How would you quantify being held up for clear running for inclusion in model?

[JerryBoyle]

Originally Posted by CBYRacer

I like this approach, Robert. Have you done this before? If so, any suggestions on where to find the person?

I wonder if this is something that could be made simple enough such that you could put it on Amazon Mechanical Turk or some other crowdsourcing platform. You'd have to keep it braindead simple, so I'm not sure it's possible. What is the simplest thing we could look for in a race replay that would have had an impact on the runner's time?

[CBYRacer]

Originally Posted by JerryBoyle

I wonder if this is something that could be made simple enough such that you could put it on Amazon Mechanical Turk or some other crowdsourcing platform. You'd have to keep it braindead simple, so I'm not sure it's possible. What is the simplest thing we could look for in a race replay that would have had an impact on the runner's time?

Yep. That would probably be an option, Jerry. You could provide a really simple checklist of things to look for and then maybe have the person spend a couple of hours initially watching examples that illustrate those things. You could then spot check a few of the person's initial notes, provide coaching, and then let them try a few more. I just wonder if there's going to be a positive ROI on hiring someone to do this. You'd probably have to pay them $10-$15 / hour (unless you did it overseas) and they could probably complete 12 races per hour (5 minutes per race) including notes? So essentially, it would be about $10 per card that you'd have to recoup. Does that seem worth it?

Obviously, it wouldn't be as effective as watching yourself, but it would at least give you some consistency. The person would simply write 0 or 1 if the trip attribute applied to each horse. They could also rate the overall trip (e.g., 'no trouble', 'perfect pocket trip', 'some trouble', 'terrible trip'). All of these things could be coded then.

[JerryBoyle]

Originally Posted by CBYRacer

Yep. That would probably be an option, Jerry. You could provide a really simple checklist of things to look for and then maybe have the person spend a couple of hours initially watching examples that illustrate those things. You could then spot check a few of the person's initial notes, provide coaching, and then let them try a few more. I just wonder if there's going to be a positive ROI on hiring someone to do this. You'd probably have to pay them $10-$15 / hour (unless you did it overseas) and they could probably complete 12 races per hour (5 minutes per race) including notes? So essentially, it would be about $10 per card that you'd have to recoup. Does that seem worth it?

Obviously, it wouldn't be as effective as watching yourself, but it would at least give you some consistency. The person would simply write 0 or 1 if the trip attribute applied to each horse. They could also rate the overall trip (e.g., 'no trouble', 'perfect pocket trip', 'some trouble', 'terrible trip'). All of these things could be coded then.

I'm not sure if it's worth it to be honest, but I've made a note to try it out. I think what I'll do is pick a particular season at any track where there are a decent number of races and a decent number of repeat runners. At the end of the day, the output of the replay will be used to adjust the time the runner ran on a particular race, so evaluating how "good" that adjustment is will be easier if the runners run on the same track next time.

From there, I'll do the first season on my own, trying to capture things that are very simple to encode in rules or steps and are very easy to spot quickly. I think I'll avoid overall analysis that requires domain specific knowledge like "bad trip". With that output, I can hopefully make some adjustments and test how significant those adjustments are. If they're significant, I'll post the job somewhere and see how the crowdsourced version compares.

My initial thought is to try and keep the "jobs" at the race level. That way I can post an arbitrary number of races and let the contractors have at them. At that point, I'm paying for race. I'll keep track of the time I take to do it when I go through the first season and pay accordingly, prorating some reasonable hourly wage by expected time per race.

[CBYRacer]

Originally Posted by JerryBoyle

I'm not sure if it's worth it to be honest, but I've made a note to try it out. I think what I'll do is pick a particular season at any track where there are a decent number of races and a decent number of repeat runners. At the end of the day, the output of the replay will be used to adjust the time the runner ran on a particular race, so evaluating how "good" that adjustment is will be easier if the runners run on the same track next time.

From there, I'll do the first season on my own, trying to capture things that are very simple to encode in rules or steps and are very easy to spot quickly. I think I'll avoid overall analysis that requires domain specific knowledge like "bad trip". With that output, I can hopefully make some adjustments and test how significant those adjustments are. If they're significant, I'll post the job somewhere and see how the crowdsourced version compares.

My initial thought is to try and keep the "jobs" at the race level. That way I can post an arbitrary number of races and let the contractors have at them. At that point, I'm paying for race. I'll keep track of the time I take to do it when I go through the first season and pay accordingly, prorating some reasonable hourly wage by expected time per race.

Cool. Sounds like a good plan to test. Good luck. Would love to hear how your experiment goes with this.

[Augenj]

Originally Posted by hpollock

How would you quantify or factor this into a model

I hesitated answering this because there is no simple answer and my answer would be long and boring. However, here's a shortened version of a complex subject.

First, I analyze 15 major tracks over a year using the THA system on a computer. As part of the race calculations, it uses two of eight base factors that could potentially be adjusted for trouble - Speed and Finish. In other words, bump Speed and Finish to some degree when a horse encounters serious enough trouble during a race.

Two issues. How much to bump and which trouble descriptions are worthy of bumping. I don't know how many chart callers there are (one at each track?) but they can't all be charting races exactly the same way. After a lot of trial and error, I found that the following descriptions were "slightly" worthy of bumping Speed and decreasing Finish (3rd to 2nd, etc) over 15 tracks. I looked for "key words" and not the whole description.
They are:
"blocked", "bolted", "boxed", "broke slowly", "clipped", "ducked", "dwelt", "forced", "impeded", "lost irons", "lost whip", "off slowly", "pinched back", "rank", "rough trip", "roughed", "shuffled", "shut off", "squeezed", "steadied", "taken up", "wide trip".

You can find all of the trouble descriptions here:
https://www.equibase.com/products/cc-comments.cfm

During test runs over 15 data bases, actually 16 due to GP being split into winter and spring/summer, I looked for bell curves in the results to determine which Speed and Finish parameters produced the most winners. Like I said earlier, the results were only slightly better but enough to keep in THA.

I may revisit this in the future, time permitting.

[headhawg]

Originally Posted by CBYRacer

...You'd probably have to pay them $10-$15 / hour (unless you did it overseas) and they could probably complete 12 races per hour (5 minutes per race) including notes? So essentially, it would be about $10 per card that you'd have to recoup...

I think that you are overestimating someone's willingness to take trip notes at what you are willing to pay. Why wouldn't someone just provide that service to everyone? Trip Note Pros charges $9 per track per day. OptixNotes is $150/month. I'm sure there are other services out there but you get the idea.

[CBYRacer]

Originally Posted by headhawg

I think that you are overestimating someone's willingness to take trip notes at what you are willing to pay. Why wouldn't someone just provide that service to everyone? Trip Note Pros charges $9 per track per day. OptixNotes is $150/month. I'm sure there are other services out there but you get the idea.

$9 per track per day...is that for the full racing card (8-10 races)? If so, that is almost exactly how much I said it would cost. I said $10 per card. What am I missing from your response?

[classhandicapper]

Originally Posted by headhawg

Trip Note Pros charges $9 per track per day. OptixNotes is $150/month. I'm sure there are other services out there but you get the idea.

This is all troubling for "trip" handicappers.

Then more people there are packaging quality trip information and selling it and the more people that are using computers to automate and test it, the less valuable it will all become, especially because it's typically people that bet larger amounts that buy these services or develop computer models.

[Jeff P]

Thus, the value in making your own observations for trip types that have statistical significance - and keeping quiet as a mouse about the exact nature of the trip types in your model.


-jp

.

[headhawg]

Originally Posted by CBYRacer

$9 per track per day...is that for the full racing card (8-10 races)? If so, that is almost exactly how much I said it would cost. I said $10 per card. What am I missing from your response?

Let's say I want to create trip notes. You offer to pay me $15/hour to create the notes. I work on three tracks at an hour each. I make $45. Why wouldn't I want to sell those notes to everyone instead of just giving them to you? @$10/per, I would only need to sell 5 tracks worth and already be ahead of your deal. My point was not how much it would cost for the completed notes but rather why would someone want to do it for your quoted $10-$15 per hour? It doesn't make economic sense unless part of the agreement is that the notes are not exclusive to you. That said, for marketing purposes I would want to do the notes my way and not yours. The short answer is my point was about trying to find someone to do it for you, not that the endeavor itself wasn't useful to some degree.

[CBYRacer]

Originally Posted by headhawg

Let's say I want to create trip notes. You offer to pay me $15/hour to create the notes. I work on three tracks at an hour each. I make $45. Why wouldn't I want to sell those notes to everyone instead of just giving them to you? @$10/per, I would only need to sell 5 tracks worth and already be ahead of your deal. My point was not how much it would cost for the completed notes but rather why would someone want to do it for your quoted $10-$15 per hour? It doesn't make economic sense unless part of the agreement is that the notes are not exclusive to you. That said, for marketing purposes I would want to do the notes my way and not yours. The short answer is my point was about trying to find someone to do it for you, not that the endeavor itself wasn't useful to some degree.

Got it. Thanks, headhawg, that makes sense. In having worked with some of these outsourcing services, a lot of the folks are not interested in fully commercializing their work or might know how to effectively do that (i.e., set up a website, handle customer service questions, etc.). They're just looking for the hourly rate. But your right, an enterprising person might ask the question that you're asking.

It seems like from your note that there are a few of these vendors already. Optixnotes looks good, but doesn't cover all tracks.

[CBYRacer]

Originally Posted by classhandicapper

This is all troubling for "trip" handicappers.

Then more people there are packaging quality trip information and selling it and the more people that are using computers to automate and test it, the less valuable it will all become, especially because it's typically people that bet larger amounts that buy these services or develop computer models.

Yep. Feels like this could be another big headwind for the average player.

[classhandicapper]

Originally Posted by Jeff P

Thus, the value in making your own observations for trip types that have statistical significance - and keeping quiet as a mouse about the exact nature of the trip types in your model.


-jp

.

I hear what you are saying perfectly well, but the more sharp people out there watching replays and selling their services, the more likely they are observing the same things as you and slowly whittling away at the value by selling that information to even more people.

I'd hate for it get like speed figures.

If you were making your own speed and pace figures in the 80s, it was easier to find a genuine overlay than it is now when everyone has high quality figures. Now I sometimes feel like I am searching for isolated cases where I disagree with the commercial figure makers.

[Tom]

Originally Posted by thaskalos

Some people post as if these are telegrams where they charge you by the letter.

(Stop)

[sjk]

I have about 15 years of the chart comments. There are around 2.1MM horses and around 1.4MM unique comments. I am sure much of the distinguishing elements are spaces and punctuation. I would find the task of building a classification and calculating the associated ROIs next time out pretty daunting and I am not hopeful enough about getting a good result to try it.