Horse Racing Strategies Megathread

Martin

@nathan-cooke not much profit past 6 months on that one would be the only concern

Nathan Cooke

Hey team, had a bit of a play around and came up with a modification of 'back the fav'. Essentially, it will only back the fav if they've won 50% of their last 6 races. Feedback welcomed 0_1682090768685_rules_export (2).json

Adam

@jonathan-jones said in Horse Racing Strategies Megathread:

@adam

Can that be something that you can turn off please?

Yes it will be optional.

Martin

@tony-hastie let us know how testing goes!

Tony Hastie

Ive been playing around and come up with a "back the leader" strategy. Its very early days and I wasnt planning to share until I had a lot more forward testing under my belt but after 2 trades and 2 wins it is currently 23 points up! The only way is down from here so sharing at its peak
0_1681852326071_rules_export (1).json

Its basically backing a horse that has lead in 80% of its last 10 races. I intend to tweak it a lot more as I log data. Any input much appreciated

? Offline

@adam

Can that be something that you can turn off please? Thinking about how it might effect the drawdown figure (eg i always do a run at 1% bank) because using stake based on current bank tells me the max drawdown regardless of the start point of the strategy. Your change should make it more accurate at the top end of realistic staking but not sure how it might mess with the start point.

Andy Donnelly

@adam Thanks, actually just saw that on the race type when I hovered over it.

@John-Folan thanks as well. Will take that into account

Adam

@john-folan said in Horse Racing Strategies Megathread:

No please do that anyway. Maybe just put in a max realistic stake?

I'm thinking of setting the stake on each bet to the lower of desired stake / amount available at that price at the time.

Adam

@andy-donnelly said in Horse Racing Strategies Megathread:

All,

What’s the most reliable filter in the software to differentiate between flat racing and jump racing?

If you use the Race Type filter, you can select Turf, A/W, NHF for flat and Chase / Hurdle for jumps.

I'll add a results grouping for flat / jumps to make the profit breakdown easier.

John Folan

@andy-donnelly said in Horse Racing Strategies Megathread:

All,

What’s the most reliable filter in the software to differentiate between flat racing and jump racing?

Just doing some data analysis and wanted to see how some strategies performed against the two different options

Pace seems to matter more of the flat if that helps

Andy Donnelly

All,

What’s the most reliable filter in the software to differentiate between flat racing and jump racing?

Just doing some data analysis and wanted to see how some strategies performed against the two different options

David Milligan

Does this mean LTBF isn't going to make me £18 million in the next 18 months?

Dan MacKinnon

@adam that would definitely be useful. I’ll have to reign in my millionaire lifestyle however

John Folan

@adam said in Horse Racing Strategies Megathread:

I've always been a fan of compounding, but I think we need some extra controls in the software to make it more realistic. I have noticed that it's possible to generate some Mickey Mouse results when using compounding, which is probably mainly because it assumes that some quite ridiculous bet sizes are getting matched.

I'm going to see about adding some liquidity controls in this week.

EDIT: Just realised the discussion was over fixed stake vs liability, not necessarily compounding. Never mind :face_with_tears_of_joy:

No please do that anyway. Maybe just put in a max realistic stake?

? Offline

@andy-donnelly I did have a feeling fixed staking would being the highest profits. I'll track the strategies for a while and see if this theory comes good.

? Offline

@adam

That could be a real boost. At what i am staking currently its not going to be a big deal but as things get bigger its obviously going to make a difference. On my excel analysis i compund based on the days returns (so stake only changes on the next day) and i have set a max stake. Thing is i dont know what is realistic for that and the reality is its not likely to be the same on each race (or each strategy). On some races you could probably get a grand on with some of the strategies but on others you would be looking at low double figures.

Given we are all playing a numbers game with these stats based strategies a better way to more realistically model the stakes we can get on is probably quite important. On a related note i was thinking that other change thats been discused at length about bet placement time values for favourite and volume (obviously i am thinking about volume in this case) could be a massive help.

Adam

I've always been a fan of compounding, but I think we need some extra controls in the software to make it more realistic. I have noticed that it's possible to generate some Mickey Mouse results when using compounding, which is probably mainly because it assumes that some quite ridiculous bet sizes are getting matched.

I'm going to see about adding some liquidity controls in this week.

EDIT: Just realised the discussion was over fixed stake vs liability, not necessarily compounding. Never mind :face_with_tears_of_joy:

Andy Donnelly

Ah ok, yeh for forward testing adding an extra column would be the thing to do.

In my experience certainly on the strategies I am tracking, fixed staking almost always has the highest returns, but as discussed in past weeks the flip side of that is it’s probably more testing on mindset than other methods as you’ll have to deal with the biggest “single trade” losses.

? Offline

@andy-donnelly back testing is easy, it's the forward/current which is more tricky. I'll add a column to my spreadsheet "this is what you could have won"

Andy Donnelly

If you have built the strategy in the strategy builder, you can change the staking in the tool to analyse both sides. In my opinion the fairest way to compare both is by matching the av loss on fixed stake to the liability used on fixed liability.

If you’ve built something outside of the strategy builder then as long as you have recorded the odds of each trade you can use excel to help. Basically all the winning trades will be the fixed stake amount which is the easy part. Then for losing trades you can create a simple formula that sums the total loss of the trade using the fixed stake amount. Something along the lines of ((odds of trade - 1)*-fixed stake amount)