a. Create a Prob Expression designed to score RIDER performance for the current race type. Such a Prob Expression might look something very much like this:

SELECT TOP 100 * FROM STARTERHISTORY
       WHERE RIDER = {RIDER} 
       AND TRACK = {TRACK} 
       AND DIST = {DIST_TYPE} 
       AND INTSURFACE = {SURF_TYPE} 
       AND [DATE] < {RACEDATE} 
       ORDER BY [DATE] DESC

b. Add the Prob Expression to my sql factor setup -- and assign it one of the valid Behaviors for Prob Expressions (18-41.)

c. Build some databases from Scratch -- first folder in Mode 3 -- Next few folders in Mode 4 -- and going forward from there build the current folder in Mode 5 as new race card files are added.

a. Run a query in the Data Window with the data broken out BY RIDER.

--Link to a text file that contains a cut and paste from my Data Window a few minutes ago after doing just that where the specific set of circumstances I decided to look at was performance by rider on the turf at Saratoga from opening day 07-21-2017 through yesterday 08-05-2017:
http://www.jcapper.com/messageboard/reports/Impact_Study_Top3Riders_SAR-Turf.txt

--Note that the rider names you use for a sql udm need to be exactly as they appear in the StarterHistory table (or exactly as they appear on the HTML Report.)

--Note that a good way to get yourself a list of valid rider names is to export them to .csv file using the JCX File Exports Module. And from there double click the resulting .csv file to open it in Excel where you can see the rider names.

--Also note that I hand keyed the rider names "ORTIZ JOSE L-LEZCANO JOSE-ORTIZ JR IRAD-CASTELLANO JAVIER" (without the quotes) just below the top n riders before saving the text file.

b. Create a UDM using the names of the riders identified as Top n based on my observations in the Data Window.

--Link to a text file that contains a cut and paste from the Data Window that can be used to create a sql udm designed to flag the top n riders on the turf so far this meet at Saratoga:
http://www.jcapper.com/messageboard/reports/UDMText_SAR_Top3Turf.txt

• You should always check the Auto Format Off box in the SQL Expression Tool of the UDM Wizard whenever you create a name based sql udm. Checking the box prevents the interface from breaking the UDM text out to the next line in the event one of the names contains the characters "AND" (without the quotes.) Example: Anderson would be one such name that would cause this to happen if you fail to check the box.

• Note that there are actually four rider names in the udm not three. That's because I used number of wins on the turf to identify who the top riders are -- and two riders were tied with seven wins.

• Also note that the top n riders (or top n anything for that matter) isn't set in stone.

You have to go back in and re-do name based udms every so often.

• One advantage Prob Expressions have over name based udms is that with a Prob Expression:

Once you get a UDM based on scoring of Prob Expression results dialed in -- you can pretty much leave it alone.

That's because a Prob Expression enables JCapper to learn from the data who the top n riders are.

Or learn what the top n anything is if you are looking at something other than rider performance.