|
JCapper Message Board JCapper 101
|
| Home |
Register
Log In |
| By | Data Window Track Weight Report - Q & A | |
| jeff 4/14/2014 11:08:58 AM | From the email inbox over the weekend... --quote: "Hey Jeff,--end quote My reply: No new programming changes are currently planned for the Data Window Track Weight Report. Just like you mentioned, the run style table displayed as part of the Data Window Track Weight Report is generic or broad category based. It isn't broken out in detail or comprehensively like the separate Run Style option on the Data Window factors drop down is for HDW subscribers. That said, below are some cut and paste's of Data Window output that (hopefully) can serve to illustrate the process by which I dig deeper into factors shaping the track weight report itself. One of the things I did very early on was create an inactive UDM (named _TrackDateExpression) that I use soley for the purpose of generating Track Weight Reports on the fly: SELECT * FROM STARTERHISTORY WHERE DATEDIFF('D',[DATE], NOW()) <= 15 AND TRACK='KEE' AND INTSURFACE <= 3 AND DIST < 1760 Note that the sql expression driving the UDM contains paramters to control the following: 1. The number of calendar days back in time from Now (in this case 15 days.) 2. Track Code (in this case 'KEE') 3. Intsurface (in this case <= 3 or everything except races run on a turf couse.) 4. Dist (in this case less than 1760 yards or races less than 1 mile in distance.) Note that the UDM itself is stored exactly as shown above. Note I can easily run the UDM for any and all combinations of track code, surface, dist, or time span simply by editing the stored parameters prior to hitting the Execute button. Let's look at some data... Here, I've selected the UDM from the drop down, selected TRACK WEIGHT REPORT from the factors drop down, and hit the Execute button:
Looking at the above Track Weight Report, I note that the algorithm analyzing the data (after comparing the ratios between the actual number of winners vs. the number of 'expected' winners for starters ranked 123 in CPace and 123 in CompoundLate respectively) came to the conclusion that the data sample was indicative of a speed tiring track - and "labeled" the data sample as 5.00 or "Speed Tiring." The only real conclusion I draw from that is that early speed horses have been getting the worst of it in sprint races at KEE while horses that don't expend a lot of energy trying to make the top in the early part of their races have been getting the best of it. More to come... -jp .
| |
| jeff 4/14/2014 11:22:33 AM | Here, I've executed the same UDM... but this time, I selected Running Style from the factors drop down before hitting the Execute button:
Keeping in mind that this is a small data sample... but it does appear that E and EP types have been getting the worst of it. (And that's really the only take-away I get after looking at this data sample.) Note that unlike the run style table on the track weight report, here, by selecting Running Style from the factors drop down prior to executing the UDM, the resulting data sample does contain a full breakout by run style. More to come... -jp .
| |
| jeff 4/14/2014 11:58:15 AM | Here, I've executed the same UDM... but this time, I selected Q Speed Points Number from the factors drop down before hitting the Execute button:
Once again keeping in mind that this is a small data sample... but it does appear that horses expending energy attempting to make the top in the early parts of their races in sprints at KEE have been taking the worst of it. And that's really the only take-away I would get after looking at this data sample in isolation. However, even though these are small data samples... This is the third data sample for the same races that seems to be saying the same thing. Each paints its own picture. My interpretation is that the pictures painted are building a case that the KEE surface over the past 15 calendar days has been shaping the outcome in sprint races in such a way that horses with early speed have had their chances of winning compromised. None of this guarantees the surface at KEE will behave that way for the remainder of the meet. Weather and track maintenance play a huge role in all of this. For example, the track super could easily decide to make the surface more speed friendly by rolling the surface to compress it. He could also accomplish the same effect by using shorter harrows in the "rakes" being dragged behind the tractors. Appraising track surface early in a race day is a skillset unto itself. In addition to looking at data samples from the recent past (like the ones posted above) I like to note things such as wind speed and direction today (sometimes a glimpse at the infield flag and/or the tails and manes of horses as they parade is a tell.) Other indicators that I find useful are how far the hooves of horses disappear into the surface as they parade in front of the grandstand along with the degree of banking built into the course itself - in the backstretch, turns, and the stretch itself. I've sometimes seen severe changes in banking implemented with no announcement in the middle of a meet - and seen it produce a 180 degree change in the way a course is playing as a result. (It pays to watch head on replays.) -jp .
| |
| NYMike 4/2/2015 12:37:41 PM | Jeff, The SQL expression that you used to target the last 15 days is interesting. Is there any way to create a "live" track weight report UDM that would fire say if the track weight changed over the last 15 days? Say I have done my research and found a track that typically is speed favoring at 6 furlongs, can I write a UDM to identify when this track moves to speed tiring at 6F over the last 15 day period? Thanks, Mike
| |
| jeff 4/2/2015 11:47:21 PM | Active UDMs aren't designed to do that. (Although having that ability with a single calc races mouse click would be a HUGE time saver.) That said, what you are describing is VERY doable as a situational query. There are lots of good examples in the Statistical programs thread: http://www.jcapper.com/messageboard/TopicReader.asp?topic=1799&forum=General Scroll down and look at the post I made about 5.5f dirt races at MNR. Hint: Timestamp is: 3/24/2015 2:36:33 PM -jp . ~Edited by: jeff on: 4/2/2015 at: 11:47:21 PM~
| |
| jeff 4/3/2015 12:17:30 AM | As mentioned in my posts from a year ago, maintaining a set of inactive track profile UDMs like the _TrackDateExpression UDM shown above can be a relatively painless way (vs. situational queries) to stay on top of things. Every few days, in the morning as part of your prep work prior to live play, bring up the Data Window, bring up your inactive track profile UDMs and execute them with the data broken out BY TRACK. Below is a cut and paste of Data Window results where I've edited my _TrackDateExpression UDM to return the most recent 15 calendar days of CPace ranks 1 and 2 in 6f dirt sprints:
A quick eyeball scan through the track codes will tell you that PEN-SUN-TAM have been speed favoring lately.
Of course, much of the time the results aren't going to be that clear cut. Before you can apply critical thinking to the results set you need a clear understanding of what the big picture looks like. To that end, the above sql expression can easily be modified to show you the big picture. Just change "DATEDIFF('D',[DATE],NOW()) <= 15" to "DATEDIFF('D',[DATE],NOW()) <= 180" Note: Making that edit will have your query showing you CPace rank 1 and 2 performance in 6f dirt sprints for the most recent 180 days - which should be enough data to give you an idea of what the big picture looks like.
By comparing individual track rows from the first data set to the big picture, the track codes with the greatest deviation (one way or the other) should stick out. For example: PEN-SUN-TAM are three track codes where the most recent 15 day performance for CPace ranks 1 and 2 sticks out vs. the big picture benchmark (above.) That said, no way to know ahead of time with 100% certainty whether or not the recent speed favoring trend at PEN-SUN-TAM continues - or whether these track codes revert back to the norm. My "lean" after looking at these results and doing a little critical thinking is that the flat bet roi (0.9902 across all tracks) for the 15 day sample is too good to be true. Even if the surfaces themselves continue to be speed favoring - in my opnion it's likely that other bettors are going to catch on - and as they do - the avg. win mutuel for early runners will drop accordingly. That said, no way to know ahead of time if my prediction (that some reverting of the norm is going to take place) is right or wrong. The best that queries like these can really do is make you aware of what the big picture looks like and when a smaller data set represents a departure from it. Note about the above queries: CPace is assigned to slot #F20 in my sql factor setup. Thus the reference to RANKF20 in the above sql expressions. -jp .
|
| Reply |
|
Copyright © 2018 JCapper Software back to the JCapper Message Board www.JCapper.com |