|
JCapper Message Board JCapper 101
|
| Home |
Register
Log In |
| By | New User Questions | |
| jeff 1/21/2013 10:41:35 PM | --quote: "Hi Jeff--end quote Word doc received in good order. Below find a cut and paste of questions from the word doc along with my attempt at an answer: JCapper Questions: Q. Odds lines - What are AProbOl, PScoreOL, JPRMLOL? A. When Calc Races and/or Build Database routines are run, JCapper performs number crunching. Number crunching in JCapper is no small thing. Each horse is first evaluated in terms of (quite literally) hundreds of proprietary handicapping factors that collectively represent what I consider to be racing's significant areas. These areas include the following: Early speed, late speed, class, form and workouts, ability of the horse gleaned from its speed figs, the abilities and tendencies of the horse's human connections (rider & trainer), influence from breeding, and misc. things such as energy requirements for surface and distance, the degree of chaos in the current race, and the pace-shape of the current race. Once each horse has been "scored" in each of the significant areas the scores for the horses in each race are then compared against each other. Probability algorithms that use a set of metrics based on a.) how the horses in each race stack up against each other are then combined with public opinion for each horse based on b.) morning line odds and c.) odds during the betting to generate a probability estimate for each horse. JCapper calculates a number of different probability estimates including: Aprob, JPRMLProb, UPRMLProb, UPRZScoreProb, UPROpinionProb, and BettorsToteProb. A description of each can be found on the supported factors page at JCapper.com. Alternately, after logging into the JCapper message board you can use the board's search function to find posts written about (I think) each of these probability estimates. The letters AProb refer to a JCapper factor known as Assigned Probability which is based on JRating from the original 100 copies of the JCapper program. The letters JPRMLProb refers to a JCapper factor known as JPR Morning Line Probability (also known as JPR ML Prob or JProb) which is probability based on a combination of JPR (in the Platinum program) and morning line odds. The letters UPRMLProb refers to a JCapper factor known as UPR Morning Line Probability (also known as UPR ML Prob or UProb) which is probability based on a combination of UPR in the Platinum program and morning line odds. In the Silver program, Silver Rating and UPR are one and the same thing. In the Silver program UPRMLProb refers to a JCapper factor known as UPR Morning Line Probability (also known as UPR ML Prob or UProb) which is probability based on a combination of Silver Rating and morning line odds. Q. What is the basis of their computation? A. The following simple formula can be used to calculate an odds line from a probability estimate: Odds Line = (1 / Estimated Probability) - 1 I'll use the following example to illustrate: Suppose for the sake of argument that a glance at the sql html report shows a horse to have been awarded a probability estimate of 0.2867 in the UProb column. The odds line for probability = 0.2867 is approximately 5/2 and can be calculated using the above formula as follows: Odds Line = (1 / 0.2867) - 1 -or- Odds Line = (3.48797) - 1 -or- Odds Line = 2.48797 -or- Odds Line = 5/2 (approx) Relating the calculated odds lines back to specific factor names that you asked about... A. AProbOl is the calculated odds line based on AProb or Assigned Probability. (Available in the first 100 copies for the program only.) A. PScoreOL is the calculated odds line based on JRating and Pscore. (Available in the first 100 copies for the program only.) A. JPRMLOL is the calculated odds line based on JPR and morning line. (Available in the Platinum program only.) A. UPRMLOL is the calculated odds line based on UPR and morning line. Note that in the Silver program Silver Rating and UPR are one and the same. In the Silver program UPRMLProb is the probability estimate that appears on the html report and it is based on Silver Rating and morning line odds. More to come... -jp .
| |
| jeff 1/22/2013 4:42:26 AM | Q. Is 1.00 100% of something and 0.00 0% of something, or is it something entirely different? A. You nailed it. If you are dealing with probability estimates a value of 1.00 would be the same thing as saying you believe the horse is 100% likely to win its race. At the other (lower) end of the scale a probability value of 0.00 would be the same thing as saying you believe the horse has a 0% chance of winning its race. Q. What do the various levels mean? A. The numbers on the html report's probability column indicate the JCapper program's assessment of each horse's "before the odds are known" chances to win the race. Here is a link to a text file that contains actual results from Jan 01, 2012 through Nov 24, 2012 for a few of the factors generated by JCapper Silver including UPRMLProb: http://www.jcapper.com/messageboard/reports/impact_study_hdw_2012ytd.txt Click on the above link and scroll to the table for UPRMLProb numeric value. Now, let's address your specific question: "What do the various levels mean?"... The UPRMLProb numbers on the Silver HTML Report are the JCapper program's probability estimate for each horse. The actual performance of those estimates (for an 11 month 46k race sample) is displayed in the UPRMLProb numeric value table where the data has been broken out in rows - one incremental row per "level." More to come... -jp . ~Edited by: jeff on: 1/22/2013 at: 4:42:26 AM~
| |
| jeff 1/22/2013 5:29:42 AM | HTML report headings Q. What is UPRProb? A. UProb and UPRMLProb as described above are one and the same. Q. CPace In feet per second or a ranking? uses early and late with different weights based on win rate and roi. Probably a regression or least squares curve fit. Higher is better??? A. CPace is a proprietary factor produced by the program's number crunching algorithms. The inputs fed into the algorithm are the early E1 (first call) and E2 (second call) pace figs as well as a horse's recent workouts. Higher is faster. Lower is slower. For the most part higher is better but not always - depending on the surface they are racing on. Here is a cut and paste of the factor definition from the Supported Factors page at JCapper.com: CPace - or Comprehensive Pace Rating. CPace is a compound number derived using an alogorithm that considers certain early and late pace elements, assigns different weights to each factor based on win rate and roi impacts of historical occurances, and then rolls all of the separately considered factors into a single number that reflects the true pace ability of a horse relative to the rest of the field. (Accessible via the Profile Table Interface) Q. PFit A pace figure of some kind or a fitness (form) value??? A. PFit is an abbreviation for CompoundPaceFit. Here's a cut and paste of the factor definition as listed in the Compound Pace Array section of the Supported Factors page at JCapper.com: 11. CompoundPaceFit - CompoundPaceFit is a number (existing on a scale of 0-100) that represents favorable vs. unfavorable pace matchups within a race. The higher the number the more favorable the pace matchup is for a given horse in today's race. The lower the number the more the horse is up against it pace-wise today. CompoundPaceFit is not a magic bullet - even though at times it may look like it should be. Sometimes riders assigned horses with pace advantages over their fields decide (or are given instructions by trainers) to ride the horse in a way that destroys the advantage. But overall, and especially after you've spent some time at the Data Window evaluating it, I think you will find CompoundPaceFit to be a VERY NICE addition to the JCapper factor set. -- Jeff Platt 4/10/2008 Q. EeCn??? A. EeCn is an abbreviation for Early Consensus. I did a search of the Supported Factors page at JCapper.com for Early Consensus and was kind of floored to discover that I (apparently) have never created an entry there for the program's consensus factors. If I were going to make such an entry, it might look like this: General Theory Almost all handicapping factors no matter how trivial can be classified as belonging to one of the following general categories: Pace... early speed, late speed, run style, energy distribution, etc. Class... an intangible representing raw ability. Ability from speed figs... Can the player get a handle on the ability of the horse from its final time based speed figs? Form... form cycles, workouts, condition of the horse itself. Human Connections... tendencies, abilities, and intent of riders, trainers, and owners. Misc... Everything else: track biases, breeding, post position, equip changes, preferred dist, age of horse, etc. Of course sometimes the lines are blurred... Class horses tend to run higher speed and pace figs than lesser horses. Horses in top form tend to run higher speed and pace figs than horses that are out of condition - But if you think about it for a bit, nearly every factor in a horse's past performance record can be thought of as fitting somewhere in the above mix. If you are asking about the Consensus Numbers in the Race Summary at the bottom of the html report for each race... Fig Consensus is a compilation of JCapper factors representing Ability From Speed Figs... a mini power rating that I created using the UPR Tools Interface. Form Consensus is a compilation of JCapper factors representing Form... a mini power rating that I created using the UPR Tools Interface. Class Consensus is a compilation of JCapper factors representing Class... a mini power rating that I created using the UPR Tools Interface. Early Consensus is a compilation of JCapper factors representing Early Speed... a mini power rating that I created using the UPR Tools Interface. Late Consensus is a compilation of JCapper factors representing Late Speed... a mini power rating that I created using the UPR Tools Interface. Power Consensus is a compilation of JCapper factors representing all of the above... a comprehensive power rating that I created using the UPR Tools Interface. -jp . ~Edited by: jeff on: 1/22/2013 at: 5:29:42 AM~
| |
| jeff 1/22/2013 6:30:23 PM | Q. WBrill Workout brilliance???? A. Here is a cut and paste of the factor definition from the Supported Factors page at JCapper.com: WoBrill - or Workout Brilliance. This number represents velocity, measured in feet per second, of a horse's fastest workout, at a distance of four furlongs or longer, in the past 30 days. If a horse has no workouts four furlongs or longer in the past 30 days, then WoBrill is zero. (Accessible via the Profile Table Interface and Preset Filters) Q. FmCn???? A. FmCn is FormConsensus (see my comments about JCapper's consensus factors above.) Q. FgCn???? A. FgCn is FigConsensus (see my comments about JCapper's consensus factors above.) Q. ClCn???? A. ClCn is ClassConsensus (see my comments about JCapper's consensus factors above.) Q. TPace What is TPace??? A. TPace Here is a cut and paste of the factor definition from the Supported Factors page at JCapper.com: TPace - or Total Pace. This number combines Early Pace (either E1 or E2 depending on race distance) with PAL in a unique way that can represent both current form and ability . It is taken from the most recent running line found in the horse's performance record regardless of surface or distance. It is calculated as follows: TPace = ((2 X EarlyPace) + PAL) / 3 (Accessible via Dynamic Filters) Q. Ped? A. Ped is is HDW's Distance Pedigree Rating. My understanding is that HDW is constantly compiling and updating a set of sire stats based on the performance of their offspring at various distances. (Their PED Rating is based on an HDW algorithm using sire stats that they keep.) My own testing shows that this factor can make a difference on a UDM by UDM basis. Q. Others you asked about: PAL, L3, RStr, PCls, B10, SC, RC, CRun, RStyl, QSpd, FCyc, REC, WO, DT, FSF, and FSP. A. Below I've posted a list of factor names to go with the abbreviations along with a quick factor def: PAL - PAL - or Pace Adjusted Late. This number represents the data vendor generated Late Pace rating adjusted for Pace. It is taken from the most recent running line found in the horse's performance record regardless of surface or distance. It is calculated as follows: PAL = LatePaceRating - ( 2 x Lengths Behind at Second Call ) (Accessible via Dynamic Filters) L3 - LatePace or L3 - or Late Pace Best 3 - This number represents the single best Late Pace rating found in the last three running lines of a horse's past performance record. The rating is taken only from races run at a similar distance and running surface as today's race. If the horse has not run at a similar distance and running surface as today's race then the rating becomes a zero. (Accessible via the Profile Table Interface) RStr - Race Strength (Fin Pos) - or Positional Race Strength. This number represents the average strength of the competition faced. All running lines found in a horse's past performance record are measured. All distances and running surfaces are considered. Race Strength for a running line is determined by adding the horse's earned data vendor generated Speed Figure plus 2 times the Finish Position. If the horse was the race winner, Race Strength is simply the horse's Bris Speed Figure from that race. PCls - PClass - This number represents an analysis of both the running positions at different points of call within a horse's past performance record and the class of race (using purse value) for each running line. As such it is a numerical representation of a horse's running style. Horses with 1's and 2's in the pace calls of their running lines and horses that have competed for more money than their competition will tend to have higher PClass numbers than horses whose records are devoid of early speed and horses that have competed for significantly lower purse money. PClass in JCapper is similar in nature to PCR or Performance Class Rating as introduced by the late Bill Finley in his book Total Victory at the Track written under the pen name William L. Scott. The key difference is that in JCapper I use differences in purse value to account for class rather than claiming price as suggested by Finley. (Accessible via the Profile Table Interface) B10 - B Bris (Best of Last 10) - This number represents the single highest data vendor generated (either Brisnet or HDW) final time Speed Figure found in a horse's past performance record. All surfaces and distances are considered. (Accessible via the Profile Table Interface and through Dynamic Filters) Note: Even though this factor is labeled as "Bris (Best of Last 10)" rest assured that if you are using HDW files, this factor is being generated by the JCapper program using HDW speed figures as read from your data files. Alternately, if you are using Brisnet files, this factor is being generated by the JCapper program using Brisnet speed figures read from your data files. SC - Sprint Cycle - This is a proprietary form indicator for horses in sprint races. (0=false 1=true) RC -Route Cycle - This is a proprietary form indicator for horses in route races. (0=false 1=true) CRun - CantRunXFormAttrib - Not listed on the supported factors page but here are two links to write ups that should give you an idea as to what the CANTRUNXFORMATTRIB is all about: CANRUN/CANTRUN Filters - June, 2005: http://www.jcapper.com/Article_JUN2005.html CANRUN and CANTRUN Preset Filter Codes: http://www.jcapper.com/MessageBoard/TopicReader.asp?topic=1264&forum=JCapper%20101 RStyl - Run Style - RunningStyle - A text descriptor describing a horse's preferred running style. (Accessible via the Profile Table Interface)
HDW Runstyle is a bit more comprehensive. Here is a link to a post on the JCapper Message Board describing HDW RunStyle in some detail: http://www.jcapper.com/messageboard/TopicReader.asp?topic=1172&forum=JCapper%20101 QSpd - QSpeedPoints - This number represents Quirin style Speed Points. Speed points exist on a scale of between 0 and 8. The higher the number, the more likely it is that a horse will try and get the lead. The lower the number, the more likely it is that a horse will not contend for the lead. (Accessible via the Profile Table Interface) FCyc - FormCycle - Here, the factor def includes tables that I can't cut and paste with any hope of getting them to line up. I have to point you to Supported Factors page at JCapper.com. Hint: FormCycle is (currently) factor #27 there. REC - Recent Activity Days Last Start - This is simply the number of days since a horse's last race. (Accessible via the Profile Table Interface) WO - Recent Activity Days Last Work - This is simply the number of days since a horse's last workout. (Accessible via the Profile Table Interface) DT - DtShift - or Distance Shift. DtShift is a number that represents the shift (or change) in furlongs from the distance shown in the running line of a horse's most recent race compared to the distance in the stated conditions for today's race. Hints~ DtShift can sometimes be used by the astute player when creating trainer specific UDMs. Some trainers excel when stretching horses out or shortening them up. Other trainers have no success whatsoever when their runners switch distances. ~ DtShift can also be used by the astute player when creating UDMs where it is desired that a horse being considered as a contender ran a distance in its last race very close to the distance of today's race. This can be done to prevent, for example, inclusion of milers or sprinters as a contenders if today's race just happens to be one of those marathon mile and a half events. (Accessible via the Profile Table Interface) FSF - Fast Slow Final - FastSlowFinal - (JCapper Platinum only) - This factor is an offshoot of the same research that produced FastSlowPace. FastSlowFinal is the result of comparing the horse's final time based speed fig from its most recent race against par for the level. The theory behind this is that horses coming out of exceptionally strong races tend to end up with running lines that are distorted and ugly. Such running lines may not be representative of what the horse may be able to achieve in today's race facing different competition. I experimented by comparing the par values used by Bris in the fields in the data file listed as "Bris Race Shape" against a proprietary JCapper par calculation that I devised. And large sample data studies clearly show that the proprietary JCapper par calculation produces a better roi than Bris race shape pars - so that's what I used. FastSlowFinal is a fully grab-bable factor in the UDM Wizard. (JCapper Platinum only) FSP - Fast Slow Pace - FastSlowPace - (JCapper Platinum only) - This factor is the result of a little research I've been doing where pace of the most recent race is compared against par pace for the level. The theory behind this is that horses coming out of fast pace races tend to end up with running lines that are distorted and extra ugly. Such running lines may not be representative of what the horse may be able to achieve in today's race when facing different competition or a different pace match up today. I experimented by comparing the par values used by Bris in the fields in the data file listed as "Bris Race Shape" against a proprietary JCapper par calculation that I devised. Large sample data studies clearly show that the proprietary JCapper pars produce a better roi than Bris race shape pars - so that's what I used. FastSlowPace is a fully grabbable factor in the UDM Wizard. (JCapper Platinum only) Whew! -jp .
| |
| jeff 1/22/2013 6:38:09 PM | --quote: "Also, when I click on PPS from either the menu or the button, I get a help file, not PPS's. Do I need to change a setting or are these not available in the Silver version?"--end quote My first best guess would be to try the following: 1. After loading card files into JCapper select a loaded track code from the track codes drop down. This will cause the PPs Generator to load that track code. Note: The video is a bit dated and shows an older version of the PPs Generator in action. The video shows the module rendering R1 whenever a track code is selected. I disabled that behavior on purpose several months ago and have not found free time to cut a new video. 2. Once a track code has been loaded into the module (in step #1 above) - to generate past performances for a desired race: Simply select that race from the Races drop down - and the module should render that race automatically. Let me know if that does the trick. -jp .
| |
| jeff 1/22/2013 6:46:22 PM | --quote: "Jeff, Is there a way to display HDW PSR in the Past Performance Generator?"--end quote Yes there is. HDW PSR is rendered on both the playlist file mode (default) HTML Report and the SQL Mode HTML Report. Both are created when you run a Calc Races. Taking things a step further, If you run a Calc Races with the program in SQL Mode, and THEN use the PPs Generator to render past performances... The PPs Generator will render past performances with the JCapper numbers table from the SQL HTML Report embedded in the past performances for each horse -- allowing you to see all of the JCapper numbers - including PSR. -jp .
| |
| rkskipster 1/23/2013 5:32:46 AM | Jeff--- is uprml the same thing as uprmlol? If not how do you test for uprmlol in dwindow? Can you tast for PscoreOL in the data window?? thanks rick
|
| Reply |
|
Copyright © 2018 JCapper Software back to the JCapper Message Board www.JCapper.com |