timeform in-running odds

hmmm - I was going to compare with the BF price file CSVs they publish online, but my link for that seems to have stopped working.

longbridge remove the slash at the end ie https://promo.betfair.com/betfairsp/prices

thanks Dave - bookmark amended!

and with that help - 2.3 is the number BF publish themselves, I guess TF just consume from them?

These were monetised years ago were they not, with the fidelity of the data depending on price.

How do you mean dusty?

Was it around 2016?
They changed it all and offered .json files instead.
The free stuff is unreliable, the more you pay the lower the refresh rate of data meaning it’s more accurate.
It’s on the historical section, something like 3-4K a year from memory for the full set horse racing

The limited data on the above link and duplicated on timeform should be accurate. It would be virtually impossible for it not to be as it is sourced by an scan not by someone typing figures.

btw as far as the OP is concerned, given the way the race was run, it is difficult to imagine 9 was matched for any amount.

But you can buy it costing more for the number of refreshes per second
It’s been a while since I looked, but I seem to recall they do a standard 1 second refresh
Or several times a second for a premium

You are bound to get more accurate representation the faster the refresh

that's why I remembered it ( not being buttons )

maybe some weird photo re-opening scooping so voided thing? But can't work out what.

The JSON is absolutely the more you pay, the better the resolution you get on the data.

The CSV files should still be as they were though; unless anyone has heard otherwise?  It's a very different data set to the JSON stuff.

I got matched at 8.4 during in play on Red Walls fwiw

around 1 1/2 furlongs out it hit the price Howard has reported

It was definitely matched at 9 turning into the straight when the runner up kicked clear. I laid some 8.96 and 6.2

The gloss on the BF price files says "only bets of a payout of more than GBP100 included" - though from the OP's comment that shouldn't be a factor here.

Though I have never worked out whether a price needs a single bet at such a payout to qualify, or that the total of bets matched at that price is filtered by that criterion.

yeah the 9 was in-running. Anyway there's a shedload of mistakes on what they put up on the Timeform page.

Interested whether those mistakes are just copies of BF data (in which case the blame lies with BF) or other things?

It says they arn’t responsible for the accuracy of the data

You can bet your arse that the more you pay the better it is

If I run a programme to look what the price is every second on every runner
It won’t look the same as the same programme running ten times a second

well 154 is traded volume isn't it so that's 77, it may be that there were no three figure sums fully matched at the bigger prices.

dusty that isn't relevant to this thread which is about timeform and https://promo.betfair.com/betfairsp/prices on which timeform is based

Do they not get it from the standard free stuff?

true kjansch but there will have been plenty of 3 figure bets at bigger than 2.3

dusty timeform is owned by betfair. The source of their prices are betfair and mirror https://promo.betfair.com/betfairsp/prices.

I know that Dave, which is why I’m asking

Why would bf offer for free a service as good as the one they charge for/
They know the value of data.

Different stuff dusty - the BF "historic data" is time-series so you can see what was matched when (at coarse, fine, and per-match intervals, depending what you pay) - I guess for training 'bots.

The SP prices files and the TF surfacing of it is just summary data, and they both long precede the new hostoric data files, so I doubt one is sourced from the other.

If it were me longbridge id run them off the same data

The SP max min would just be a line of script that looked at each runner off the slowest IR refresh data

I’d be very surprised if that’s not what they do

the simple stuff on promo etc is not related at all to the granular data. It is or should be accurate nonetheless. The highest price matched with £1OO+ payout is not something that should be inaccurate no matter where it comes from. It isn't interval data and should be the same no matter where it is sourced from.

This isn't "prices available" - which would obviously run the risk of inaccuracy if taken from less granular data.

Jan 5, 2024 -- 2:23PM, howard wrote:

true kjansch but there will have been plenty of 3 figure bets at bigger than 2.3

There was 203 eur looking to lay at 4.8, i can see about 40 of that matched, then 115eur at 3, which looks as though near to fully matched.
The price jumps around like a 'count the number of hops program' in between so its not certain that all the 115 was taken but that would be perilously close to 100 pounds on currency exchanges.
There are no other three figure prices until the horse goes odds on- It only hit 2.4 for 20 quid from the start of the race untl it was odds on from the first turn in the race until the home bend when it went out to 9. So no three figures there either.

Its more like a modelled set of prices for the horse with additional in play bets. This happens for a number of races daily where a fav will just collapse to half its odds just because its in a good position

£77 matched at 9.0 is a larger payout than £100, Kjansch, ergo, it should be included in the data.

The phrase used to decribe the data is "only bets of a payout of more than GBP100 included". It seems likely that this means that at least one single bet has to have a payout of more than £100. So if the £77 comprised of small bot staking and keep bets, there might not be an individual bet that met the criterion to be included in the data.

Though the main takeaway imo is that the in running liquidity is very poor if no substantial money was bet above 2.3, even though some players thought the horse was an 8-1 chance.

Fair one, Dave, though that is frankly ridiculous logic (from BF), if that is the case.

think it is just the IT bods doing what they were asked and when the spec for the scan was written it would be a rare occurence that there would be a big gap between the high £100+ and the actual high. But nowadays....