I understood they'd done away with the data request charges since moving to the new API so no need to throttle your requests.
The amount of data you can retrieve in one request is limited though. If you exceed their maximum weighting of 200 you'll get the TOO_MUCH_DATA error response. The weighting is explained here https://api.developer.betfair.com/servi ... est+Limits
Obviously the more data you request in terms of price depth etc per market the higher the weighting is.
API-NG Direct Access Pros/Cons
You sometimes get 'TOO_MANY_REQUESTS' if you send to many requests (max should be 5 a second)
https://api.developer.betfair.com/servi ... Exceptions
http://forum.bdp.betfair.com/showthread.php?t=2838
I used to get this error quite a lot but have moved to a 3 thread approach sending requests every 200ms on the second in the hope to sync with betfair's batch processing. No idea if it is beneficial but I have stopped getting these errors
https://api.developer.betfair.com/servi ... Exceptions
http://forum.bdp.betfair.com/showthread.php?t=2838
I used to get this error quite a lot but have moved to a 3 thread approach sending requests every 200ms on the second in the hope to sync with betfair's batch processing. No idea if it is beneficial but I have stopped getting these errors
I just thought Id post an update on this
Although I use Betangel each and every day and will continue to do so. I had a real desire to get under the bonnet and fine tune some of my BOT's to explore how I might capitalise on any tiny increases on speed.
I took a great c# course (on a websire called UDEMY) to enable me to make my own tweaks to existing bespoke BOT's I use, really enjoyed that course.
More recently I bought James Butlers book - Programming for Betfair. Initially I was wishing that it was centered around C# rather than Visual Basic, but having part worked my way through this, I can see that just understanding the concepts of how you interact with the API, means you could (if needed) easily port the application to another lanuage.
The book assumes little programming knowledge and takes you step by step through building your own application. Moreover, although I havent got there yet, it seems a small step to be able to extract data for offline analysis. All in all I would recommend this book and credit wher it is due to James.
You will encouter errors, ( i spent two days finding the last one!), but I learned how to debug and find them myself.
Finally I would like to publically thanks Xitan for his help in getting my SQL database up and running
All in all, a steep learning curve but looking forward to building my knowledge on this platform. Worth the time investment, if you have the inclination.
Regards
Peter
Although I use Betangel each and every day and will continue to do so. I had a real desire to get under the bonnet and fine tune some of my BOT's to explore how I might capitalise on any tiny increases on speed.
I took a great c# course (on a websire called UDEMY) to enable me to make my own tweaks to existing bespoke BOT's I use, really enjoyed that course.
More recently I bought James Butlers book - Programming for Betfair. Initially I was wishing that it was centered around C# rather than Visual Basic, but having part worked my way through this, I can see that just understanding the concepts of how you interact with the API, means you could (if needed) easily port the application to another lanuage.
The book assumes little programming knowledge and takes you step by step through building your own application. Moreover, although I havent got there yet, it seems a small step to be able to extract data for offline analysis. All in all I would recommend this book and credit wher it is due to James.
You will encouter errors, ( i spent two days finding the last one!), but I learned how to debug and find them myself.
Finally I would like to publically thanks Xitan for his help in getting my SQL database up and running
All in all, a steep learning curve but looking forward to building my knowledge on this platform. Worth the time investment, if you have the inclination.
Regards
Peter
No thanks needed, Peter. I'm sure if I didn't help you with that database many others on this forum would have done.
In the end, setting up a database is just the start! The hard bit is finding any sort of edge in the data, with which I wish you the very best of luck!
A very good next step, as you mention, is to start collecting timestamped data. Especially for in-play if that's what you're into. Then you can look into minuscule detail of what's going on to refine strategies.
As a few people have requested, a really cool feature for BetAngel would be to be able to load a historical race and backtest your automation rules offline. That's never likely to happen due to market data licensing rules, so will have to be left as an exercise for each individual.
In the end, I think you've got just the right attitude, which is to have fun learning and playing with it as you go along!
In the end, setting up a database is just the start! The hard bit is finding any sort of edge in the data, with which I wish you the very best of luck!
A very good next step, as you mention, is to start collecting timestamped data. Especially for in-play if that's what you're into. Then you can look into minuscule detail of what's going on to refine strategies.
As a few people have requested, a really cool feature for BetAngel would be to be able to load a historical race and backtest your automation rules offline. That's never likely to happen due to market data licensing rules, so will have to be left as an exercise for each individual.
In the end, I think you've got just the right attitude, which is to have fun learning and playing with it as you go along!
I really hope a route via a 3rd party can be found to allow the historical backtesting to occur. There is a competing product does exactly this, whereby a whole meeting can be purchased as a download for 50p!! Like yourself, I timestamp the in-play data using the api-ng and save it as a json file for later playback and examination. It's really interesting to see the extent of the width in back/lay when looking sequentially thro the collected data. So many ideas have been spawned by looking back at this data that I really strongly feel that finding a solution should be at the top of BA's agenda now, given that there are ways around this.xitian wrote: As a few people have requested, a really cool feature for BetAngel would be to be able to load a historical race and backtest your automation rules offline. That's never likely to happen due to market data licensing rules, so will have to be left as an exercise for each individual.
I made a little plea to this effect here: viewtopic.php?f=20&t=11983