The BTC Horse Racing Thread

? Offline

@david-milligan
I think thats probably a good way forward. It 'sounds' good anyway.

I have some data on this (but not the double liabilty bit). I downloaded all the individual results for all the strategies including versions of the lays at liablity. The results might interest you.

With Normal Lays:

Biggest Loss day: -27.51
Biggest Win day: 60.71
Average Points per day: 4.5
Max DD: 47.1%

With All Liability Lays:

Biggest Loss day: -19.86
Biggest Win day: 55.46
Average Points per day: 2.98
Max DD: 38%

With Liability Lays (but short odds lays as normal):

Biggest Loss day: -19.65
Biggest Win day: 55.56
Average Points per day: 3.05
Max DD: 37.44%

This is at 1% compounding based on daily returns (not per race).
Shows 2 things you need to think about.
2% is bloody risky as hell. A flat 2% doesnt quite blow your account but the max drawdown is 76% (and, that occured a couple of days ago and after yesterdays results i think thats probably increased).
The short odds lays influence is marginal.

David Milligan

@jonathan-jones this is what I've done - I'm backing at 2% bank and laying the short horses at the same. The longer odds lays I have programmed to lay with liability up to 4% bank

David Milligan

@karl-pick they certainly are, I'm running them with 2% stake. I only started running the presets about 2 1/2 weeks ago fully automated, at the moment C&D winner, back the fave and short lays are all in profit. Lay strategies are down but as I said above I have been laying to stake so some big losers have disproportionately affected the balance. Pace backs down v slightly but only had 4 bets

? Offline

@david-milligan

I was thinking something similar in terms of balancing lays and backs in a portfolio. Laying to liability wont 'smooth' things out. It just makes everything on that side smaller and gives a larger bias to the performance of the backs. Though, at the moment its an odd one because profit is scewed to the performance of the backs but drawdown is dependant on those big lays. Was wondering if lay to liabilty of twice the stake of the backs or something like that would give a better equity curve.

? Offline

@david-milligan .....Course and Distance are doing very well at the minute, if your building.

David Milligan

@jonathan-jones I'm not too worried about the strategies themselves at the moment as they have good results over the long term and I haven't forward tested them anywhere near long enough to draw any conclusions.

The issue I have at the moment is balancing my greed and desire for my bank to grow exponentially, with realistic and conservative staking that will protect my bank and allow it to continue to grow over the long term. I need to find the sweet spot then trust the models and let them run for an extended period of time.

? Offline

@john-folan said in The BTC Horse Racing Thread:

@sam-ratcliffe said in The BTC Horse Racing Thread:

Hello all. I was wondering if anyone has ever looked at laying a specific jockey?

Rachel Blackmore…. Oh wait, that’s not what you mean

Made me chuckle!!!
Never a truer word said in jest!

? Offline

@david-milligan

The big lays are causing serious damage this month.

Thing is i am wondering if there isnt something more long term going on. LBF for example has had steadily reducing returns for a while. If you look at the red graph (with the trades/strike rate) you will see its average strike rate has been subtly declining for several months. Pace lays is similar though the change is much more subtly (apart from this month so could be vairance). Both are seeing their worst performance ever. Over almost 2 years of data, steadily declining performance with the worst performance ever now. Think they might be done (at least in their current form).

Fundementally they seem like good strategies so i am wondering what has changed. Any of the more experienced members want to offer some ideas because those graphs show more than just variance and its a lot more than just 'winter'. I experimented with re-optimizing the courses giving a bias to more recent data (so september onwards or last March onwards). Both do sort the problem but had a bigger negative effect on 2021 data than i hoped. Is that too much curve fit or is a bias to recent data a valid idea? Wondering because i think out of all the variables one that could change over time is course layout.

Paul Foster

@david-milligan said in The BTC Horse Racing Thread:

@chris-osborne I assume that the strategies will run similar to the past - I don't think that I'm overstaking (2% - could be too aggressive) and would like to see if the strategy recovers from the big losers. Having said that, laying to liability should smooth out some of the variance but ultimately the compounding will be slower. I can handle variance but don't want to be greedy if laying to liability is a no-brainer! Any overwhelming reasons you can see to lay to liability?

My understanding is that laying 2% stake at odds of 15 is a liability of 28% of your bank.

Laying at 10% liability would mean a low profit on high odds (c 0.7% at odds of 15) but would be a higher stake at lower odds (c 2% at odds of 6). Even 10% may be too high for some strategies. I used this example as the lower odds still hits the 2% you're currently trading at.

I've changed nearly all of the lay strategies I'm following to liability. I just tweaked the % of bank to lay in the software until I was at a comfortable level based on return v risk. Some are at 5% and one I'm playing with is at 15%. I tend to increase the % to the point that it blows the bank and wind it right back to a comforable level.

John Folan

@sam-ratcliffe said in The BTC Horse Racing Thread:

Hello all. I was wondering if anyone has ever looked at laying a specific jockey?

Rachel Blackmore…. Oh wait, that’s not what you mean

David Milligan

@chris-osborne I assume that the strategies will run similar to the past - I don't think that I'm overstaking (2% - could be too aggressive) and would like to see if the strategy recovers from the big losers. Having said that, laying to liability should smooth out some of the variance but ultimately the compounding will be slower. I can handle variance but don't want to be greedy if laying to liability is a no-brainer! Any overwhelming reasons you can see to lay to liability?

? Offline

@david-milligan said in The BTC Horse Racing Thread:

Got stung a bit today by continuing to lay to stake for the higher odds strategies. Going to continue to lay to stake until my road test is complete but will definitely be considering laying to liability once I'm up and running properly!

why wait?

Andy Donnelly

Thanks all so far. It’s a condition where I was ruling out horses that finished 1st or 2nd in their previous race. Obviously I was already ruling out the first place horses as it’s a beaten favs strategy but was also trying to rule out horses who had near misses in the previous race.

Dan MacKinnon

@andy-donnelly what is the condition?

David Milligan

Got stung a bit today by continuing to lay to stake for the higher odds strategies. Going to continue to lay to stake until my road test is complete but will definitely be considering laying to liability once I'm up and running properly!

? Offline

@andy-donnelly
I would say widening should be better but with recent experiments in optimisation timeframes forward testing is king whatever.

Andy Donnelly

After a bit of advice.

I’ve been following my own customisation of the lay the beaten fav strategy for the last 6 months, which has been profitable. However I’ve just noticed that one of the conditions I was applying was actually causing the strategy to earn less than it should. So removing that seems the logical decision.

Question should I remove this condition straight away?

Or should I run both variants side by side and not switch to the version with the condition removed until testing for a period of time?

My gut feel is that as it’s widening a criteria rather than narrowing, that the back test results is probably sufficient to move in this case, but interested in anyone’s thoughts before I change.

? Offline

Hello all. I was wondering if anyone has ever looked at laying a specific jockey?

Nick Allan

@adam thanks for the clarification adam

Adam

@nick-allan "Official" is the planned start time, "Actual" is the exact time the race actually started. Often races are delayed, especially the Irish ones with ridiculous numbers of runners.