• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Benter 2 Step Example

MySportsAI

MySportsAI

Filly
I thought I would post an example of a set of tissue prices derived using Bill Benters 2 step process ie - fundamental features used to create a rating eg trainer jockey etc as a first step followed by a second step of regression suing the market odds. The idea is that the splitting them up like this stop the algorithm getting swamped if bot steps were executed in one step. The algorithm would take little notice of the fundamental features because the market odds are so dam good at predicting win probabilities. Anyway I have chose for no good reason other than it looks like a contentious races (see my other thread on this) namely the 18:30 at Redcar

horseTissue
Always Blue
3.503274​
Ceinture D'orion
6.656624​
Yafaarr (IRE)
7.142806​
Sweet Kiss
7.448556​
Chesneys Charm
8.032017​
Whiskey Pete (IRE)
15.09425​
Boubyan (IRE)
16.15873​
Cosmos Raj
26.71438​
 
FWIW, here are my tissue price(s) for that race.

The Value Efficient confirmation is based on the ARAZI91 ARAZI91 model/market coefficients that incorporate the live market.

Bill Benter's second, and successful try, at Hong Kong betting, occurred when he took the live market into account, just before the off.

He also got some favourable rebates too!


1787395252962.png
 
Well done. :handgestures-thumbup:

Which one didn't get matched? What were the net odds?

Edit: I can work out the net odds if I know which selection didn't get matched. My Calculator only goes up to 6 selections!
 
Last edited:
BC said:
Well done. :handgestures-thumbup:

Which one didn't get matched? What were the net odds?

Edit: I can work out the net odds if I know which selection didn't get matched. My Calculator only goes up to 6 selections!
Click to expand...
The one that came second, for a moment it looked like spoiling the party
 
Ah, so not one of the 7. My calculator only goes up to Dutching 6. So I asked Google's AI

From scenario 2 (Dutching), the average odds were 1.22.

And you do this close to the off? No wonder it needs to be automated.

How do you split the stake? Dutched?




***********************

AI Overview

How your £100 total stake behaves depends entirely on how you choose to split the money across the 7 horses.

The two most common strategies are Equal Stakes and Dutching. Here is how the profit/loss and net odds work out for both:


Scenario 1: Equal Stakes (Dividing the £100 evenly)
If you place an equal bet of £14.29 on each horse (£100 / 7), your net profit or loss will depend on exactly which horse wins:

    • Horse at 3.65 odds: £47.86 loss (Total return: £52.14)
    • Horse at 7.00 odds: £0.00 break-even (Total return: £100.00)
    • Horse at 7.40 odds: £5.71 profit (Total return: £105.71)
    • Horse at 8.40 odds: £20.00 profit (Total return: £120.00)
    • Horse at 17.00 odds: £142.86 profit (Total return: £242.86)
    • Horse at 18.00 odds: £157.14 profit (Total return: £257.14)
    • Horse at 29.00 odds: £314.29 profit (Total return: £414.29)
    • Net Average Odds: If you look at the raw average of all 7 individual prices, the average is 12.92. However, because your money is split 7 ways, your effective net odds on the total £100 stake if a horse wins varies from 0.52 (for the 3.65 horse) up to 4.14 (for the 29.00 horse).


Scenario 2: Dutching (Proportional stakes for equal profit)
If you adjust your stakes so you make the exact same profit no matter which of the 7 horses crosses the line first, your stakes will look like this:


Selection (Odds)Distributed StakeTotal ReturnNet Profit
Horse 1 (3.65)£33.42£121.97£21.97
Horse 2 (7.00)£17.42£121.97£21.97
Horse 3 (7.40)£16.48£121.97£21.97
Horse 4 (8.40)£14.52£121.97£21.97
Horse 5 (17.00)£7.17£121.97£21.97
Horse 6 (18.00)£6.78£121.97£21.97
Horse 7 (29.00)£4.21£121.97£21.97
    • Net Average Odds: In a Dutching scenario, your 7 selections function as one combined bet. The net decimal odds for the entire bet together is 1.22 (returning £121.97 on a £100 stake), which equates to fractional odds of roughly 2/9.
 
BC said:
And you do this close to the off? No wonder it needs to be automated.
Click to expand...

Not sure how any one else would do it, but for mere mortals like myself, I'd have my excel sheet at the ready, input the live odds as close as possible near the off into the sheet.

Let it cook:

The £1000 would be the running bankroll.
Column Y is the staking for each runner. Y5 = Total Stake.
Column Z is the results. Z5 = Max liability @ % of bank.
Column AA are the odds available. AA5 = Total odds for a winning coup.
Column AB is the profit/Loss. AB5 = Profit after 2% commission.

1787466250072.png

After inputting the live odds near the off, I'd input the staking into Betfair @ BSP. The theory is, the live odds being used are inferior to the exchange odds, so accepting BSP, if a win, then you could end up with a brucey bonus, or egg on face.

Yesterday would have been just normal payout, Last show 20/1, BSP 21.33.

The overall winning coup after 2% comms would have been 3.05. (£6.90 Staked v Profit £14.14)

----


BC said:
How do you split the stake? Dutched?
Click to expand...

Dutched.

---

Bill Benter had/has a syndicate of people with live access to the HK tote ticket machine. Once the live odds were input, the instruction to stake were done pretty much instantaneously, including all the exotic bets. The HK jockey club initially gave him the computers and rebates for being such a big customer.

----


On another sheet, I have a staking plan that covers 20:80 variable staking, if there's a single stand-out value runner. Needless to say, ARAZI91 ARAZI91 posted that one up after I was flailing around with a very basic one I was using.

----
 
You must log in or register to reply here.
Back
Top