Historical Data

[Shaun]

I have downloaded some data from betfair and was wondering if any one knows of a free program or excel sheet to do searches to understand the information better.

[Smokin Joe]

Shaun,
Don't think you will find anything out there to understand it. It is quite simple really.

If you want to open the data in Excel you (probably) need Office 2007 as there are more than 64,000 rows of data.

If you get it opened in Excel, ask any questions here about anything you don't understand.

[Shaun]

Got it open ok, only using the racing info for australia just not sure what to do.

I want to be able to search each race and check the last price traded so i can test a laying idea i havE, i only need info on 3 runners max in each race plus the resul of the race so i can work out the stakes.


Maybe i am asking a bit much of the data i have.

[manager1]

Shaun,
Some Chrome advice

The data is time stamped with the price first taken and last taken, but you need to know the last price taken. Therefore, it's a bit confusing to try and get the last traded price.

I would also suggest that there are huge gaps of missing data in the historical data, when I first noticed this, I started recording in excel and comparing it with the historical data forward. There was quite a bit of it missing, matched amounts and prices that never make it into the data.

So for lay purposes, you can either access the historical SP prices.

http://promo.betfair.com/betfairsp/SP_history.html

Or you could set up a macro to save your worksheet at suspension of the market.

[Smokin Joe]

The big problem with the Betfair historic data (in terms of getting the last price matched pre off) is that Betfair's "Keep" Bet has buggered up the data.

The data is flagged as either PE or IP (Pre Event of In Play) this relates to when the bet was submitted, not matched.

Betfair data has a line for each odds for each flag, so that 1.01 will have 2 lines a PE and IP flag (assuming that bets at 1.01 have been matched against both a bet submitted pre event and also in running).
Each line shows the price that it was first matched and last matched.

The problem with the Betfair data is that there are plenty of bets that are flagged as PE (placed before the event) but are subsequently matched in running - so they will have a last matched time of later than the off time. This means that the visibility is lost of the last time a bet flagged as PE was actually matched pre off time if it has been matched after the off.

That fact is what makes the Betfair historic data less useful than it should be, and is probably the reason why manager1 thinks that a lot of the data is missing - I would wager that the odds that are "missing" has PE bets matched after the off.

The most infuriating thing is that the problem can be easily fixed - by making the PE and IP flag relate to when the bet is matched, and not when it is placed.
Unfortunately betfair say that if they change the data it will inconvenience those who currently use the data

I would have thought that current users of the data would have liked to be able to accurately tell what the last price a selection was matched at at the off

[Captain Sensible]

Got to agree the betfair data is pretty much useless as the timings of when bets were matched is misleading. Doubt they'll ever change it as it'd be competing against the likes of fracsoft who pay them a healthy fee to gather and sell the data.

You might want to switch to the BFSP data at http://www.betfairpromo.com/betfairsp/prices/

I've found it useful for the in play minimums and maximums and it throws up some interesting stuff but I guess if you want pre off stuff you'll have to use the existing data files.

bit of a pain there's 457 files but you should be able to automate something to scrape the site and add them to a database

[petsol]

Is that possible that Betfair rebuilt his data? because I have not found any data in the database described as below. if a row is marked PE there is no row that has a later latestTaken than the actualOff. Please someone download a file and check to verify this. Or am I missing something?

The problem with the Betfair data is that there are plenty of bets that are flagged as PE (placed before the event) but are subsequently matched in running - so they will have a last matched time of later than the off time. This means that the visibility is lost of the last time a bet flagged as PE was actually matched pre off time if it has been matched after the off.