Has anyone found any issues with 2016 SP data?
I'm pretty sure I loaded it properly (as per 2015,14,13 etc) and that my sheets are correct. But when I run some basic strategies the results are very different for 2016 vs those for 2010-2015.
Probably my bug but thought I'd ask before spending another session pulling my hair out.
NH or Flat, which is best for Laying The Field?
When I downloaded it last year it was fine (just checked and mine is around 116K rows once cleaned and merged)ShaunWhite wrote: ↑Thu Jan 04, 2018 3:29 pmHas anyone found any issues with 2016 SP data?
I'm pretty sure I loaded it properly (as per 2015,14,13 etc) and that my sheets are correct. But when I run some basic strategies the results are very different for 2016 vs those for 2010-2015.
Probably my bug but thought I'd ask before spending another session pulling my hair out.
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
Cheers buddyDallas wrote: ↑Thu Jan 04, 2018 3:52 pmWhen I downloaded it last year it was fine (just checked and mine is around 116K rows once cleaned and merged)ShaunWhite wrote: ↑Thu Jan 04, 2018 3:29 pmHas anyone found any issues with 2016 SP data?
I'm pretty sure I loaded it properly (as per 2015,14,13 etc) and that my sheets are correct. But when I run some basic strategies the results are very different for 2016 vs those for 2010-2015.
Probably my bug but thought I'd ask before spending another session pulling my hair out.
- northbound
- Posts: 737
- Joined: Mon Mar 20, 2017 11:22 pm
I imported 2015 and 2016 Betfair SP data (UK/IRE horses) a while ago into my database. Didn’t notice anything weird in 2016’s compared to 2015’s.ShaunWhite wrote: ↑Thu Jan 04, 2018 3:29 pmHas anyone found any issues with 2016 SP data?
I'm pretty sure I loaded it properly (as per 2015,14,13 etc) and that my sheets are correct. But when I run some basic strategies the results are very different for 2016 vs those for 2010-2015.
Probably my bug but thought I'd ask before spending another session pulling my hair out.
Sorry for the delay, here is the 2017 data:
https://www.dropbox.com/s/vkp2kg5nfr14d ... 7.zip?dl=0
Please note that the format will be slightly different as its the raw file I 'would' upload to my database, I have stopped loading SP data from here, instead taking it from the marketbook once a race has completed. The file size seems good and I only got one missing file (17012017) but use with caution as betfair don't seem to care about the quality of this data.
https://www.dropbox.com/s/vkp2kg5nfr14d ... 7.zip?dl=0
Please note that the format will be slightly different as its the raw file I 'would' upload to my database, I have stopped loading SP data from here, instead taking it from the marketbook once a race has completed. The file size seems good and I only got one missing file (17012017) but use with caution as betfair don't seem to care about the quality of this data.
Thanks Linus
How accurate do you guys think the data is?
The reason I ask is that there are items of data that appear wrong - like the horses with a PPWAP of 1, or the horses with one or two figure BSPs whose INPLAY MIN figure is 1000 (OK, it's possible that the horse went straight to 1000 as it refused to leave the stalls, but it's still odd).
Are these rare anomalies that can be safely disregarded when you're analysing many thousands of entries, or do they hint at a bigger problem?
However, the BSP figures seem to be accurate, judging by their efficiency
Jeff
How accurate do you guys think the data is?
The reason I ask is that there are items of data that appear wrong - like the horses with a PPWAP of 1, or the horses with one or two figure BSPs whose INPLAY MIN figure is 1000 (OK, it's possible that the horse went straight to 1000 as it refused to leave the stalls, but it's still odd).
Are these rare anomalies that can be safely disregarded when you're analysing many thousands of entries, or do they hint at a bigger problem?
However, the BSP figures seem to be accurate, judging by their efficiency
Jeff
-
- Posts: 12
- Joined: Sun Jun 02, 2019 9:22 pm
The link doesn't exist anymore, is it somewhere else or deleted?LinusP wrote: ↑Sat Dec 31, 2016 9:52 amA few have asked for this now:
Contains 2009-2016 in one csv.
https://www.dropbox.com/s/zp3a9r5vlg9eh ... v.zip?dl=0
Contains 2009-2016 in separate csv's.
https://www.dropbox.com/s/ls8d5sdih5hkx ... R.zip?dl=0
Both 46.2mb but excel will struggle with the top one.
COUNT(*): 962425
MAX(EVENT_DT): 2016-12-30 16:20:00.0
MIN(EVENT_DT): 2008-12-31 12:00:00.0
With a few if statements you can easily backtest, although I recommend chucking the data in a database, I plan on doing a short tutorial on how to do this soon.
More data can be found here, http://www.betangel.com/forum/viewtopic.php?f=5&t=11883