Jockey Comparison Tool - Horse Racing Forum - PaceAdvantage.Com

Running Amok · 10-04-2023, 10:40 PM

Trainers that ship a horse in usually do not bring the jockey that's been riding the horse with it. Yes, sometimes, especially in stakes races, you may see the same jockey that rode the horse last time riding the shipper again. But for most other races it's not the norm.

So my question is, how can quickly compare whether the jock riding today would be considered an upgrade, downgrade or a neutral move?

The only place I found where you can perform comparisons is the Equibase comparison tool and perhaps I'm doing something wrong but it really doesn't tell me what I wanted.

For one, it doesn't even list the win % or in the money %. Nor does it rank them by the track they mostly ride at.

Also, at the bottom it explains how to share the results but it doesn't work:

"Note: To share the comparison results, right click on this link - Compare Results - and choose "Copy link address"

Here is a comparison I did and even though the numbers seem much higher for one jock compared to the other, that could be simple due to Corrales riding in races with larger purses. Again, no win%, ITM %, or Track Rank. And some of those columns I'm not even sure what they are measuring. For example it lists the Starts (Sts) and then lists Starters (Strs) and then it lists Starts Per Starter (St/Str).

compare.PNG

Why would I even need that info and what is the significance of those stats?

I just wanted a simple way to quickly evaluate if the jockey riding today is better, worse or about the same as the jockey who rode the shipper last time.

But just like everything else in this sport, the answers are ambiguous and difficult to determine. Now I'm beginning to see why expensive software programs exist that only run on costly data.

Running Amok · 10-05-2023, 04:30 PM

OK, after further review I think I understand what these column headings represent:

Starts (Sts): Total number of races the jockey started in
Starters (Strs) Total number of horses the jockey has rode
Starts Per Starter (St/Str) Total races (Sts) divided by total horses (Strs)

This information is actually quite useful now that I understand it. It would be helpful if there was a column for win% but that's easily calculated.

BTW, the jockey Arrieta, despite much lower Per St$, actually has the higher win%.

Sorry for the rant in my original post. I been away from racing for some time and recently got back into the game. Needles to say I have been losing money and it's frustrating. Especially when I try and improve by doing more research and finding temporary roadblocks.

This is indeed a tough game. I've had some nice hits but still have a negative ROI. But I'm not giving up!

Running Amok · 10-05-2023, 06:07 PM

Also, for what it's worth, I just wanted to mention one more little rant...

The Starts Per Starter (St/Str) category, which represents total races (Sts) divided by total horses (Strs)...

I believe this indicates the horse the jockey rode started more than once in the time frame being evaluated. So it's basically an average of how often the jockey rides same horse again.

Why we need to know that and it's level of significance seems minimal compared to Wins/Starts.

Ahh but nope, lets include worthless info instead. Gotta love it!

ARAZI91 · 10-18-2023, 06:23 AM

One way i developed here in the UK for analysing jockeys was actually comparing "Jockey's Head to Head" statistics - data only concerns "common races" where each jockey has a ride in the race so it's a direct comparison - i spent over 6 months building matrices within Python to include Trainers, Jockeys & Sires to analyse them in this way and also built in a market function (which if you take the consensus of BFSP(Betfair Starting Price, which paradoxically is their "closing odds") is arguably "quality of ride" which calculates "Expected H2H Wins" , compares them with actual and uses an additive metric called WAx (Wins Above Expectation) - Heres an example of a comparison between our top jockey just now William Buick and a good up and comer Rossa Ryan - clearly Buick comes out on top on descriptive statistics like "H2H Win %" etc but when you factor in the market Rossa is around 8.27 wins under-rated than Buick.

jOCKEYS 2H2.JPG

ARAZI91 · 10-18-2023, 07:25 AM

One thing i have learned about jockeys (and it's common sense really) is that the higher proportion of Last Race Placed horses they get on the more their strike rate should improve - considering here in the UK where over all Flat Turf/Synthetic races the delta between placed /not placed LR is 8.68% in favour of being placed - if you keep statistics on this kind of thing again you can find jockeys who are over or under-rated accrordinf to their strike rate

Heres an example i did only a few days ago with jockeys >=500 rides for the UK Turf/AW - i split Hcaps and Non-Hcaps as the win%'s overall are different (it's the overall win % for ALL placings i use as weighting) i even add in First Time Starter winners which has a lowly weight of around 8.66 - this is done over a 3 year rolling average so requires a datafeed every day but it's a decent method of looking for jockeys improving/declining. if you have trouble reading the tables, right click then save as a JPEG then you can zoom in and out of the pic. - obviously the output is a "Rating" and as you would expect William Buick who has a 24.27% strike rate is way clear on a Rating of 224.1735 but he has the highest amount of LR Placings incuding Wins at a total of 11.30% (next best is only 8.60%) but these %s are an almgamation by counts of weighted Wins, 2nds and 3rds - regarding underrated jockeys look at Jim Crowley who has a LR Win/Place % of 6.43% but achieves a strike rate of 20.21% (only 4% behind Buick) or Andrea Atzeni who has a lowly LR Win/Place % of 4.79 but achieves a win % of 15.56.

RIDERS lr PLACINGS.JPG
RIDERS lr PLACINGS 2.JPG

ARAZI91 · 10-18-2023, 08:55 AM

And then there's the very simple WAx metric (Wins Above Expectation) - you can also do WPAx (which is WinsPlaces Above Expectation) though you need Betfairs place price data - this is strictly a "market" metric and designed to be used with Betfair' SP - hardly any overround and more accurate prices at the off but there are ways to use other markets which would involve normalizing the final Pari-Mutuel prices to extract the over-round and convert them to a 100% book (the "Shin" method is apparently the best for this)
WAx is just simply a cumulative/additive method very similar to calculating A/E where you convert all prices to a prob (1/odds) - add them up and compare with each win which gets a value of 1.00 , you then ratio the two and it spits out A/E - with Wax you don't ratio , you just let it run and run in a cumulative fashion again heres this years Flat Turf/synthetic jockeys with a "Positive" WAx - Up and coming Rossa Ryan is out in the clear with a figure of 26.70 over 800+ rides (i limited the ride count to 250 this time) which equates to an "edge of 0.030204 per ride but marginal gains and all that - it also equates to an above expectation strike rate of + 3.02% (18.67 to an expected 15.64 - rounding accounts for the 0.01 difference)

POSITIVE WAX 1.JPG

Beause the difference in sample sizes as it is a cumulative metric you should normalize them - here i've took the average of all the positive WAx's and normalized them to that figure - here you just divide the original WAx into the Count/Runs and then multiply by the normalization figure - here 412.15, it can change the rank order quite a bit

NORMALIZED WAx.JPG

This is another decent mathod to highlight over or under-rated jockeys from a market point of view. Formula for WAx is derived by calculting Expected Wins from the market (converted to probabilities 1/odds), then simply deducting the Expected wins from the Actual Wins. (aWins-xWins)

MJC922 · 10-18-2023, 10:33 AM

I once did a very elaborate study of riders based upon performance move ups with a rider switch. When it turned out the best rider in north america based upon that metric was a bozo at my local track known for being plugged in I just shook my head and moved on...

In other words based upon those results you wont be seeing North American Jockey Rankings anytime soon. All those would be indicative of is who most likely needs to be searched for Duracells.