Trakus at Aqueduct

[Tom]

Trakus Tnet has Aqueduct data for today posted.
Actually, from the 28th. Charts only at this point.

[cj]

Cool, thanks for the heads up.

[FantasticDan]

12/11 the target date for Trakus @AQU.

[cj]

12/11 the target date for Trakus @AQU.

Did they give up on the main track, now trying for the inner?

[Jasonm921]

Can they turn on the switch for the speakers as well?

[Tom]

Must be some snag - only three days left on the website, 6,7, and 8.
The rest are gone.

[cj]

Must be some snag - only three days left on the website, 6,7, and 8.
The rest are gone.

I don't think it is "officially" in use yet. Pretty sure that is set for next week.

[Tom]

They had posted Belmont data for about two weeks prior to officially starting.
When they took down days of Aqu, I assumed they found out it was not accurate? To be safe, I scrapped what I had saved from those days.

[cj]

Just checked the page, back to a mid-November launch date, which is odd to see in mid-December.

http://www.nyra.com/aqueduct/handicapping/trakus/

[Tom]

I hadn't looked at that page - I was using the T-Net site.
Thanks for the link!

[cj]

I hadn't looked at that page - I was using the T-Net site.
Thanks for the link!

Pretty sure that page has had data a few times too, though I don't think it was correct.

[FantasticDan]

12/11 the target date for Trakus @AQU.12/18 now the target date.

[Tom]

I was looking for the chicklets yesterday.

[FantasticDan]

12/18 now the target date.Didn't happen, but tomorrow is supposedly the day..

[Tom]

Full menu up at T-Net for today.
Love the overhead view.

[EMD4ME]

https://www.nyra.com/aqueduct/videos/race-replay/AQD/2015/20151106/8/pan/

How does this happen? Watch the stretch drive. The 1 clearly is holding off the 4 and Trakus has the 4 getting by the 1.

I see many errors (chicklets leaving the screen and coming back) with Trakus bot none like this before.

[Steve 'StatMan']

Odd seeing that, but glad the official results & payoffs are correct.

[cbp]

There are two types of TRAKUS participating track sites

1) The data appears in the page source and you can easily scrape the site. Gulfstream is an example.

2) The data is hidden and does not appear in the page source. Much harder to scrape. NYRA is an example.

This would appear to be controlled by the individual track/jurisdiction. I then wonder why a track would make it HARDER for bettors to make use of the data.

Moreover, the data on TRAKUS' site is much more comprehensive. Why the condensed version on track sites?

All this data and so many obstacles to using it.

[woodbinepmi]

cdp, did you get in touch with them at Trakus for an account?

[cbp]

Tried calling a few times but never get anyone live. Haven't left a message.

Ironically, I had an account ages ago and forget the password.

[woodbinepmi]

That happened to me also, sent them an email and they sent the password back to me.

[cbp]

I'll try that, thanks

[Tom]

All this data and so many obstacles to using it.

Amen.
You can't fix stupid.
Racing is run by greedy morons.

[EMD4ME]

Thanks for the effort everyone. Was just curious.

[cbp]

Here's something interesting:

WO R1 10/24/15

The :6:, High End Gal, NEVER appears as one of the chicklets. At any point in the race.

And her running lines, except for the finish, final time and beaten lengths, are empty.

I suppose her device was not working.

[EMD4ME]

Sat March 26th,2016. Race 4. I need your opinion.

Watch the 5 and 6. Trakus has the 6 travelling 28 more feet than the 5.

Am I crazy or does it sound like they switched the saddle cloths in their assessment?

[cj]

Sat March 26th,2016. Race 4. I need your opinion.

Watch the 5 and 6. Trakus has the 6 travelling 28 more feet than the 5.

Am I crazy or does it sound like they switched the saddle cloths in their assessment?

Shouldn't be too hard to see if that happened by checking BLs and positions against the Equibase chart.

[EMD4ME]

Shouldn't be too hard to see if that happened by checking BLs and positions against the Equibase chart.

I checked the replay for 5th time, realized their chicklet for the 6th was OFF. They had the 6 4 wide the whole way, horse was near the rail the entire trip.

Wasn't even like the 6 was switched with someone, just the usual, total incompetency.

That's why visually checking everything is so important.

[Cratos]

I checked the replay for 5th time, realized their chicklet for the 6th was OFF. They had the 6 4 wide the whole way, horse was near the rail the entire trip.

Wasn't even like the 6 was switched with someone, just the usual, total incompetency.

That's why visually checking everything is so important.
Trakus "chicklets" (sensors) are not turned off/on during a race.

However they can and do malfunction due to susceptibility which is a failure mode with any electronic monitoring.

What would be good would be for you to contact Trakus directly and explain your concern; they are admendable to comments about their product.

There is no way to correlate Trakus measurements to Equibase's metrics because they time the race entirely different. Trakus time each horse in the race with respect to distance travelled. Equibase time the race at a fixed distance with specific POCs.

[EMD4ME]

Trakus "chicklets" (sensors) are not turned off/on during a race.

However they can and do malfunction due to susceptibility which is a failure mode with any electronic monitoring.

What would be good would be for you to contact Trakus directly and explain your concern; they are admendable to comments about their product.

There is no way to correlate Trakus measurements to Equibase's metrics because they time the race entirely different. Trakus time each horse in the race with respect to distance travelled. Equibase time the race at a fixed distance with specific POCs.

That's good to hear. I was so annoyed that after 15 minutes, decided to post it on here to see if I'm going crazy.

I'll let them know as I hate wasting time triple checking things.

Thank you

[cj]

Trakus "chicklets" (sensors) are not turned off/on during a race.

However they can and do malfunction due to susceptibility which is a failure mode with any electronic monitoring.

What would be good would be for you to contact Trakus directly and explain your concern; they are admendable to comments about their product.

There is no way to correlate Trakus measurements to Equibase's metrics because they time the race entirely different. Trakus time each horse in the race with respect to distance travelled. Equibase time the race at a fixed distance with specific POCs.

So customers are QA for them? Lol, ok. They screw up pretty regularly and nobody is really checking. I was doing some work on Meydan today and found a 1600 meter race that has comical fractions and final time. The first 200 meters is posted as sub 21 with no run up. How can anyone take them seriously if they don't catch that? The race was months ago. There data is littered with errors.

It would be easy to check EMd's concern with the Eqibase chart or a replay. What you say in the last paragraph isn't really relevant.

[cj]

If they aren't turned on/off during a race, why do riderless horses disappear from the screen?

[Cratos]

So customers are QA for them? Lol, ok. They screw up pretty regularly and nobody is really checking. I was doing some work on Meydan today and found a 1600 meter race that has comical fractions and final time. The first 200 meters is posted as sub 21 with no run up. How can anyone take them seriously if they don't catch that? The race was months ago. There data is littered with errors.

It would be easy to check EMd's concern with the Eqibase chart or a replay. What you say in the last paragraph isn't really relevant.
It is very relevant; you should know the difference if you want to use either of the providers.

I come from a damn good engineering background and Quality Assurance (Q A) was never done in the manner you suggested, but I will allow you to be wrong and cynical as usual.

[Cratos]

If they aren't turned on/off during a race, why do riderless horses disappear from the screen?
Do you understand the I//0 process of electronic signaling?

[Tom]

Let me guess, it is flawed and no one is QC on it?

Errors like the one CJ reported are inexcusable.
I say that from a 45 year career in Quality Assurance, where 0 PPM is the expectation.

Trakus is incompetent.

[cj]

It is very relevant; you should know the difference if you want to use either of the providers.

I come from a damn good engineering background and Quality Assurance (Q A) was never done in the manner you suggested, but I will allow you to be wrong and cynical as usual.


All I'm saying is that I can look at both charts and see if chips were placed in the wrong horse. It is simple to do. That is the part that wasn't relevant. And it has happened before, certainly wouldn't be the first time and won't be the last.

I didn't suggest how QA is to be done. I said it ISN'T being done, not very well anyway. I was telling EMD4ME how he could do it since the people running the store aren't.

So no explanation how a blatantly wrong timing error could be left in the charts for months? I'm cynical for a reason, but I'm not wrong.

[cj]

Do you understand the I//0 process of electronic signaling?

Yes.

[classhandicapper]

I use Trakus extensively to help with my trip notes and bias determination. Having the data in front of me helps me find and follow horses, especially in some of the bigger fields.

That said, I've found a couple of distances to be way off in terms of ground loss at certain tracks. I also find that being approximately 1 path wide (the way I personally view it) seems to represent a slightly different number of feet depending on the track.

I still use paths in my notes (ex. 2wt1; 3wt2 out 4wlt2). 2wt1 (2 wide turn 1) seems to represent approximately about 12 feet at some tracks the way I measure paths, but more than that elsewhere.

[Cratos]

I use Trakus extensively to help with my trip notes and bias determination. Having the data in front of me helps me find and follow horses, especially in some of the bigger fields.

That said, I've found a couple of distances to be way off in terms of ground loss at certain tracks. I also find that being approximately 1 path wide (the way I personally view it) seems to represent a slightly different number of feet depending on the track.

I still use paths in my notes (ex. 2wt1; 3wt2 out 4wlt2). 2wt1 (2 wide turn 1) seems to represent approximately about 12 feet at some tracks the way I measure paths, but more than that elsewhere.
Let me understand this, Trakus uses a XY coordinate system to define the points of travel of the horse during the race and by definition a line is connecting the points.

Therefore I am curious to understand how you can correct or invalidate any Trakus race distance value without having a comparable measuring system.

Am I saying that all of Trakus data is correct? No I am not, but what I am saying is that verification and validation depends on fact and not anecdotal rhetoric or supposition.

[classhandicapper]

Let me understand this, Trakus uses a XY coordinate system to define the points of travel of the horse during the race and by definition a line is connecting the points.

Therefore I am curious to understand how you can correct or invalidate any Trakus race distance value without having a comparable measuring system.

Am I saying that all of Trakus data is correct? No I am not, but what I am saying is that verification and validation depends on fact and not anecdotal rhetoric or supposition.

Take a look at all the 1 Mile races at Aqueduct on the Inner Track over time. Even though I don't have a technical way of measuring it, I've seen plenty of examples from watching the races that the ground covered on some of those races has been too high relative to the distance.

The rest of the cases I am talking about are more anecdotal. I watch a race visually and mark the path each horse took. Then I compare my notes to Trakus. It just seems like certain tracks may be a 1-2 feet different per path around a turn than others. There's no way I can prove that though.

[Cratos]

I am not going to try and convince you that what you are referring too is subjectivity.

Using Trakus data the minimum distance or the most efficient distance that a horse would travel around Aqueduct Inner Track at the mile distance is 5,302.4 feet.

However this is very difficult if not impossible for a horse to do.

While I agree that the RFID chips in the horses’ saddle cloth used by Trakus for monitoring distance and speed of each horse in the race can be subjected to conducted EMI; I am quite sure it doesn’t happen.

Again, without proof or some rational theoretical calculation to support your assumption I find it difficult to believe or understand.

[classhandicapper]

Using Trakus data the minimum distance or the most efficient distance that a horse would travel around Aqueduct Inner Track at the mile distance is 5,302.4 feet.



All I can do is tell you is what my experience was.

I used to enter the Trakus ground covered for the top 3 finishers for every race in NY into a spreadsheet to see how much further than the official distance each horse traveled. At a certain point it was clear that Trakus had some overstating issue even on the straights. You, I, and CJ discussed it back then. I built an adjustment into my spreadsheets to account for that (much as you are saying 1M is 5302.4 vs. an actual 5280). Even after that adjustment, some of the Mile races were overstating the actual ground run. If the typical horse that ran on the rail all the way (based on my observation) was running 5-10 feet more than my adjusted value, the same trip might be 30-35 feet more for the mile races.

I just looked at few races from this year and they look much better. So maybe it was corrected at some point.

[whodoyoulike]

Let me guess, it is flawed and no one is QC on it?

Errors like the one CJ reported are inexcusable.
I say that from a 45 year career in Quality Assurance, where 0 PPM is the expectation.

Trakus is incompetent.

Now some of your comments are making sense to me.

[Cratos]

Now some of your comments are making sense to me.
And I say that with 30+ years of Consulting Design experience (Mechanical Engineering & Telecom) and Certified BB 6 Sigma Trakus is not incompetent.

What the horseplayers have with Trakus data is the vector info long needed, but the horseplayer has been indoctrinated with the Equibase/DRF data for such a long time that it is difficult to change.

For example, a race run at Santa Anita on the main track in 1:09 for 6F as published by Equibase/DRF will always be a time figure because by definition speed is the ratio of distance/time and Equibase/DRF only gives time with respect to a specified race distance.

Yes, an estimated speed can be derived from the Equibase/DRF data, but it would be inconsequential.

Do a vector analysis and calculate the horse's displacement; it cannot be done from the Equibase/DRF data.

[cj]

And I say that with 30+ years of Consulting Design experience (Mechanical Engineering & Telecom) and Certified BB 6 Sigma Trakus is not incompetent.

What the horseplayers have with Trakus data is the vector info long needed, but the horseplayer has been indoctrinated with the Equibase/DRF data for such a long time that it is difficult to change.

For example, a race run at Santa Anita on the main track in 1:09 for 6F as published by Equibase/DRF will always be a time figure because by definition speed is the ratio of distance/time and Equibase/DRF only gives time with respect to a specified race distance.

Yes, an estimated speed can be derived from the Equibase/DRF data, but it would be inconsequential.

Do a vector analysis and calculate the horse's displacement; it cannot be done from the Equibase/DRF data.

Santa Anita data in Equibase charts is from Trakus. Trakus has a ton of potential. They just don't have much QA, particularly since Pat Cummings left.

[Tom]

Do a vector analysis and calculate the horse's displacement; it cannot be done from the Equibase/DRF data.

Oh my, I will have to find a way to live without it then.:rolleyes:
Competence is the quality of the product to put out.
When your product is incorrect or missing, I call that incompetence.

Trakus has a lot of potential, but potential ain't real life.

[Cratos]

Santa Anita data in Equibase charts is from Trakus. Trakus has a ton of potential. They just don't have much QA, particularly since Pat Cummings left.
What is strange is neither you or Tom identified a failure mode due to the Trakus process.

Also in the product development process Quality is designed in and not added on ; the V&V performed by QA/QC is against a spec from engineering that is typically customer driven.


I am quite sure Trakus doesn't make their physical chips and their software is implemented into them.

Given that Trakus integrated its technology into an existing physical structure their EMC by any objective measure is damn good.

[cj]

What is strange is neither you or Tom identified a failure mode due to the Trakus process.

Also in the product development process Quality is designed in and not added on ; the V&V performed by QA/QC is against a spec from engineering that is typically customer driven.


I am quite sure Trakus doesn't make their physical chips and their software is implemented into them.

Given that Trakus integrated its technology into an existing physical structure their EMC by any objective measure is damn good.

None of which explains why they miss so many errors. The sub 21 quarter at Meydan in a 1600 meter race with no run up was one I mentioned. How is that not caught by even the most rudimentary quality control?

If you publish the data, you should take steps to ensure accuracy. Period.

[Cratos]

None of which explains why they miss so many errors. The sub 21 quarter at Meydan in a 1600 meter race with no run up was one I mentioned. How is that not caught by even the most rudimentary quality control?

If you publish the data, you should take steps to ensure accuracy. Period.
It is apparent from reading your posts that analytics is not your strong suit.

A timing metric wouldn't be the QA/QC function; what should be investigated is "cause" and not "effect." If the timing was wrong at Meydan then it probably was a system miscue, but if it was within in the system's acceptable accuracy metric the system is qualitative.

Having worked in two of the most quality driven regulated environments in the world in aerospace and medical devices I don't become alarm by ridiculous claims like yours.

Trakus provide data for 17 racetracks worldwide and to the best of my knowledge I haven't read or heard of a ground-swell complaint about their data.

Therefore if you have a bona-fide issue with the quality of their data, I would suggest that you contact either Jim Conant or Rob Griffin; both are VPs with responsibility for Engineering and Product Development at Trakus.

If they don't respond then go straight to Bob McCarthy the CEO who have an EE at the Masters level and attended the Sloan School from my Alma mater, MIT.

[Tom]

A timing metric wouldn't be the QA/QC function; what should be investigated is "cause" and not "effect." If the timing was wrong at Meydan then it probably was a system miscue, but if it was within in the system's acceptable accuracy metric the system is qualitative.

Apparently, the real world is not your strong suit.
It is the PRODUCT, not the PROCESS that matters.
What you do makes zero difference if your product is flawed. Trakus's product if frequently flawed, ie, dead wrong.

Would you defend the process of food processing if you ate a bad steak and got Mad Cow?

[cj]

It is apparent from reading your posts that analytics is not your strong suit.

A timing metric wouldn't be the QA/QC function; what should be investigated is "cause" and not "effect." If the timing was wrong at Meydan then it probably was a system miscue, but if it was within in the system's acceptable accuracy metric the system is qualitative.

Having worked in two of the most quality driven regulated environments in the world in aerospace and medical devices I don't become alarm by ridiculous claims like yours.

Trakus provide data for 17 racetracks worldwide and to the best of my knowledge I haven't read or heard of a ground-swell complaint about their data.

Therefore if you have a bona-fide issue with the quality of their data, I would suggest that you contact either Jim Conant or Rob Griffin; both are VPs with responsibility for Engineering and Product Development at Trakus.

If they don't respond then go straight to Bob McCarthy the CEO who have an EE at the Masters level and attended the Sloan School from my Alma mater, MIT.

Call it what you like, those bad times should be caught every time. If those times were within the accuracy metric then whoever wrote the metric doesn't know horse racing.

[classhandicapper]

Despite the things that may still need some work, I'm always disappointed when I'm doing my trip and bias notes and it's not for a Trakus track.

[cj]

Despite the things that may still need some work, I'm always disappointed when I'm doing my trip and bias notes and it's not for a Trakus track.

I like it too despite the issues at times. It is a good tool.

[steveb]

It is apparent from reading your posts that analytics is not your strong suit.

A timing metric wouldn't be the QA/QC function; what should be investigated is "cause" and not "effect." If the timing was wrong at Meydan then it probably was a system miscue, but if it was within in the system's acceptable accuracy metric the system is qualitative.

Having worked in two of the most quality driven regulated environments in the world in aerospace and medical devices I don't become alarm by ridiculous claims like yours.

Trakus provide data for 17 racetracks worldwide and to the best of my knowledge I haven't read or heard of a ground-swell complaint about their data.

Therefore if you have a bona-fide issue with the quality of their data, I would suggest that you contact either Jim Conant or Rob Griffin; both are VPs with responsibility for Engineering and Product Development at Trakus.

If they don't respond then go straight to Bob McCarthy the CEO who have an EE at the Masters level and attended the Sloan School from my Alma mater, MIT.

most of the above is double dutch to me, but would you care to comment on the below?

i was really keen to use trakus for singapore racing for a group i was once helping, and was in the midst of doing some research, when for some reason i went back to trakus site to check something.
what i was not ready for was what i actually saw.....
this was NOT an isolated example i later discovered.
the headings are mine, the data is all trakus.
one set of numbers was grabbed several weeks before the other set.
how could one possibly trust the info in light of that??
it is old, and it is possible they have lifted their game since, but when i saw that i immediately stopped what i was doing.

I decided to remove the attachment, as I am not sure I am allowed to post the Trakus data i have for this particular race.
if you want a look PM me.
I also decided to go to Trakus and again check this particular race.
Amazingly, it is different again.
That is 3 versions of data for the same race!

[Tom]

That is 3 versions of data for the same race!

Oh, I see.
The pick three? :lol:

[Cratos]

most of the above is double dutch to me, but would you care to comment on the below?

i was really keen to use trakus for singapore racing for a group i was once helping, and was in the midst of doing some research, when for some reason i went back to trakus site to check something.
what i was not ready for was what i actually saw.....
this was NOT an isolated example i later discovered.
the headings are mine, the data is all trakus.
one set of numbers was grabbed several weeks before the other set.
how could one possibly trust the info in light of that??
it is old, and it is possible they have lifted their game since, but when i saw that i immediately stopped what i was doing.

I decided to remove the attachment, as I am not sure I am allowed to post the Trakus data i have for this particular race.
if you want a look PM me.
I also decided to go to Trakus and again check this particular race.
Amazingly, it is different again.
That is 3 versions of data for the same race!
Steve, I will send you a PM because i am curious to see the Trakus problem you experienced.

[Cratos]

I think this thread is about horseracing and not about the QA/QC methodology.

However I do want say that an inferior process will yield an inferior PRODUCT.

Yes, it is the product that irritates the customer, but it is the "bad" process which produces the bad product and should be corrected; and that is the real world.

Therefore if Trakus "PRODUCT" is bad (and I don't believe that it Is) then their "PROCESS" should be corrected which I don't believe needs any more correcting other than ongoing maintenance.

[steveb]

Steve, I will send you a PM because i am curious to see the Trakus problem you experienced.


i guess you decided that you did not want to know?
if you are a user and a staunch defender of trakus, then it is logical to assume you have historical data on your computers in some format.
why don't you check some of the older, and compare with what it says now?

as i am writing this reply, i am rereading what you said here.
i did not experience any problems as such, the problems are all theirs.

any other interested parties are welcome to this file too.
it just shows the differences, between 2 different instances, of retrieved trakus data.
every field has differences, sometimes big ones.
an example of each 100 metre section time for one horse

100-8.68-8.11
200-5.82-5.85
300-5.97-5.94
400-6.06-6.06
500-6.06-6.06
600-6.17-6.17
700-6.32-6.29
800-6.17-6.20
900-6.00-6.06
1000-5.97-6.14
1100-6.00-6.17
1200-6.08-6.26
1300-6.08-6.08
1400-6.08-6.08
1500-6.23-6.23
1600-6.55-6.55


that is for the winner and its sections(both of them!) add to something a lot different than the overall winning time which was 100.11

how can one defend that?

[Tom]

how can one defend that?

Wait for it............ :lol: :lol:

[Cratos]

i guess you decided that you did not want to know?
if you are a user and a staunch defender of trakus, then it is logical to assume you have historical data on your computers in some format.
why don't you check some of the older, and compare with what it says now?

as i am writing this reply, i am rereading what you said here.
i did not experience any problems as such, the problems are all theirs.

any other interested parties are welcome to this file too.
it just shows the differences, between 2 different instances, of retrieved trakus data.
every field has differences, sometimes big ones.
an example of each 100 metre section time for one horse

100-8.68-8.11
200-5.82-5.85
300-5.97-5.94
400-6.06-6.06
500-6.06-6.06
600-6.17-6.17
700-6.32-6.29
800-6.17-6.20
900-6.00-6.06
1000-5.97-6.14
1100-6.00-6.17
1200-6.08-6.26
1300-6.08-6.08
1400-6.08-6.08
1500-6.23-6.23
1600-6.55-6.55


that is for the winner and its sections(both of them!) add to something a lot different than the overall winning time which was 100.11

how can one defend that?
Steve,
You are incorrect; I didn't decide not to send you a PM.

I travel a lot and I access this forum via my cellphone, but for whatever reason I cannot open my PMs on my phone.

I should be home tomorrow and I will try from my computer.