Jockey RS DB

[George]

Wanted to report some preliminary results from the Jockey Runstyle DB that I have been working on for a couple of months.

I have been tracking jockey starts/wins by runstyle (F,E,P,S,R) by dirt sprint or dirt route for all tracks that have been running for March thru April.

Calculated win percent and roi for each item above for each jockey.
Built a simple table which I have not as yet "joined" to my main db.
Therefore must query each jockey one at a time at present.

Ran a real world test 2 weeks ago whereby attempted to handicap the best jockey runstyle matchups. Seemed I always handicapped myself "off" the winning matchup. Upon review it seemed best to not handicap but simply wager the jockey matchups with high percent being best and ties broken by roi. That looked good on paper. Surprisingly top 4 rankings all showed a paper profit.

Have played that approach lightly ($2 win) over last 4 days.
Rank1........BET $46......PROFIT $11.80
Rank2........BET $46......PROFIT $129.20
Rank3........BET $42......PROFIT $40.60
Rank4........BET $26......PROFIT $52.20
80 BETS 18 WINS 22.5% ROI 246% AVG MUTUEL $21.87

Bets by rank vary because not every race has 4 jockeys with success in a runtype.
Tracks played were HAW, MNR, LS (good longshot tracks).

It is too early in my testing to say if this is real, but sure looks exciting so far.

[Glen]

George,

Outstanding prelim work. Sounds really inspiring. I have made some minor work, in that direction and will hopefully pick up on it one day.

[Carl]

Interesting George, thanks for the post.

I seem to remember that you started out this project with Qpts and then moved over to running styles.

Dispite my "fame" or being written up in in Barry Meadow's newsletter for my analysis or ROI and running styles, I really don't trust the designations that much, and feel more comfortable with Qpts. Do you feel better using running styles or were there other considerations in doing your study?

later,
Carl

[George]

Carl..you have a good memory. Have kept jockey stats by qpts for many years. Tried the runstyle just out of curiosity.

At least short term, am much more impressed by the runstyle db, but not sure I can explain why. Gonna try below.

It appears visually in scanning the db that most jockeys have not only a specific runstyle that they do better with, but that they also seek out and primarily ride that runstyle animal. For a given race type most of their starts are on the runstyle that they have the highest win% with.

For some reason I do not see that pattern in the qpt db. Maybe because there is finer seperation that pattern is not obvious but may really be there inthe qpt.

In over seven years have never found a pattern in the qpt db that you could rank and bet on with a positive roi expectation. Used that info simply to pick between two closely matched horses. Usually wagered on the one whose jockey rode best on it's matching qpts.

Live testing of the rs db has been a blast. Think I now know where all those strange high odds place horses that kill my exacta's come from.

Going to automate the thing this weekend so testing will go faster.
Doing one jockey query at a time is wearing me out.

[Glen]

The folk over at HDW, Jim Cramer and his RS-POS group track how thier jockeys do according to thier version of Ev1.

ie How abc Jock does with
ev1
ev2
ev3
ev4
ev5 and so on.

Wonder if that might be a better measurement than RS. I dunno...

[George]

Glen......they must translate that internally as the reports show E2, P5, S3 etc.
I thought about doing that with the HTR F,E,P,S,R codes by adding the FR1 rank as a suffix. Decided that would end up very close to QPT's plus would be a lot of work.
Being very lazy just decided to test pure runstyle codes. May go back and try the other
later.


Update on jockey runstyle test thru last night.
Rank1........BET $100......PROFIT -$24.20
Rank2........BET $100......PROFIT $128.20
Rank3........BET $94........PROFIT $92.20
Rank4........BET $68........PROFIT $51.40
181 BETS 33 WINS 18.2% ROI 68% AVG MUTUEL $18.47

Made a mistake on earlier posting, roi was really 146% not 246%.
Has dropped from 146% to 68% but still nice return from betting up to 4 horses.

[hurrikane]

George....nice stats.
Curious...are you actually betting up to 4 horses in a race. Also....thinking of tracking them for ITM for exotic play. Have you done any of that yet?

[George]

As Glen would say......yeop! Most races have only 3 or 4 jockeys that have a success rate on their mounts runstyle. Have been $2 win betting up to 4 of them per race. If there are more than 4 have just bet top 4 by percentage rank. Have hit winner on 33/50 races while actual betting 181 bets. 181/50 equals 3.6 bets per race avg.

Am tracking the place and show as well. Have hit 2 monster exacta's boxing top 3.

Have also developed 2 spot plays around RS and nESD that are doing MUCH BETTER than above stats.

Trying not to get too pumped up! Figure it will fall apart about the time I raise the bet level.

[George]

Realized that my answer to Hurrikane about playing 4 horses
was misleading if you had not read earlier in thread. Am playing
multiple horses only on selected longshot tracks. Doing that at
any track would be a disaster.

I am becoming convinced that jockey/horse RS matchups are
important based on 2 spot plays using it with espeed diff greater
than plus 8. Pulled in old data and created a six-month All_impc
db joined to jky/rs db.

Checked the spot plays by month. Averaged 7 plays a day on all
tracks. Worst month hit 42.6% win with 2.07 roi. Best month
hit 46.9% win with 2.18 roi. Like the narrow range.

Automated spot plays Thursday thru Saturday this week have hit
10/22 for 45.4% with 2.00 roi. Not too bad going forward.

It is way too early to tell if this really works. Have seen many
approaches look good early and fall apart going forward.
If by some chance this is solid after extensive testing will make
the database available or at least write a detailed desc of how
to make your own.