I'm working on a new Greyhound lay automation. I've got one version letting the lay bet ride, and another that greens up to SP. Both are placing the original lay at the same stake and price.
1000 races in and I'm seeing an inverse relationship between letting the lay bet ride, and greening up to SP. I can't get my head around why that might be.
Any ideas?
Illogical inverse relationship between letting positions ride and greening to SP
-
jamesedwards
- Posts: 6288
- Joined: Wed Nov 21, 2018 6:16 pm
You do not have the required permissions to view the files attached to this post.
nothing obvious i can determine, i'll go with spurious correlation for now as you still have periods where it's not correlated.
if there was some sort of fast player spiking price at off then you would imagine BSP would be significantly lower on those dogs which you layed that actually won, however as you're taking a profit on lay bet to back bet BSP it suggests drifters are often winning
ive got a lot of 1,000 mkt live samples of gh mkts and it's not enough to draw much of a concrete conclusion about anything.
on another note takesp is just gonna make P&L worse as anywhere you force a bet into the mkt just for the sake of reducing liabilities and the more size you're just making your price worse. of course when you're playing for tiny amounts hedging is possible without giving too much profit back using BSP but you'd likely be better off to take the variance on the chin and have higher underlying +EV on every market, scale up bet size over time. there's only so many pounds you can put through those markets anyway so the max drawdown should be absorbable for you (if not just limit stakes as needed until you can afford to up it)
if there was some sort of fast player spiking price at off then you would imagine BSP would be significantly lower on those dogs which you layed that actually won, however as you're taking a profit on lay bet to back bet BSP it suggests drifters are often winning
ive got a lot of 1,000 mkt live samples of gh mkts and it's not enough to draw much of a concrete conclusion about anything.
on another note takesp is just gonna make P&L worse as anywhere you force a bet into the mkt just for the sake of reducing liabilities and the more size you're just making your price worse. of course when you're playing for tiny amounts hedging is possible without giving too much profit back using BSP but you'd likely be better off to take the variance on the chin and have higher underlying +EV on every market, scale up bet size over time. there's only so many pounds you can put through those markets anyway so the max drawdown should be absorbable for you (if not just limit stakes as needed until you can afford to up it)
At a very granular level, the following scenario (or equivalent) is unfolding.
- Lay one unit of stake at decimal odds β (liability β − 1).
- Let α be SP
- Let Y = 1 if the selection wins, 0 otherwise.
The back that greens the book is β/α. Both outcomes then pay the same amount:
g = 1 − β/α {g: 'Green-At-SP'}
Let-It-Ride, same opening lay, held to settlement:
r = 1 − βY {r: 'Let-It-Ride'}
which is +1 if the selection loses and −(β − 1) if it wins. Subtracting gives the residual:
r − g = β(1/α − Y)
The right-hand side is exactly the settlement of a fresh lay of stake β/α struck at SP. So, on every race,
Let-It-Ride = Green-At-SP + (Lay-At-SP held to the result)
All divergence between the bots lives in that residual (lay at SP, held to the result).
- Lay one unit of stake at decimal odds β (liability β − 1).
- Let α be SP
- Let Y = 1 if the selection wins, 0 otherwise.
The back that greens the book is β/α. Both outcomes then pay the same amount:
g = 1 − β/α {g: 'Green-At-SP'}
Let-It-Ride, same opening lay, held to settlement:
r = 1 − βY {r: 'Let-It-Ride'}
which is +1 if the selection loses and −(β − 1) if it wins. Subtracting gives the residual:
r − g = β(1/α − Y)
The right-hand side is exactly the settlement of a fresh lay of stake β/α struck at SP. So, on every race,
Let-It-Ride = Green-At-SP + (Lay-At-SP held to the result)
All divergence between the bots lives in that residual (lay at SP, held to the result).