Cabbie.
The 80/20 has been around on this site for a bit,probably on Flatstats for even longer,however it's like all available info,it's how you use it to the best advantage that counts.
Carry On posting.
Q.

Hi All,
I've just spent the time reading this and the PWR thread -- I'm proposing to try and systemise the selection criteria to take the leg work out of the basic number crunching each day. I'm currently working on a look up for the 80:20 Jockeys and Trainers -- forgive me if I'm asking anyone to go back over stuff I may have missed on the way through.

Early in the thread it suggests the top 10% of jockeys and trainers as a prime filter and there were inclusions of listings of those trainers and jockeys. In my model I'm using the stats from the Superform site for 2001 Flat (although once modelled I can insert any year/season). My real question is what constitutes "top"; Is it an absolute number of winners which I have as K Fallon with 166 wins out of 1003 rides (17% win rate)or best % winners which I have as D Flores with 5 wins from 13 rides (38% win rate) -- and do we care about place rates?

In an ideal "system" would we want to adjust the rating according to the actual win rate % of the trainer/ jockey rather than just saying they are in or out of the top 10%?

Hi All,
You'll probably tell me this is redundant work because it is already available somewhere -- but I've spent some time today building a systemised calculator for working out where top jockeys and trainers conicide.

My base data is the Superform figures for trainers and jockeys in 2001 Flat -- I've then selected Jockeys who have a minimum win rate of 9%and a minimum number of winners of 25, and trainers who have a minmum win rate of 9% and a minimum number of winners of 10 -- this gives me a top jockey list of 41 and a top trainers list of 75.

Against this list I can just dump the next days complete fixture list and it instantly spews out the horses where top jocks and trainers "meet".

The following is the list for tomorrow: Is this useful? I know it is only one part of Max's approach.
3 AprilHorse Trainer Jockey Meeting Time
Luxor W M Brisbourne K Dalgleish RIPO 4.20
Dispol Peto T D Barron K Darley RIPO 2.10
Simeon M Johnston K Darley RIPO 3.15
Shanook M Johnston K Darley RIPO 6.00
Dalblair J A Glover G Duffield RIPO 5.30
Chandris J A Glover G Duffield RIPO 6.00
Cadeaux Cher B W Hills Pat Eddery RIPO 3.50
Shagraan J L Dunlop Pat Eddery RIPO 5.30
Dispol Rock T D Barron K Fallon RIPO 4.20
Cereus B W Hills M Hills RIPO 3.15
Canlis B W Hills M Hills RIPO 4.55
Sheer Guts B W Hills M Hills RIPO 5.30
Thaqib J L Dunlop R Hills RIPO 6.00
Miss Mougins G Wragg D Holland RIPO 4.55
Grampian J G Given J P Spencer RIPO 3.15

Doesn't layout very well but I hope yopu can read it!

CAN-DO
i have only just seen you to-day on this thread i welcome you for having a go.

i have printed your system calculator and i will read it and come back some time to-night

going on your earlier q about top trainers/jock there is a list for train/jocks for the 80/20
when we are looking at a race the train/jock, must be in form last 14 days, they must be 10% on class,course,dist,race type,jock, if ou download the r/p go to any race, by the side of the race, you will find,top train, then top jocks, tou filter these to find the 10%

i hope this is of help
regards cabbie

CAN-DO
I WILL GO THROUGH THE RACES AT RIPON FOR WED
THESE ARE THE ONES THAT SHOULD BE FILTERED I WILL ALSO PUT IF THEY ARE 80/20
2.10 TRAINERS
BA MCMAHON 80/20
TD BARRON 80/20
M TODHUNTER

JOCKS
K DARLEY 80/20
J FANNING 80/20
F LYNCH 80/20
D HOLLAND 80/20
S DROWNE 80/20
THIS RACE IS A MAIDEN RACE AND ONLY 2 HAVE RUN

2.40 TRAINERS
KA RYAN 80/20
W H TINNING
C SMITH

JOCKS
K DARLEY 80/20
K FALLON 80/20
JP SPENCER 80/20
ALEX GREAVES
J MCAULEY
SW KELLY

3.15 TRAIN
B W HILLS 80/20
J G GIVEN 80/20

JOCKS
NONE ARE SHOWING A PROFIT
SO IM LEAVING THIS ALONE

3.50 TRAIN
BW HILLS 80/20
KA RYAN 80/20
JL EYRE 80/20
JS GOLDIE 80/20
RM WHITAKER 80/20
TD BARRON 80/20
MRS GS REES

JOCKS
K FALLON 80/20
D HOLLAND 80/20
JP SPENCER 80/20
T HAMILTON
ALEX GREAVES
P FITZSIMONS80/20
DALE GIBSON 80/20

THAT IS ALL I CAN DO BECAUSE I MIGHT LOSE THE LOT
WHAT I HAVE ALSO DONE ON THE ABOVE LIST IS I HAVE ADDED THE IV FACTOR
I HOPE THIS IS OF HELP
YOU WILL STILL HAVE TO FILTER THE TRAIN/JOCKS FOR R/F,CLASS,COURSE,AGE,DIST,JOCK,
WHAT I ALSO ADD IS THE 80/20 SIRES

BE LUCKY CABBIE

wednesday 3/4/02 ripon

4.20 dispol rock
6.00 thaqib


be lucky cabbie

Just stumbled on this thread and wondered if anyone could answer a couple of questions for me.

Also, if I can help automate any of the steps I will. The 3 race form looks do-able.


Trainer/Jock performance indicators
===================================
[ class/course/dist/race type/jockey ]

Obviously derived from the rp website.

How many years are you meant to go back???
Just use previous season and completely ignore current and other previous seasons???


Can-do, you'll get back slaps all round if you are kind enough to post your shortlist on a regular basis. I notice your lists are slightly shorter, which may be better or worse than the original, who knows? If you are interested in arriving the basic list using the calculations as I understand them, let me know.

Dear Cabbie,
Have tried to reply to you on the thread. It was all lost as it does sometimes. Bit too tired now. Will do it again later,
Swish

Per Amore
trainer/jockey pass all filter inc 80/20.recent form ok.
closest danger
Saloup
trainer passes most filters and just squeezes 80/20 but jockey is a non starter.

F1

Hi Cabbie and All,
The advantage a computer has over people is it loves to do drudge work accurately. For instance Cabbie in the 3.50 at Rippon you missed Pat Eddery (in combination with B W Hills)

I've also come to the conclusion that although the 80:20 is a good and accurate philosophy we are not just interested in betting on horses where the Jock or train has numerically high numbers of wins -- we're much more inteested in their propensity to win. In other words the jockey with the highest win rate of 38% doesn't feature in mine or Stephen's Top Jock list -- but aren't we interested to know when he is out on one of his rare outings?

So I haven't done it yet but I'm going to put the full list of jocks and trainers back in my system.

What I have done since yesterday is show the %win of the Jock and Trainer and then their COMBINED win % and then I've made a cut off of a combined win rate of 20%. This then gives you a "Human Factor Rating" for the horses in a race and does more than simply say they are or aren't top jock/trainer combinations -- it gives a feel for how top they are. Let me know what you think and if you want me to continue this line of thinking.
Horses Trainers Jockeys Track Time JockWin%TrainWin% Comb%
Moth Hil J L Dunlop R Hills LEIC3.25 0.18 0.17 0.35

Budoor J L Dunlop R Hills LEIC4.00 0.18 0.17 0.35

Rasid E A L Dunlop R Hills LEIC5.00 0.18 0.17 0.35

Izmail E A L Dunlop K Fallon LEIC 4.30 0.17 0.17 0.34

Old California J L Dunlop Pat Eddery LEIC3.25 0.16 0.17 0.33

Snow Leopard J L Dunlop Pat Eddery LEIC 4.00 0.16 0.17 0.33

Greater Glory J L Dunlop Pat Eddery LEIC 5.00 0.16 0.17 0.33

Systematic M Johnston K Darley LEIC 4.00 0.17 0.15 0.32

Revelino E A L Dunlop G Carter LEIC 4.00 0.12 0.17 0.29

Rose DOr J L Dunlop G Carter LEIC 5.00 0.12 0.17 0.29

Afeef E A L Dunlop W Supple LEIC 5.00 0.11 0.17 0.28

Eastern Breeze P W D'Arcy D Holland LEIC 5.00 0.16 0.11 0.27

Wayyak J W Payne J P Spencer LEIC 2.50 0.15 0.12 0.27

Castle River B W Hills M Hills LEIC 4.00 0.12 0.14 0.26

Donegal Shore B W Hills M Hills LEIC 4.30 0.12 0.14 0.26

Espada J A Glover D Holland LEIC 5.30 0.16 0.10 0.26

Kaparolo Mrs A J Perrett S Sanders LEIC 5.00 0.13 0.13 0.26

Sky Dome M H Tompkins G Duffield LEIC 5.30 0.12 0.12 0.24

College Rock R Brotherton J P Spencer LEIC 5.30 0.15 0.09 0.24

Hmmm could get used to this picking winners lark
Per amore
1st 6/5

Dear Cabbie,
I was well aware when you talked about adding up form figs that you did not mean last run only.
I don't record my horse's last 3 figures, but what I was trying to say was suppose my horse had, say, 128 as its last 3 form figs but another one had 121. Well if you added up first figs =11, whilst second horse =4. So if you were choosing between the 2 you would go for the one with 4.
Well I am saying that you will lose a lot of winners that way (in my opinion).

Regarding class, the way I do it is based on knowledge and experience and there are no exact rules, but I prefer prize money drop to A,B,C, etc. For example if you saw a horse running in an E and then was in another E race next time most people would think no class drop. But if that horse was in a £4000 E last time but is now in a £3000 E I would count that as a class drop (although not a big one). On the other hand if a horse was in a £40000 A last time and is now in a £35000 A this time I would not count that as a class drop. But as I say I don't have an exact table or anything like that, it's just experience.
But some are simple to see. e.g if a horse has dropped from a £10000 to a £2000 well it's obvious. As for which race to use, well as I don't want to go into long and complicated detail that I might find hard to explain anyway, why not just look at a horses last race for the time being. I would suggest if a horse passes the filters AND has a class drop as well it has more chance than one in same class or up in class.

I hope that doesn't sound like "clear as mud" stuff,
Cheers
Swish

Further to the above as I realise it is difficult to explain, perhaps as a starting point it might be better to use A, B,C,D, just make sure prize money is down as well. I must add, that not all people agree with the money thing it just suits me,
Yours
Swish

And a little more to the above.
Today you had Dispol Rock. Last time ran in £2180 (F). Today ran in £3584 (E)=UP in class.
Also Thaqib Last time £4192 (D) Today £4225 (D) = SAME class.
Now I realise it is easy for me to just pick 2 results and say "I told you so" but that is not my intention. We need a lot of results to see if only doing ones dropped in class improves things, I just think it will,
Yours
Swish

Swish
Cant remember where I saw the stats but horses up in class showed a good return in certain classes/race types.I tried a down in class/speed system that proved rather fruitless so I dont think down in class is necerssarily a major consideration as a filter.
F1

Oh well,
Just an idea, never mind.
Swish

I shall leave you to it.
Swish

The thing about dropped in class as I pointed out in my thread about 4 day winners is that it is supposedly made easier for them to win by dropping them in class with lesser opposition a big advantage when done right.
Im wondering whether the stats you thinking of is the one I think that Raceform did where horses that won readily, easily eased down etc then upped in class gave a good return because of enhanced SP for being upped a step in class.
Regards.......

BRICK

Brick
From what I can remember the stats I saw were on one of the boards and came from a respected source (stephen or Adrians site or somewhere similar) where they were compiled from straight results.
I also saw figures for class drop by money on one of the other boards but cannot find the thread agin.
It basicallly went along the lines of a higher percentage of returns were gained when a drop of a certain amount was present (think it was somewhere between £2000 and £10000 drop) anywhere outside this gave poor returns.This was also allied to another selection criteria for the strike rate given.
I personally have noted in the past that a drop in class does not guarantee a good strike rate.This is from monitoring my speed system and just noting classs drops in general.I am not denying that a class drop when added to a selection that has already passed the filters would only srengthen the selections chances but I do not think that a filtered selection should be overlooked in favor of a horse that is a class drop.
One must allow for the obvious though.I asked Max his opinon on an all weather race where the fav was dropped from A to E claimer but all the stats showed the trainer/jock to be v poor at course etc.MAx advised against the horse in favor of a regular a/w horse with known form but the fav romped in.I dont claim to be anywhere as knowledgeable as Max but I was a bit surprised at his answer.Such a class drop must make up for t/j shortcomings but the price was fairly inhibitive to most.
F1

hope i am doing this right,my husband max has e-mailed me and asked if i would type a message to this gummy board,i have no idea what any of this means 2.35 ain westender to beat in contrast
3.10 ain native upmanship and a e/w on fadalko
4.50 ain iznogoud with barton and artic jack to run him close
2.25 sed bergamo
3.00 sed keen to the last
3.35 sed island warrior
4.10 sed cill curnain

hi everyone at gummy
not able to post up so the good lady doing it for me,hope you are all winning.
the above horses are ones that pass filter but as i have no access to the r/p online i have had to do it the old fashioned way.the only thing i would be concerned about is the ground as it is drying out and that may not suit all of above but at the time of writting all i can go on is the info at hand.
good luck and be lucky,max. but here is the exact message-