• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Erin Question

Tufnel said:
Yes that makes sense and seems logically based. Have you tried that exercise on any of VDW's selections?

I can certainly do a query for that to test the outcomes. Though I don't think using the entire field would be very enlightening because A: VDW advocated concentrating on the business end of the race and B: Horses out of the business end tend not to be making much effort at that stage and placing become meaningless. Especially in handicaps where if you're beaten and out of the frame you might as well get well beaten and hopefully get dropped a few pounds in the weights. The amount of times I've seen horses who finished well behind another in a race where one of them was in the running for a place and the other wasn't and then next time they meet the one down the field turns it round. The Silver Cup winner for example. A week earlier not even sighted behind one who ran it's race.

So I'd try capping it at 10 places or maybe even 6. I'll give it a go.
Click to expand...
6 sounds good, since that was the limit of previous placings in the RFUD results pages.
 
JennyK said:
"Horses out of the business end tend not to be making much effort at that stage"

Agreed, and I think it is clear that VDW recognised that, for example in how he treated races which turned into matches. With only two left, once it was clear which was going to win, there was no reason why the jockey on the second should press his or her horse to get as close as possible to the winner. Why worry about being beaten 30l rather than 10l? Probably the most obvious example where distance beaten should not be taken literally is Pegwell Bay. There are others and a not wholly dissimilar logic appears elsewhere, for example with Ascenia.

The exercise of judgement in such cases is just one of the reasons VDW described his approach as methodical rather than a system.
Click to expand...
i agree with you totally jenny, i just cant shake his " and you will have the same horses as myself" comment
 
Tufnel said:
R rave154 Ok this is 4 years data on Flat Handicaps Class 2 only. The queries take quite a while to process because every runner in a race has it's last race ranked by price and checked on fin pos etc which creates some big CTE's. So I thought running it by race class would break it down into more manageable chunks.

View attachment 172011

The individual ranks, with lowest totals being the top of course show a steady decline as you go down the ranks, which is as one would hope.
The top groups start at 15% same as rank 1 obviously, and then go up in 15% jumps right up until past top 4 where it slows down. Top 6 on 82% of races won by a horse in the top 6 is possibly the peak area as top 4 is 60% but going to top 6 grabs another 22% winners.

The one issue that springs to mind though is should a negative total for the place vs market rank totals be considered the best when really what we should be looking for is whatever is closest to zero from either side. If too many horses ran better than their market rank then surely that could be a fluke result? And if too many ran below their rank then again the question is was that an unreliable race.

I'll consult the bot to see what it comes up with on that issue. Even so, just the first impressions are positive I think.
Click to expand...
oh, i forgot to mention....you should take the ABS ( absolute ) value of the difference between price-rank and finished-rank... ie as below

pr_rank =1, finished =7th , 1-7 = -6, but ABS(-6) =6

pr_rank = 7, finished = 1st, 7-1 = 6, ABS(6) = 6..

in other words, in both of those cases, the RESULT wasnt what was 'expected'

basically what were totting up and measuring is the disparity between expectation ( price_rank ) and actuality ( fin_pos )
 
Having said all of that, just to mention of course, when talking about price_Rank, its not an easy thing,

for example, say you have these horses in the betting

A) 2-1
B) 3-1
C) 3-1
D) 3-1
E) 3-1
F) 5-1

by the numbers, horse F is 3rd in the betting, but i would disagree and say he is 6th in the betting cos he has
5 other horses on shorter odds than himself, but by the numbers he is "3rd"

of course this also bleeds into "take the first 5 in the forecast " etc etc
 
No he is 6th if in any sport you finish like golf just like you have there you get 6th price money the winner the erst share all the price money down to 6th which you would get your not third and there all joint second you would end up with more than them.
 
R rave154 So the solution was to rank those totals closest to zero from either side. Took the bot 1 second to solve that and provide me with the alteration to the code to change it. As I've said before AI is brilliant for saving heaps of time on coding issues.

Anyway this looks very promising. Same set of data but very different results I'd say. By that I mean for the first few rank groups. Top 6 is much the same but top 1,2 or 3 much better. Less horses to consider.

Flat Hcps Class 1 pos v merket rank closest to zero.png
 
Last edited:
rave154 said:
Having said all of that, just to mention of course, when talking about price_Rank, its not an easy thing,

for example, say you have these horses in the betting

A) 2-1
B) 3-1
C) 3-1
D) 3-1
E) 3-1
F) 5-1

by the numbers, horse F is 3rd in the betting, but i would disagree and say he is 6th in the betting cos he has
5 other horses on shorter odds than himself, but by the numbers he is "3rd"

of course this also bleeds into "take the first 5 in the forecast " etc etc
Click to expand...
I treat ranks as if 2 are tied on 1st then the next is 3rd.
 
Tufnel said:
R rave154 So the solution was to rank those totals closest to zero from either side. Took the bot 1 second to solve that and provide me with the alteration to the code to change it. As I've said before AI is brilliant for saving heaps of time on coding issues.

Anyway this looks very promising. Same set of data but very different results I'd say.

View attachment 172012
Click to expand...
very nice numbers, esp top 3,

btw what Ai are you using? just curious
 
And the thing you said you cant shake of R rave154 about you would end up with same horse as him that can only be a system as the minute you go in to a method no mater how clear it seems we would possibly not have same horse not less you followed a clear system you had no choice and with the way our brain works i think its logical to try get a system where we dont interfere with it at all as long as it works long term , i think thats why the method comes in so we can interfere when its not looking good instead of changing the system a little.
 
Tufnel said:
I treat ranks as if 2 are tied on 1st then the next is 3rd.
Click to expand...
you could say that horses B,C,D,E are the sum of 2,3,4,5 = 14 / 4 = 3.5

since horse B only has ONE horse better than it ( horse A )......BUT he also has 3 other horses EQUAL to it...

so..

A) =1
B) =3.5
C) =3.5

etc

F)=6

see what i mean when i said its not easy

ASSUMING VDW was legit.... i doubt he would have gone to these lengths, prob kept it simple
 
for a wee (mental) break ...any of you guys heard the first "script" of "The Matrix"?

as in...in the movie...all the worlds population were ACtUALLY in pods, plugged in, to be a battery, to provide power
for the machines...

but....that wasnt the first script, the first script was way more terrifying !!
 
ok here is something to think about i done yesterday before i fell out with AI again for mixing it up about 10 times got me a winner today at nice price and out of the 3 races i done i got winner in two of them from top two in list.
So now i have to think of it on my own and not sure what i will do but the point was following him we had ability rating ,then consistency rating, then we had the betting market that ahs to be done close to time not actually at of but late morning as night before i have noticed are miles out now.
so there are three things he swore by ,now i thought about this and my question was whats the most important of the three ability consistency or odds.
ability is in past from beginning, consistency is from lately form, and market is from the here and now today of race.
I will leave it to individual to work out what they think not by being smart but just its up to them,
Now ability can be a big number, consistency the opposite a small number, and odds well we all know how they work percentage of winners fav best up obviously.
So i did ask AI and he suggested but i told him no your wrong and talking nonsense which when i told him why agreed but what does he do agree to any crap, the answer was how do we get these figures to adjust to live beside each other and thats where my idea arrived from so i done my sums well he did from what i told him and came up with my list not as a ability rating or a consistency or odds rating but a number where its all inside it and thast what i done today.
 
gerry said:
ok here is something to think about i done yesterday before i fell out with AI again for mixing it up about 10 times got me a winner today at nice price and out of the 3 races i done i got winner in two of them from top two in list.
So now i have to think of it on my own and not sure what i will do but the point was following him we had ability rating ,then consistency rating, then we had the betting market that ahs to be done close to time not actually at of but late morning as night before i have noticed are miles out now.
so there are three things he swore by ,now i thought about this and my question was whats the most important of the three ability consistency or odds.
ability is in past from beginning, consistency is from lately form, and market is from the here and now today of race.
I will leave it to individual to work out what they think not by being smart but just its up to them,
Now ability can be a big number, consistency the opposite a small number, and odds well we all know how they work percentage of winners fav best up obviously.
So i did ask AI and he suggested but i told him no your wrong and talking nonsense which when i told him why agreed but what does he do agree to any crap, the answer was how do we get these figures to adjust to live beside each other and thats where my idea arrived from so i done my sums well he did from what i told him and came up with my list not as a ability rating or a consistency or odds rating but a number where its all inside it and thast what i done today.
Click to expand...
sounds a little like......omg.....what was his name....A Baker?...where he ranked each figure, then totted up the totals for each horse?

probably....early 90's?
 
gerry said:
ok here is something to think about i done yesterday before i fell out with AI again for mixing it up about 10 times got me a winner today at nice price and out of the 3 races i done i got winner in two of them from top two in list.
So now i have to think of it on my own and not sure what i will do but the point was following him we had ability rating ,then consistency rating, then we had the betting market that ahs to be done close to time not actually at of but late morning as night before i have noticed are miles out now.
so there are three things he swore by ,now i thought about this and my question was whats the most important of the three ability consistency or odds.
ability is in past from beginning, consistency is from lately form, and market is from the here and now today of race.
I will leave it to individual to work out what they think not by being smart but just its up to them,
Now ability can be a big number, consistency the opposite a small number, and odds well we all know how they work percentage of winners fav best up obviously.
So i did ask AI and he suggested but i told him no your wrong and talking nonsense which when i told him why agreed but what does he do agree to any crap, the answer was how do we get these figures to adjust to live beside each other and thats where my idea arrived from so i done my sums well he did from what i told him and came up with my list not as a ability rating or a consistency or odds rating but a number where its all inside it and thast what i done today.
Click to expand...
btw gerry, its really easy to anthropermorphise the Ai...as in....start treating it like its a real person...i fell for that,....its just...well.....a black-box solution
 
rave154 said:
btw tufnel, your presentation and tabulation of data.........is insanely good !! i wouldnt have a CLUE how to do that
Click to expand...
It's just the sql editor I use. It exports as csv or html, etc. It's been very useful for working out my racing database. The bigger problem now is the results data is getting throttled and harder to grab. When it was just a few folks doing it they didn't seem to mind. But I guess the user ship has gone up many fold with so many scrapers on code sharing sites.
 
You must log in or register to reply here.
Back
Top