[AusRace] Jockeys generally
L.B.Loveday
lloveday at ozemail.com.au
Fri Jan 25 12:03:59 AEDT 2019
I find analysing trainers, let alone incorporating them into ratings, much
more problematical than jockeys. The data I receive is far less consistent,
and while I hunt down who the jockey actually was when there is the all too
often discrepancy, and ensure, as far as I can, accuracy and consistency, I
don't undertake the same effort wrt trainers.
>From Roman's 7,516 runs for Weir, I figure he's using data from about the
start of 2017, which raises the question of how far back do you go - in both
time and runs.
Here's some trainers in my data who may be the D Hayes Roman refers to -
D Hayes
D J Hayes
D & B Hayes & T
David Hayes
David Hayes & To
D, B & T Hayes &
Apart from the difficulty of determining who's who, there is the question of
who is actually training and the continuity of "dynasties" - CS, Peter and
David Hayes, the Cummings's, Hawkes's . When an heir apparent takes over as
the listed trainer, do you start rating afresh?
How different from jockeys - there is just one, discrete jockey on each
runner. MBeadman did not continue the riding of his father - he started from
scratch and there was no phasing in or continuity from D to B.
I presume these are all for Gai Waterhouse, but it's beyond the ability of
this number cruncher to be sure:
G Waterhouse & A
Ms G Waterhouse
G & A Waterhouse
Then, what about when Gai is at Royal Ascot or on a skiing holiday and she
does not run her expert eye over the horses or give final instructions to
jockeys? You can be the named trainer while 20,000km away, but is the same
job done? With a jockey, who is named is who you get.
D K Weir is at least unambiguous in my data, BUT, I looked at several
public-domain sites and their figures are all over the place. So, I looked
at darrenweirracing.com and they claim 3535 winners at a strike rate of
20.58, from which I infer 17,177 runners. BUT, my data has D K Weir with
24,740 runners since 1/1/1995 for 3,896 winners (I can't be bothered looking
at my archived data to go back further).
I'll leave rating trainers to those who have personal, as distinct from
digital only, data.
LBL
From: Racing <racing-bounces at ausrace.com> On Behalf Of Roman
Sent: Friday, 25 January 2019 8:51 AM
To: 'AusRace Racing Discussion List' <racing at ausrace.com>
Subject: Re: [AusRace] Jockeys generally
Hi Tony,
I respect the fact you have your way that works for you as Sean has his.
That's how the punt goes for those keen enough to go past "pluck a duck"
with a cursory ten minute look or listen to various tipsters.
The one constant I can quote is that the racetrack market is what I call
"linear". I assume that's the correct term where I mean favs win more times
than 2nd favs who win more than 3rd favs and so on. Thus the SP figure is
linear in that $2.50 chances win more than $4 chances who win more than $7
chances and so on. I am sure we all agree that this general premise is
correct in the high 90's percentage wise.
Therefore, the rating of jockeys and trainers can be aligned to this premise
and their LOT or POT should give a reasonably clear picture of whether they
are up to the market assessment. Where this can go asunder a fraction is
that top trainers runners are overbet by a lazy public who think the likes
of Waller Hayes and Weir can train every favourite to win. As most on this
site realise many of their horses are "unders". Nevertheless that can be
factored in.
In the file I have DKWeir 7516 runners for minus 23.8%LOT, D Hayes 4710 for
-17.6% LOT however at $3 or less Weir 1053 runners for -5.1% whilst Hayes
with 529 runners is -11.2%.
So, if betting all odds, as I assume you do, you would use the larger set
you would credit Hayes with more points. The favs punters would give DKW a
better figure.
These figures are, of course, open to all sorts of personal interpretation
if I add that overall from 7513 runners at $3 or less covering all trainers
the LOT is 8.1%. I am not sure but would 5.1 divided by 8.1% give a figure
or should it be vice versa.
Naturally, a similar process for jockeys would find some riders of $3 or
less chances, for instance, better than others. From there some
jockey/trainer combos would be another facet i.e. Yendall/Weir, Allen/Weir,
Bowman/Waller et al but a downside for some combos would be not enough runs.
However, all said above is just one way!!
Cheers
Roman
From: Racing [ <mailto:racing-bounces at ausrace.com>
mailto:racing-bounces at ausrace.com] On Behalf Of Tony Moffat
Sent: Friday, January 25, 2019 1:59 AM
To: <mailto:racing at ausrace.com> racing at ausrace.com
Subject: Re: [AusRace] Jockeys generally
Len - thanks
Kozzi's assertion that the iv are poles apart does not hold up here - I
have similar scores to yours. I leave mine at 1.15 for Guymer and you
promote him by multiplying by a hundred (de-decimate?) to get 115 (I guess).
I wanted a score in the here and now and that is how/why I came to derive
the iv, it is contemporaneous with other riders in this race, their presence
affects its score somewhat, a little, and never majorly. It is not uniquely
mine, by the way. It involves the use of all the placings, I had included
the win record only, then added second place(s) to see how that ran and have
reverted to this input now.
My calculations are in the mould of 'ok, what can you do', looking forward,
and others can be described as 'look what I done'.
The inclusion of performance at price bands might be the best but I don't
have that data, the prices of past endeavors. I can access it, the prices,
but choose not to manually enter it, and who would do that.
Yes, I do iv for jockeys (as you know) and also trainer, horse, distance and
form and multiply these to get a value for each runner - highest is best.
Form is a two part process. I involve their last 4 runs by multiplying the
places together, remove the worst result, then rank that - this appears to
be strong information, and has always been.
As a factor in a weight rating process used, I again involve their places
but this time I start from a base of 9 (the worst there can be) then
subtract each succeeding run from the previous product until I get a score
from which I can calculate a rating to win. So 6214, comes out as -3,-4,-1,3
and when summed this is -3+-4+-1+3 = -5. The -3(minus three) came from 6-9 =
-3, the -4 (minus 4) came from 6-2 = -4, the -1 (minus 1) came from 2-1 =-1,
and the (+)3 came from 4 minus 1 = +3. The -5 for this runner, and the
calculated scores for all runners is then multiplied by 1.5 to give a weight
rating variation and this product is then added to the limit weight for this
race and the allocated weight deducted from that. The best result, the
highest/biggest number resulting from that is considered the best for this
race, and you can zero that against the other calculated weights to sort out
the weight rated best ranking.
I use a variation of this method in my own punting, having streamlined a few
of the calculations, but the principles are the same, and the selections
also. I back more than one runner in each chosen race, often a quinella now,
and for several years, with a saver on some of the quinella inclusions.
I don't use or include the iv selections in my punting yet, I may do soon,
and include it here only for information and comment.
Cheers
Tony
FROM THE ARCHIVES
From: <mailto:ausrace-bounces at ausrace.com> ausrace-bounces at ausrace.com [
<mailto:ausrace-bounces at ausrace.com> mailto:ausrace-bounces at ausrace.com] On
Behalf Of Nick at Twonix
Sent: Thursday, 5 November 2015 1:29 PM
To: 'AusRace Mailing List' < <mailto:ausrace at ausrace.com>
ausrace at ausrace.com>; 'L.B.Loveday'
< <mailto:lloveday at ozemail.com.au> lloveday at ozemail.com.au>
Subject: Re: [AusRace] Michelle Payne
I did an analysis of 271 K Aus races rides over last 2-3 years and
discovered that Male jockeys have a 2% better strike rate and a 3% better
A2E (think POT betting to prices).
However Apprentices ( both Male and Female) have the same Strike Rate and
A2E . Licensed Male jockeys have a 6% better A2E compared to Female jockeys.
Category Rides Wins S/Rate ExpW
A2E
Aus Races 271,662 35,340 13% 40,474 -13%
Female 40,478 4,626 11% 5,448
-15%
Apprentice 21,840 2,549 12% 2,930 -13%
Licensed 18,638 2,077 11% 2,518 -18%
Male 231,184 30,714 13% 35,026 -12%
Apprentice 54,329 6,789 12% 7,840 -13%
Licensed 176,855 23,925 14% 27,186 -12%
AN
Len, I was able to distinguish Female jockeys in AAP data as they all start
with "Ms ". I am assuming that MS Dhoni doesn't ride in Aus :-)
From: Racing [ <mailto:racing-bounces at ausrace.com>
mailto:racing-bounces at ausrace.com] On Behalf Of L.B.Loveday
Sent: Tuesday, January 22, 2019 9:50 AM
To: 'AusRace Racing Discussion List' < <mailto:racing at ausrace.com>
racing at ausrace.com>
Subject: Re: [AusRace] Jockeys generally
"Raw wins and wins and placings" don't mean much in absence of prices - it's
easy to back winners; just back every runner at 1/1 or less and you'll back
around 56% of winners, and "just" lose about 5.5%.
Nor is just looking at past returns enough - factors such as those you list,
and eg, track, trainer should be considered.
Here's a simplistic look at some figures that could be used:
Considering the last 1000 rides for jockeys who have had at least 1000 rides
in the past 14 years (a somewhat different picture arises if only
considering since the advent of SOP rather than traditional SP as SOP
markets have lower market%s, especially away from Sydney/Melbourne tracks):
Best returns @ SP:
SThornton 101
MJWalker 103
WD'Avila 103
CParnham 104
VWong 104
DMoor 105
PWells 105
DWBallard 107
SFawke 113
SGuymer 115
JOliver 117
Considering only rides on horses "in the market" - gets rid of outliers like
125/1 winners:
JPStanley 100
JPracey-Holm 100
JTaylor 100
MWeir 100
RFradd 100
RonStewart 100
KWalters 102
SLisnyy 102
LJMeech 103
TPannell 103
CGallagher 104
RMaloney 106
CHall 107
BWerner 108
DWBallard 108
JLyon 109
PWells 109
SThornton 109
CNutman 110
VBolozhinsky 112
Worst returns @SP:
LGHenry 21
JeffKehoe 31
DPitomac 33
TJeffries 33
SBayliss 34
JMissen 36
MJStephens 37
ABadger 38
NRose 38
SStarley 38
ECockram 39
JKeating 39
MHackett 39
RYetimova 39
SParnham 39
Considering only rides on horses "in the market" (as I've previously said
LGHenry is in a class of her own):
LGHenry 27
MJStephens 32
SBayliss 36
CBryen 41
JMissen 43
SGalvin 45
SStarley 45
ABadger 46
DPitomac 46
BPowell 47
MHackett 47
SParnham 47
BStower 48
PaulPayne 49
CQuilty 50
The big gaps - All "in market"
SFawke 113 79
WD'Avila 103 75
MJWalker 103 76
JOliver 117 91
BMertens 88 63
JTaylor 69 100
NPunch 60 95
JeffKehoe 31 72
SLisnyy 61 102
CHall 66 107
VBolozhinsky 70 112
From: Racing < <mailto:racing-bounces at ausrace.com>
racing-bounces at ausrace.com> On Behalf Of Roman
Sent: Monday, 21 January 2019 9:34 PM
To: 'AusRace Racing Discussion List' < <mailto:racing at ausrace.com>
racing at ausrace.com>; <mailto:tonymoffat at bigpond.com> tonymoffat at bigpond.com
Subject: Re: [AusRace] Jockeys generally
Hi all,
The fascination of it all is that two raters could have the same jockey
literally poles apart depending on criterias chosen.
I have never rated jockeys nor trainers as I wonder if there is all that
much between a number of them at the top level. If the SP figures is a solid
determinant of the overall structure of horse racing does it not figure
those jockeys that ride well on well fancied horses are giving the horses
the chance of winning the market determines. Say Jockey A has 100 rides in
races in town on favs and scores 35% of the time is he not a fraction better
than Jockey B who rides 32%. So the next time the two jockeys meet on say
favs at 2/1 and 9/4 (close) but the 32% jockey rides an on pacer and the 35%
jockey rides a chronic get back type where does the ratings look now. It
would be best to rate them all on their ability with leaders, on pacers, mid
fielders and get back types and another set of figures comes up far more
accurate, imho, than just a raw wins and wins and placings.
I look forward to Len's reply.
Roman Koz
From: Racing [ <mailto:racing-bounces at ausrace.com>
mailto:racing-bounces at ausrace.com] On Behalf Of L.B.Loveday
Sent: Monday, January 21, 2019 6:12 PM
To: <mailto:tonymoffat at bigpond.com> tonymoffat at bigpond.com;
<mailto:racing at ausrace.com> racing at ausrace.com
Subject: [AusRace] Jockeys generally
Tony,
Did not get to me and I just saw it in the archives - a very different
rating method to mine; I'll evaluate and comment anon.
LBL
790*150-93-96 is the revealed racing stat for Linda Meech tomorrow - to
expand this Ms Leech has had 790 rides for 150 wins in the time frame
covered by this stat. My IV for that is 1.4, essentially she is 40% advanced
on some others in this race.
No rider gets less than 1, although the calculation is often less than
that, John Keating has .6 (scores a one in the scheme). Why? - he is on a
horse in the race and Bradbury's have happened, although I use the 1 for
statistical pureness, and to get rid of some decimals. To be factual, off a
calculation, Keating is somewhere like 80% more unlikely of producing a good
ride than Meech - he has 395*17-25-33 and is .6 against Meech at 1.4 (1.4 -
.6 is the basis of the claim for 80%).
Jason Maskiell is also on 1.4 in this race, off 347*54-46-41. The factor is
0.300552251 (the average of all jockeys riding) and my fall back value is
.31 - if a jockey can't be rated (the data is missing e.g.) then I assign
that value to it early in the calculation.
Roger Biggs wrote that he used .2595, which may be the statistical base of
all jockey placings across many rides. This has changed somewhat, there is a
jockey db. on RB Ratings. I am unaware of another method to rate and rank
jockeys against all their rides. They can only ride one horse in a race so
that the iv concocted from a large number of rides seems to be correct, and
I total all the rides for all jockeys in the race then divide that into all
the places achieved by all the jockeys, and from that sub-total I
individually determine an iv.
There is a place system for ranking jockeys when on favorites, but that is
not the jockey at all. Another time perhaps. Who likes, or wants, dividends
in the sub $2 range, most of us really.
This upcoming race has riders which have achieved 4708 rides totally under
the period of review, and of those rides those riders scored, placed, in
1415. So, 1415/4708 = .300552251 is the factor to be used. Individually
Keating has 395*17-25-33 (17+25+33/395 = .1898734) and this product is again
divided by the total score .3005522512 to give the score of .6. These
numbers seem minimal, mickey mouse almost, but are a significant part of the
overall stat picture
Trainers may have two or more runners in the race. I score them the same as
jockeys, total rides into total places (123) and develop a iv score from
that.
Involving riders and trainers, getting a score from them combined, I
multiply their ivs and work with the product, ranking that.
Meech 1.4, trainer 1.3 (1.4 * 1.3 = 1.82)
Keating 1, trainer 1 (1 * 1 =1) actually .6 * .1. The trainer is yet to win
a race
Maskiell 1.4, trainer 1 (1.4 * 1 = 1.4.
Dylan Dunn = 1.1
There is some upside to Linda Meech ability, trainer ability.
This is R2 Kyneton tomorrow, a maiden and I'm not betting
in it, nor do I suggest you do.
<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_cam
paign=sig-email&utm_content=emailclient>
Virus-free.
<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_cam
paign=sig-email&utm_content=emailclient> www.avg.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ausrace.com/pipermail/racing_ausrace.com/attachments/20190125/4f515582/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 350 bytes
Desc: not available
URL: <http://ausrace.com/pipermail/racing_ausrace.com/attachments/20190125/4f515582/attachment.jpg>
More information about the Racing
mailing list