Hi guys, just found out that after the endless hours I spent collecting data on in running horse prices it was a complete waste of time as the Timeform in running highs and lows are NOT CORRECT
Its amazing how I managed to get a trigger on the winner at Lingfield (3.0 bsp) in play when its last traded price was greater than 5.0 and actually traded a matched bet at 4.5 when at the highst point this horse traded at 3.5 according to Timeform, just amazing. This is not a one off, I've only been watching races in real time for a couple of days and this is consistent behaviour.
Just a heads up so you guys don't waste your time on collecting very inacurate data from this poor website.
Timeform in play horse prices
So the £63 matched today at 1000 won't count!
It should be based on layer's liability rather than matched stake.
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
As Dallas says, it's £100. It's a much better measure of a horses performance than showing random price spikes for a £1 which would be meaningless.
Admittedly £63@1000 isn't trivial. But then again it's not typical, so you have to set the parameters of an automated system somewhere. Seems to work fine in the majority of cases.
Admittedly £63@1000 isn't trivial. But then again it's not typical, so you have to set the parameters of an automated system somewhere. Seems to work fine in the majority of cases.
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
There's nothing like the facts for getting to the correct answer.
I always understood it to be a £100 payout ie. stake+winnings (calculated weirdly as per the Daq approach) and some time back (years) validated this using some data for odds-on losers in running taken from graphs straight after the race, and then comparing it with the Timeform IR-low subsequently provided on their data page.
Eg, if graph after race gives you:
Price Traded/matched
1.2 £150
1.3 £100
1.4 £150
Then, at 1.2, £75 staked at 1.2 (75+15 winnings =£90, less than 100, not included).
At 1.3 £50 staked at 1.3 (50+15 winnings = £65, less than 100, not included).
At 1.4 £75 staked at 1.4 (75+30 winnings = £105, =>100, given as price).
Certainly for long odds on it can make a significant difference based on a few tenners at very short prices, and a seemingly much bigger price given on the data page.
Maybe nowadays, accuracy has gone down the spout, costcutting/lapsadaisical ...or just changed the method
Eg, if graph after race gives you:
Price Traded/matched
1.2 £150
1.3 £100
1.4 £150
Then, at 1.2, £75 staked at 1.2 (75+15 winnings =£90, less than 100, not included).
At 1.3 £50 staked at 1.3 (50+15 winnings = £65, less than 100, not included).
At 1.4 £75 staked at 1.4 (75+30 winnings = £105, =>100, given as price).
Certainly for long odds on it can make a significant difference based on a few tenners at very short prices, and a seemingly much bigger price given on the data page.
Maybe nowadays, accuracy has gone down the spout, costcutting/lapsadaisical ...or just changed the method
That would make far more sense and make more appropriate data for traders working on in-play strategies, as not many people would want £100 on at 1000.Andriy wrote: ↑Sun Dec 23, 2018 4:34 pmI always understood it to be a £100 payout ie. stake+winnings (calculated weirdly as per the Daq approach) and some time back (years) validated this using some data for odds-on losers in running taken from graphs straight after the race, and then comparing it with the Timeform IR-low subsequently provided on their data page.
Eg, if graph after race gives you:
Price Traded/matched
1.2 £150
1.3 £100
1.4 £150
Then, at 1.2, £75 staked at 1.2 (75+15 winnings =£90, less than 100, not included).
At 1.3 £50 staked at 1.3 (50+15 winnings = £65, less than 100, not included).
At 1.4 £75 staked at 1.4 (75+30 winnings = £105, =>100, given as price).
Certainly for long odds on it can make a significant difference based on a few tenners at very short prices, and a seemingly much bigger price given on the data page.
Maybe nowadays, accuracy has gone down the spout, costcutting/lapsadaisical ...or just changed the method