Kelly calc - sort of

Don't chase your losses, it doesn't work. You will eventually bust your bank.
Post Reply
User avatar
megarain
Posts: 2040
Joined: Thu May 16, 2013 1:26 pm
Contact:

A friend (bookie) has asked me for some advice.

He has a punter who is backing 6 horses a day at Sp, some singles, but a lot of trebles. Lets say the Betfair price at Sp, is (after comms etc), 12% better than Sp. (There is no settlement risk)

The friend has a bank of say 400,000

The average bet by the punter is 3000. At odds normally between evens and 2/1.

Is there an optimal amount he should hedge? I sort of feel the punter has to lose, but some of his trebles quickly become v large liabilities.

So, feel, he should hedge maybe 2500 per event, at either Betfair Sp, or hitting the market just before they jump, to maybe further depress the actual Sp.

Is there a calculation that defines optimal staking ?

Thx
ly18
Posts: 5
Joined: Wed Aug 29, 2018 1:19 pm

Hi there

An old mate of mine sold his chain of Bookmakers many years ago for a fantastic valuation, he had a similar "Punter", she got very lucky one day.

A relatively complex maths and stats model could be built to manage Hedge Staking.
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

megarain wrote:
Tue Mar 07, 2023 4:04 pm
A friend (bookie) has asked me for some advice.

He has a punter who is backing 6 horses a day at Sp, some singles, but a lot of trebles. Lets say the Betfair price at Sp, is (after comms etc), 12% better than Sp. (There is no settlement risk)

The friend has a bank of say 400,000

The average bet by the punter is 3000. At odds normally between evens and 2/1.

Is there an optimal amount he should hedge? I sort of feel the punter has to lose, but some of his trebles quickly become v large liabilities.

So, feel, he should hedge maybe 2500 per event, at either Betfair Sp, or hitting the market just before they jump, to maybe further depress the actual Sp.

Is there a calculation that defines optimal staking ?

Thx
I think a Kelly-style calculation would involve making a judgement of both, the ISP and the true chances of the treble. Both would be highly speculative but they should give a rough guide to an appropriate liability.

£3K treble at evens, 6/4 and 2/1 wins £44K, so he's betting at odds of 1.07 approx.

If he thought his bet had a 1.06 of winning, Kelly suggests risking about 11% of his bank which happens to be spot on. :shock:

But off the top of my head, I'd be inclined to hedge half of it, or maybe all of it if there's enough price gap between ISP and BSP. Perhaps I'm not being entirely mathematical but I wouldn't want to see two of those trebles come in in the same week. :)

My understanding of the Kelly Criterion is that it's only applicable if you can continue to bet indefinitely. Your mate may not know if the punter intends to bugger off after a few winning bets.
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

Just thinking out loud....
He doesn't have to accept a stake he's not comfortable with (unless he's not into bookie stake restrictions)

What does he do of someone comes in and wants 3000 on a horse at 8/1 or 12 grand on a 2/1. (or whatever the treble is priced at?)

He doesn't need to worry about hedging until the 1st (and maybe 2nd) has landed..... Maybe that's it, pocket the guys money when the 1st or 2nd loses, fully hedge even to breakeven/small loss if it gets to the last leg. Although getting enough on the last race might be tricky however he approaches it so only accept these on big markets where you can get a decent amount at bsp?

There's no accas on Betfair so commision has to be paid on each leg if he's covering from race 1.

...... Absolutely no concrete advice there Megarain but sometimes a comment sparks an idea. :).
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

I heard a funny story on the old Channel 4 Morning Line about a bookie who got absolutely hammered by a massive bet.

His motto was, hedges are for gardens. :lol:
User avatar
megarain
Posts: 2040
Joined: Thu May 16, 2013 1:26 pm
Contact:

Thx for the replies.

He pretty much has to accept every bet - and I don’t see any reason why he wouldn’t - the bets are always Sp and knows enough on-course guys to send money to the track if
necessary.

He got the punter as punter got tired of waiting 20 mins in
betting shops for bet acceptance.

The current model of most high street bookies is not
accommodating to fairly large staking individual who just want to gamble.

I feel a set percentage to hedge would be best.

Invariably, if you pick and choose what to hedge, you end up talking to yourself on the drive home.

Modern on-course bookie software has features where
you can simulate hedging into an exchange. (98% of on-course
hedging bookies end up paying the exchange chunks over a year, and all exchanges want bookie action both from a liquidity standpoint and because they lose).

I knew one operation on-course who had the figures analysed and decided to simulate the hedging, so 60% went to an exchange (cheapest comm deal, not always Betfair), and 40%
was personally hedged by 3rd parties.

After 2 months, they had to pack it in - variance is a bitch - and thou u know over the year you are better off not hedging,
mentally it can be exceedingly tough.
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

thx mega.

Personal situation and risk aversion are the biggest factor in what he'll decide, and his margins. Variance isn't just tough it can be a killer unless you're doing a LOT of horses instead of a couple of dozen a day.
Anbell
Posts: 2004
Joined: Fri Apr 05, 2019 2:31 am

megarain wrote:
Wed Mar 08, 2023 1:45 am
Thx for the replies.

He pretty much has to accept every bet - and I don’t see any reason why he wouldn’t - the bets are always Sp and knows enough on-course guys to send money to the track if
necessary.

He got the punter as punter got tired of waiting 20 mins in
betting shops for bet acceptance.

The current model of most high street bookies is not
accommodating to fairly large staking individual who just want to gamble.

I feel a set percentage to hedge would be best.

Invariably, if you pick and choose what to hedge, you end up talking to yourself on the drive home.

Modern on-course bookie software has features where
you can simulate hedging into an exchange. (98% of on-course
hedging bookies end up paying the exchange chunks over a year, and all exchanges want bookie action both from a liquidity standpoint and because they lose).

I knew one operation on-course who had the figures analysed and decided to simulate the hedging, so 60% went to an exchange (cheapest comm deal, not always Betfair), and 40%
was personally hedged by 3rd parties.

After 2 months, they had to pack it in - variance is a bitch - and thou u know over the year you are better off not hedging,
mentally it can be exceedingly tough.
Why not make 12% risk free by betting 100% on BSP?
User avatar
megarain
Posts: 2040
Joined: Thu May 16, 2013 1:26 pm
Contact:

That’s a fair question.

But the punter logically has to lose and stop (thou funds seem
unlimited).

So, you would want to maximise etc before the golden eggs
got cooked.
Anbell
Posts: 2004
Joined: Fri Apr 05, 2019 2:31 am

megarain wrote:
Wed Mar 08, 2023 6:51 am
That’s a fair question.

But the punter logically has to lose and stop (thou funds seem
unlimited).

So, you would want to maximise etc before the golden eggs
got cooked.
Yeah, but unless you yourself have an edge that beats BSP, and BSP is basically 100% overround, and that you can get it all on at BSP without impacting it too much (and I'm confident that is true, given what you've said (a few thousand pounds at shortish odds)) then you've already got your golden egg. There's no extra value in keeping the bet in-house, and all the costs of stress + variance.

(I'm referring to single bets here. Accas might require some extra consideration)
Anbell
Posts: 2004
Joined: Fri Apr 05, 2019 2:31 am

Anbell wrote:
Wed Mar 08, 2023 5:51 am

Why not make 12% risk free by betting 100% on BSP?
When I say 100%, I mean, 94% or whatever it is greened-up
ly18
Posts: 5
Joined: Wed Aug 29, 2018 1:19 pm

SMarkets would be a logical Stake Hedging Platform.

I would think that with some form of optimised but more complex Stake Hedging, 20% plus could be made consistently on the 3K.

Interesting 'problem'.
sionascaig
Posts: 1053
Joined: Fri Nov 20, 2015 9:38 am

ly18 wrote:
Wed Mar 08, 2023 10:03 am
SMarkets would be a logical Stake Hedging Platform.

I would think that with some form of optimised but more complex Stake Hedging, 20% plus could be made consistently on the 3K.

Interesting 'problem'.
I've been scraping SMarkets prices (small sample so far) and they are way of exchange (even the "boosted" ones). If I could lay at those prices would take it like a shot.
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

Anbell wrote:
Wed Mar 08, 2023 5:51 am
Why not make 12% risk free by betting 100% on BSP?
BSP is 0ev so great in theory ..... except for the commission esp when covering multiples and you're paying on each leg.
Anbell
Posts: 2004
Joined: Fri Apr 05, 2019 2:31 am

ShaunWhite wrote:
Wed Mar 08, 2023 1:35 pm
Anbell wrote:
Wed Mar 08, 2023 5:51 am
Why not make 12% risk free by betting 100% on BSP?
BSP is 0ev so great in theory ..... except for the commission esp when covering multiples and you're paying on each leg.
You're agreeing with me again :lol:
Anbell wrote:
Wed Mar 08, 2023 6:59 am
(I'm referring to single bets here. Accas might require some extra consideration)
Post Reply

Return to “Loss recovery systems and methods”