something broad and robust. if it flattens, you know the engine is really just a small number of monsters, and you'd want to size a lot more cautiously for the dry spells in between. what does yours look like with the top 20 or so pulled out - still trending up, or does it go flat?

last one, worth doing as you scale up: keep a running log of the price you actually got matched at vs the BSP you'd have expected, esp on the bigger-priced runners. at £1 it won't show, but somewhere north of that your own money starts nudging the very prices you're chasing - and that's the silent killer for high-odds systems. seen any of that slippage creep in yet going from £1 to £2-3, or is it still filling clean?

anyway, keep the updates coming - one of the more interesting threads on here in a while. rooting for it

Thanks, I appreciate the vote of confidence!

I've not done that specific test about removing top winners etc. However, my top 15 winners average odds are ~150 (and i do block bets above a certain point, don't want to be chasing 1,000 winners!)

But my main system figures in my data are currently totaling ~£480K (based on £50 back stakes, similar Lay equivalent) without even adding in July/Aug data yet - so those winners are only contributing ~£7.5K each, or 112K collectively, which amounts to less than 1/4 of the total. And obviously those stakes are silly, i just didn't want to use small numbers, to avoid rounding issues.

For your other suggestion... Difficult to measure BSP taken versus expectation - I'm not modelling anything to be able to get any kind of expectation, so can't really tell on that front. But like you said, at the stakes I am doing so far it's unlikely to make much difference thus far, but as I increase, it's possible it will. That is one of the reasons I am working on other variations on my system, such as one that focusses only on the favourites, and one just on low odds in general etc. so that there won't be much slippage.

One thing I just discovered today though, which is pretty disappointing - I didn't realise how extortionate 'Rest of World' markets are for commission. I had the same 2% comm applied to my ROW data which had similarly impressive results, but now i've updated it to be 9% average for AU and 6% for everything else, it's nuked those figures by 50%. On the flip side though, because there's so many more races, it's still looking likely to be returning similar to my UK main system. Currently showing £150K but just from 6 months data - so very early days - but the drawdown is decent, only 8k so far, versus 21k on the main UK system.

I have 7 in total on the go now, bust most of them only running for a couple of weeks. About £400 of that £1200 weekend was from the ROW system. Hopefully as 1 system has a bad run it will be made up for by some of the others (and not compounded!) too early to tell.

It is pretty funny though now, watching my software's race screen. When it places the bets, i get a mixture of about 15 bets on a 10 runner race, usually about 20% as Back bets. It's chaos.
Absolutely delicious last Saturday when it picked a winner @ 212 and layed every other horse

ha, the 212 winner with everything else layed - that's the dream scenario right there, must've been a proper moment watching that one land.

the commission thing is a great catch, and honestly it's the bit that'd have me double and triple checking everything else. if a wrong rate was doubling the ROW figures, i'd be having a hard look at the UK 2% as well. the one that gets most winning UK accounts eventually is the premium charge - if you're consistently green across a lot of markets it can be 20% of profits (up to 40% at the top end), and the standard 2% doesn't cover any of it. worth modelling in now while it's cheap to find out, rather than it landing as a nasty surprise once you scale.

on the expectation point - fair enough you can't model a fair price, but there's a simpler version you can measure with no model at all: log the price you actually got matched at against the BSP that was eventually returned on that runner. no expectation needed, just filled price vs returned SP. that's the one that'll quietly tell you whether slippage is creeping in as you go up, esp on the bigger-priced ones.

and just a thought on the 7 systems - if they're all variations on the same core read, they might be more correlated than they look, so a bad patch could hit a few of them at once rather than one covering for another. worth keeping half an eye on how they move together.

anyway, still enjoying the updates - that race screen chaos sounds *interesting*.