TPD Feedback and Issues
TPD-Zone wrote: ↑Fri Apr 19, 2024 11:40 amThank you! FYI, here's another one.
All sorted now - 16:25 Newbury.
@TPD-Zone
Hey TPD-Zone,
As you know, the actual race winner is often recorded as being behind the leader at the finish. Can you share any plans to enhance the accuracy of the data, particularly in terms of tracking the true position at the finish line? Really keen to know about any fixes or updates you guys are working on.
Hey TPD-Zone,
As you know, the actual race winner is often recorded as being behind the leader at the finish. Can you share any plans to enhance the accuracy of the data, particularly in terms of tracking the true position at the finish line? Really keen to know about any fixes or updates you guys are working on.
To provide a summary of our GPS tracking solution; the technology was designed, developed and continues to be administered by Gmax Technologies Ltd, who work with technology and engineers from the world famous Cambridge tech cluster.lotora wrote: ↑Fri Apr 26, 2024 2:32 pm@TPD-Zone
Hey TPD-Zone,
As you know, the actual race winner is often recorded as being behind the leader at the finish. Can you share any plans to enhance the accuracy of the data, particularly in terms of tracking the true position at the finish line? Really keen to know about any fixes or updates you guys are working on.
Gmax will always embrace improvements in the underlying technology and TPD themselves have taken a number of steps to improve the service - the introduction of two GPS transponders per horse, upgraded radio stations at a number of UK tracks etc.
Our data highlights that 85% of the 1.5 million horses tracked worldwide have been reported to an accuracy of within 1 meter and this statistic transposes to the UK.
It is worth bearing in mind that the GPS tracking is rarely going to be 100% accurate and we know that there is an average positional error of a head. However, two errors of that magnitude in opposite directions means an error of half a length. The error is random, each positional update being independent of the last.
Hope this provides some context.
To minimize the error , is there a specific set up that's favorable to certain races ?TPD-Zone wrote: ↑Mon Apr 29, 2024 10:29 amTo provide a summary of our GPS tracking solution; the technology was designed, developed and continues to be administered by Gmax Technologies Ltd, who work with technology and engineers from the world famous Cambridge tech cluster.lotora wrote: ↑Fri Apr 26, 2024 2:32 pm@TPD-Zone
Hey TPD-Zone,
As you know, the actual race winner is often recorded as being behind the leader at the finish. Can you share any plans to enhance the accuracy of the data, particularly in terms of tracking the true position at the finish line? Really keen to know about any fixes or updates you guys are working on.
Gmax will always embrace improvements in the underlying technology and TPD themselves have taken a number of steps to improve the service - the introduction of two GPS transponders per horse, upgraded radio stations at a number of UK tracks etc.
Our data highlights that 85% of the 1.5 million horses tracked worldwide have been reported to an accuracy of within 1 meter and this statistic transposes to the UK.
It is worth bearing in mind that the GPS tracking is rarely going to be 100% accurate and we know that there is an average positional error of a head. However, two errors of that magnitude in opposite directions means an error of half a length. The error is random, each positional update being independent of the last.
Hope this provides some context.
what do you recommend as an ideal usage ...
should we avoid picking a winner in a highly competitive race , especially in the final stages of a race but rather try and pick a lay .
or go for long distance hurdles .
much appriciated
decomez6 wrote: ↑Wed May 01, 2024 6:15 pmThe general advice is to avoid the business end of races, either using the positional metrics or the efficiency metrics. The TPD guys who are close to this have had success using a combination of the three efficiency metrics, Cadence Error, Velocity Error and Velocity Frequency - with greater weighting to each of the three metrics depending upon the race being on the flat or over obstacles. Our YouTube channel contains a few videos describing the rationale; https://www.youtube.com/@TPDZone/videosTPD-Zone wrote: ↑Mon Apr 29, 2024 10:29 amTo provide a summary of our GPS tracking solution; the technology was designed, developed and continues to be administered by Gmax Technologies Ltd, who work with technology and engineers from the world famous Cambridge tech cluster.lotora wrote: ↑Fri Apr 26, 2024 2:32 pm@TPD-Zone
Hey TPD-Zone,
As you know, the actual race winner is often recorded as being behind the leader at the finish. Can you share any plans to enhance the accuracy of the data, particularly in terms of tracking the true position at the finish line? Really keen to know about any fixes or updates you guys are working on.
Gmax will always embrace improvements in the underlying technology and TPD themselves have taken a number of steps to improve the service - the introduction of two GPS transponders per horse, upgraded radio stations at a number of UK tracks etc.
Our data highlights that 85% of the 1.5 million horses tracked worldwide have been reported to an accuracy of within 1 meter and this statistic transposes to the UK.
It is worth bearing in mind that the GPS tracking is rarely going to be 100% accurate and we know that there is an average positional error of a head. However, two errors of that magnitude in opposite directions means an error of half a length. The error is random, each positional update being independent of the last.
Hope this provides some context.
To minimize the error , is there a specific set up that's favorable to certain races ?
what do you recommend as an ideal usage ...
should we avoid picking a winner in a highly competitive race , especially in the final stages of a race but rather try and pick a lay .
or go for long distance hurdles .
much appriciated
Long distance hurdles seem to work well with Velocity Fluctuation graph providing a quick historical over view.TPD-Zone wrote: ↑Thu May 02, 2024 4:09 pm
The general advice is to avoid the business end of races, either using the positional metrics or the efficiency metrics. The TPD guys who are close to this have had success using a combination of the three efficiency metrics, Cadence Error, Velocity Error and Velocity Frequency - with greater weighting to each of the three metrics depending upon the race being on the flat or over obstacles. Our YouTube channel contains a few videos describing the rationale; https://www.youtube.com/@TPDZone/videos
The other three efficiency metrics including the splits could be made better when viewed in the graph format.
in the short distance races i tend to use the tool as a measurement of performance v/s the wisdom of the crowd.
and you can sometimes see races where the crowd is emotionally attached to selection and will not dump their positions irrespective of poor TPD performance.
Much appreciated
Thanks