TPD Feedback and Issues

[TPD-Zone]

Sam Houston race 3 -

• No data throughout the race for 'position', distance to leader' or 'race leader speed'.

• Data flow error triggered using "race information has NOT been updated from TPD in the last 0.8 seconds".

17/03/2022 21:03:46: Guardian has detected that the market is suspended
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: position = 0
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: speed = 10.63
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: stride = 2.55
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: distance to finish = 1701.7
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: distance to leader = 0
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for market: race leader speed = 0
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: lat = 29.93 = 30
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: long = -95.53 = -96
17/03/2022 21:03:46: [G_Auto 1] : Store Value (Shared) for Comedian: market_metres_to_finish = 1709.9

We can see the exact same thing in our logs - passed to the Developers for investigation. Many thanks for highlighting.

[jamesedwards]

Penn race 1 - something odd with the data for 'Yes Its Roo' which seemed to be several metres out throughout the race. eg with 100m to the finish TPD data showed he was last but he was actually in 3rd and several metres in front of 4th and 5th. As they crossed the line TPD had 'Yes Its Roo' beaten 10m into 5th but he actually finished 2nd.

[jamesedwards]

Penn race 2 - Something odd with the data for 'Pride In The Biz' and/or 'Simmard Shenanic' which seemed to be several metres out throughout the race. eg as they crossed the line TPD had 'Pride In The Biz' a metre ahead of 'Simmard Shenanic' but in actual fact he was beaten into 2nd by a good length.

[jamesedwards]

Penn race 3 - Something odd with the data for 'Colour Guard' which seemed to be several metres out throughout the race. eg as they crossed the line TPD had 'Colour Guard' in 2nd position but in actual fact he was well beaten into 4th.

Please can these issues at Penn be investigated asap. For now I'm switching off my automaton for this track.

[ShaunWhite]

James I was considering a TPD sub for my own stuff but can't find a Service Level Agreement on their site. I guess you saw one when you subscribed so I wondered what it said about data integrity, accuracy and reliability? I'm not going to sign up to anything unless I know it's not just an agreement for me to pay and no agreement about what I'm paying for.

[jamesedwards]

James I was considering a TPD sub for my own stuff but can't find a Service Level Agreement on their site. I guess you saw one when you subscribed so I wondered what it said about data integrity, accuracy and reliability? I'm not going to sign up to anything unless I know it's not just an agreement for me to pay and no agreement about what I'm paying for.

There is no service level. Their FAQs state the average error is 40ms (down from 70ms the previous year) so about a head in either direction. In my experience so far this seems about right. There are other technical issues especially with US races but they are slowly being fixed and overall stability is improving. It's a product still in it's infancy so I'm willing to cut them some slack for now.

When combined with Bet Angel it potentially becomes really powerful and with so many potential threats to my future trading income I'm not letting a technological advancement be the one that brings me down!

[jamesedwards]

Laurel race 8 - no data throughout.

[ShaunWhite]

It's a product still in it's infancy so I'm willing to cut them some slack for now.

It's a long infancy, you could get TPD data via Betfair back in 2018 and they say themselves....
"Our experience gained from tracking more than 16,000 races over 3+years ....blah blah blah"
...so it's not exactly teething trouble.

I dunno James, I'm just a bit bored with the way things are these days. You shell out good money to basically be a beta tester and you get zero back with regard to any sort of service. SLAs used to be an absolute given years ago until firms realised they could get customers without them and avoid any sort of costly responsibility. You're not my only source of opinion btw, a number of people elsewhere are pretty disillusioned and I'm actuially fishing for a positive or two believe it or not.

But there's only one thing that matters....are you managing to make decent money with just BA & TPD yet? I'm not seeing many success stories being posted.

[jamesedwards]

It's a long infancy, you could get TPD data via Betfair back in 2018 and they say themselves....
"Our experience gained from tracking more than 16,000 races over 3+years ....blah blah blah"
...so it's not exactly teething trouble.

Infancy in terms of integration and data flow through Bet Angel.

But there's only one thing that matters....are you managing to make decent money with just BA & TPD yet? I'm not seeing many success stories being posted.

Yes. I have been a full-time in-play trader for several years and it's absolutely complementing my profitability.

[ShaunWhite]

Yes. I have been a full-time in-play trader for several years and it's absolutely complementing my profitability.

Thx. That's good to hear.

I'll have to see if they do any historical data, if not then I guess my only option is to sign up, collect the tpd api for a couple of months, sync it with the price data api, and get on with some strategy testing. The todo list for this year gets bigger and bigger, this is a hard way to earn an easy living.

thx again.

[jamesedwards]

Sam Houston race 8 - multiple data errors throughout

[PeterLe]

I'll have to see if they do any historical data, if not then I guess my only option is to sign up, collect the tpd api for a couple of months, sync it with the price data api, and get on with some strategy testing. The todo list for this year gets bigger and bigger, this is a hard way to earn an easy living.

thx again.

Shaun Well worth giving it a go at £60 (say for 450 races). The opportunity cost could be huge. There are many who cant make this work, but a few say they are doing very well.
I originally got this this via the Betfair API around Aug 2020 and I incorporated it into my own program.I found it really hard to create a system that relies on this data alone to create a successful strategy. however, I found it useful to supplement an existing strategy.
More recently Ive been testing via Betangel on a different account as I find this much easier than doing so in my own stuff (programs). You can implement an idea in minutes on Betangel and get real feedback. Ive always been of the opinion that a well though out strategy wont lose much in testing anyway.
Before I got it via Betfair, the cost to get it direct rom TPD was around £35+VAT a meeting! so a bargain now (but only if it pays for itself of course)
Id say give it a go!
Regards
Peter

[The Silk Run]

I had a very similar experience to PeterLe reply. We found the TPD product worked well in support of other stratergies, but not stand alone. We used it as a hedging tool.

[ShaunWhite]

Thanks PeterLe. The thing that put me off was the dev effort because even though Guardian is handy for knocking something up quickly we all know its weaknesses. But if you think there's an advantage to be had despite Guardian, then that's an option I'm duty bound to take notice of. Like you say adding it to an existing strategy is a good start, the only issue is that I'm not sure if I can get Guardian to do what i've been doing.

I think we're in a similar position wrt tech, we can do it but rather we didn't have to, so I'm always interested in workarounds or a pick-n-mix approach. If only I had Liam's youth and brains, these obstacles are a 20min job for him rather than the 20 weeks I'd need to schedule

[ShaunWhite]

I had a very similar experience to PeterLe reply. We found the TPD product worked well in support of other stratergies, but not stand alone. We used it as a hedging tool.

Thanks, I'm surprised though about using it for hedging as hedging is something I wouldn't usually do deliberately. Imo bets are either value or not and I don't use my current position in my decision to have subsequent bets (other than it being a factor in stake/total exposure of course).