News You Can Use -- "The Jockey Switch"

[Que]

It's time to put an end to all the nonsense that's been going on on this board over the past few days and return to the subject that brought us here in the first place, i.e. thoroughbred handicapping.

Although the little voice in my head said not to post this information, i.e. "you're giving too much information to the competition," however if we all don't share on occassion then none of us will learn anything. To begin, while updating my jockey stats this past week, I was amazed at some of the differences a simple jockey switch can make to their overall ROI, i.e. by blindly playing a jockey switching horses, or blindling playing a jockey riding the same horse consecutively. Overall, there appears to be only a slight increase in ROI (approx. 1.9%) with jockeys making the switch; but, the disparity in ROIs for many jockeys is simply amazing. For example, this year's overall results (1/1/01 to 6/24/01) are shown below.


Jockey Switch Win ROI Count Wins Win Pct
Yes 0.783 128,398 14,892 11.6%
No 0.764 113,516 15,323 13.5%


However, more revealing is whether a particular jockey benefits from making a switch or not. For example, the following shows this year's top ten jockey's ROI for each of the last three years. (Note. My data may not be 100% accurate/complete, but its close enough for handicapping purposes):

2001 Ten Leading jockeys (by number of wins)

Jockey Switch Win ROI Mounts

Dunkelberger Travis L
2001
Yes 0.909 401
No 0.973 285

2000
Yes 0.779 420
No 1.189 236

1999
Yes 0.711 124
No 0.862 61

Observation: Travis Dunkelberger does better when riding the same horse for consecutive mounts.

Dominguez Ramon A
2001
Yes 0.906 463
No 0.937 351

2000
Yes 1.220 958
No 0.807 625

1999
Yes 1.013 468
No 1.019 340

Observation: Inconclusive, except for 2000--wow... what a difference that year was.

Baze Russell A
2001
Yes 0.755 393
No 0.724 312

2000
Yes 0.758 845
No 0.849 640

1999
Yes 0.877 767
No 0.740 512

Observation: Don't bet Baze until he can at least beat the track take.

Lumpkins Jason P
2001
Yes 1.192 390
No 0.884 309

2000
Yes 1.252 610
No 1.067 255

1999
Yes 1.057 503
No 1.063 284

Observation: The "little voice" really wanted me to delete this query.

Day Pat
2001
Yes 0.933 381
No 0.854 266

2000
Yes 0.834 715
No 0.739 503

1999
Yes 0.693 716
No 0.884 534

Observation: Over the last two years, Pat Day has done substantially better after switching horses--not so in 1999 though.

Velazquez John R
2001
Yes 1.056 406
No 0.949 257

2000
Yes 1.071 603
No 0.867 468

1999
Yes 0.947 754
No 0.928 659

Observation: "Delete... delete."

Gomez Estaban A
2001
Yes 0.725 1359
No 0.886 1670

2000
Yes 0.706 359
No 1.070 423

1999
Yes 0.667 362
No 0.895 391

Observation: Don't play Estaban Gomez if he's switching horses.

Corbett Glenn W
2001
Yes 0.815 310
No 0.969 410

2000
Yes 1.145 462
No 0.959 555

1999
Yes 0.992 457
No 0.944 534

Observation: Inconclusive... but overall a pretty nice set of ROIs.

Cora David
2001
Yes 0.956 376
No 0.778 282

2000
Yes 0.525 343
No 0.598 273

Observation: Give extra credit to horses/trainers switching to David Cora.

Chavez Jorge F
2001
Yes 1.123 442
No 0.853 246

2000
Yes 0.954 822
No 0.850 604

1999
Yes 1.005 940
No 0.992 633

Observation: A pretty impressive record for such a popular jockey.


A few more observations:

- Overall, you will probably gain a slight advantage, approx. 2%, by blindly playing horse's making jockey changes--although your win percentage will probably decrease. For example try this experiment, divide your past plays into two separate sets--one set with horse's making a jockey switch, and the other set with horse's not making a jockey switch. Compute your ROI for each set--then see which set has the highest ROI.

- Although I can't yet prove this, but my suspicion is that the more popular the jockey, the less likely the jockey switch will produce a postive change in ROI, i.e. just look at Travis Dunkelberger's stats.

- Although not shown in this study, the results of a jockey switch or lack of it, is more pronounced with the lesser known jockeys.

- Jason Lumpkins must be one hell of a jockey. Also, if you had blindly played Jorge Chavez when switching mounts over the last three years you would have broken even.

Although, this simple technique alone can't make anyone a winner--it might help someone get a little bit closer. Anyway, has anyone else noticed/researched this?

regards,

Que.

[SAL]

I've noticed Jason Lumpkins only since he came to ride in Northern California. He gave Baze a run for the riding title, riding considerably higher priced winners. I cashed more than a few tickets with that particular jockey switch, though I never gave thought to the study you undertook. Very interesting angle, another viewpoint made possible through databases.

[GR1@HTR]

Excellent work Que. Jim Cramer of HDW ran a similiar test for one of Barry Meadown newsletters. If I recall correctly, it matches up with your work.

[ceejay]

Que:

As always, thanks for the data (and for getting us back on-track). For some of the samples could one or two huge mutual payoffs be strongly influencing the numbers? Can you test for this?

For example, if you show a 1.06 ROI for a jockey with 400 starts in a category, and assume a 20% win rate (80 wins) that implies a average mutual of $10.60. The same results could be a factor of an average $9.00 mutual for 79 starters and one $137 bomb. The ROI for the 79 races is 0.89 in this scenario.

[smf]

Que,

As always, good stuff.

If you could make this jock/ trainer/ age/ class specific, you may find a few nuggets.....and keep the info for yourself. You'll have earned the gained edge!

A few years back, Asmussen rode Casey Lambert on a lot of his babies, one being Heritage of Gold. In fact, Lambert was a good bet on an Asmussen well bred youngster but if he was on one of Steve's older runners, it stood no shot.

[Tim]

Que,

Outstanding stuff. Impressive dataset.

I've looked at some jockey fundementals by trying to split out competence from juice(getting mounts on good horses). Isolating factors like a jockeys performance when winning or losing by a head or less. I've always looked at switches from the trainer's side.

It would be interesting to hear about the splits:

- jockey switch with the former jockey on another mount in the same race
- jockey switch with the former jockey on another mount at a different track that day
- jockey switch with this jockey riding this horse 2 races back

I would think that splitting out jockey switches when the change was dictated by events or perceived trainer choice would be interesting.

Nice post. Thanks for sharing.

Tim

[Rick Ransom]

Que,

You might check switching jockeys where the last time today's jock rode, the speed rating of that race was higher than in the last race (with a different jockey). I checked a couple of hundred races and had good results with this.