New GPS Charts

[cj]

There is some really good info in these, much better than the Trakus data.

Twitter

I also learned something new. It is possible to run less than the published distance of the race. Distances are measured three feet from the rail. That is why you will see some horses run less than the actual race distance overall if they stay on the rail on the turns, and it applies to individual segments as well.

[classhandicapper]

Originally Posted by cj

There is some really good info in these, much better than the Trakus data.

https://twitter.com/TimeformUSfigs/s...96115651481600

I also learned something new. It is possible to run less than the published distance of the race. Distances are measured three feet from the rail. That is why you will see some horses run less than the actual race distance overall if they stay on the rail on the turns, and it applies to individual segments as well.

Something like that may also explain why Trakus data is not consistent from track to track for horses running along the rail. They may not be adjusting "3 feet" consistently from track to track and distance to distance.

The one thing I don't like about the layout is that you have to scroll left/right to see the entire race (unless I'm missing something).

[CheckMark]

Talked about this in my thread I posted on Friday http://www.paceadvantage.com/forum/s...d.php?t=157326

Seems like a neat idea! Way better in-depth data than Trakus

[cj]

Originally Posted by classhandicapper

Something like that may also explain why Trakus data is not consistent from track to track for horses running along the rail. They may not be adjusting "3 feet" consistently from track to track and distance to distance.

The one thing I don't like about the layout is that you have to scroll left/right to see the entire race (unless I'm missing something).

I'm not a fan of the scrolling either, hopefully it will get better, still really new.

The Trakus thing is even more puzzling IMO. They are always over 3.5 to 4 feet per furlong, but that was assuming horses had to run the full distance of the segment or race. This means the error is even bigger than I had thought.

[maddog42]

This begs a few questions. While I believe that a few horses can run less than the posted Trakus data. How wide is a horse? 30 inches? Where on the saddle is the sensor located? How far from the rail does a horse usually run?
12 inches? 14? I would be very surprised if a horse can average running only
a foot from the rail.

[dlivery]

When the gate pops is this not considered a constant threat.
Or the Evil eye.

[ubercapper]

Originally Posted by maddog42

This begs a few questions. While I believe that a few horses can run less than the posted Trakus data. How wide is a horse? 30 inches? Where on the saddle is the sensor located? How far from the rail does a horse usually run?
12 inches? 14? I would be very surprised if a horse can average running only
a foot from the rail.

The sensor is typically located on the saddle pad in a pocket located just behind the saddle.

[MJC922]

Good stuff.

[cj]

I've been working on adding ground loss to the TimeformUS Speed Figures for a few weeks and it was just a coincidence these charts started appearing recently, but I'm finding them accurate and a real time saver. This is an example of how I'm looking to use it, still in the very early stages:

Twitter

Twitter

[classhandicapper]

Won't adding ground loss in make new figures with ground loss included and those without ground less comparable?

The new ground loss figures will all be bumped up by some amount unless the horse was on the rail the whole way. The old ones will all remain the same.

I have more faith in ground loss figures in theory than in practice because of biases and the way tracks play differently in general, but I think you are doing it correctly by looking at when and how the ground loss occurred.

I spend a ton of time on ground loss watching replays now, but if I was honest with myself I'd have to say it has been a horrible use of my time.

I'm really only interested in the extremes, but it takes lot of time to just get that.

I'm mostly interested in those perfect up the rail closer trips against a fast pace (negative), horses that were used hard early while wide, horses that were hung out wide just as the pace starting picking up late, and extreme ground loss from start to finish.

The rest of it feels more like another area of noise in the figures because the impacts are hard to understand. I see races all the time where horse A beat horse B, B was wider, they meet again, this time A was wider, but A beats him again as the favorite. It's like the ground loss was a total non event even though there was huge net change in roles.

[pandy]

Ground loss is tricky because of the good and bad rail tracks. On a closer's track the best place to be is outside so the inside horses are saving ground but seem to get more tired because the rail is dead.

[cj]

Originally Posted by classhandicapper

Won't adding ground loss in make new figures with ground loss included and those without ground less comparable?

The new ground loss figures will all be bumped up by some amount unless the horse was on the rail the whole way. The old ones will all remain the same.

I don't think they would be bumped up, because the variant is going to adjust to the new method. I'd still be using projections from the old scale. In the example I posted, maybe Tom's d'Etat is and Improbable were both projected to run 125 for example. Former calculations gave Tom a 140 raw figure and Improbable a 141. The variant would be about 15 or 16. Of course this is just using two horses, there would be other races and other horses from this race even.

Now, with ground loss, the raw figures changed to 142 and 148. The variant jumps to about 20.

Originally Posted by classhandicapper

I have more faith in ground loss figures in theory than in practice because of biases and the way tracks play differently in general, but I think you are doing it correctly by looking at when and how the ground loss occurred.

I spend a ton of time on ground loss watching replays now, but if I was honest with myself I'd have to say it has been a horrible use of my time.

I'm really only interested in the extremes, but it takes lot of time to just get that.

I'm mostly interested in those perfect up the rail closer trips against a fast pace (negative), horses that were used hard early while wide, horses that were hung out wide just as the pace starting picking up late, and extreme ground loss from start to finish.

The rest of it feels more like another area of noise in the figures because the impacts are hard to understand. I see races all the time where horse A beat horse B, B was wider, they meet again, this time A was wider, but A beats him again as the favorite. It's like the ground loss was a total non event even though there was huge net change in roles.

Like I said, still very early. Could very well be that we only use it in some situations or not at all with a bias. I really don't know. I'm already being very conservative. I'm not using near the amount others do, one path on the turn equal to one length. I started with half of that as the base and adjust up or down a little based on the pace during the segment being measured. The adjustment still brings it nowhere near the 1 to 1 others use.

[cj]

Originally Posted by pandy

Ground loss is tricky because of the good and bad rail tracks. On a closer's track the best place to be is outside so the inside horses are saving ground but seem to get more tired because the rail is dead.

I agree. But how often does that really happen? And when it does, we would always have the option to not use and ground loss adjustment.

[classhandicapper]

Originally Posted by cj

I don't think they would be bumped up, because the variant is going to adjust to the new method.

OK.

I get it.

[classhandicapper]

Originally Posted by cj

Like I said, still very early. Could very well be that we only use it in some situations or not at all with a bias. I really don't know. I'm already being very conservative. I'm not using near the amount others do, one path on the turn equal to one length. I started with half of that as the base and adjust up or down a little based on the pace during the segment being measured. The adjustment still brings it nowhere near the 1 to 1 others use.

It's interesting that you started with 1/2.

I never had the data to study it, but just from day to day experience it always seemed like 1 length per path tended to overrate the impact. But I still wanted something to get me in the ballpark in the cases where the horse lost a lot of ground. I usually use 1/2 because the math is easy and that's as good a guess as any.