Both KLASSLEADER & ELMONJED disappointed yesterday but AL AASY won 4/9I think it is more about the last couple of weeks or sopawras , maybe it's not the case but i'm happy to stear clear until we can be sure, FRACTIONAL the latest doubt and there's a few running today.
Does trainer form really matter? Interested in people’s views.
- Thread starter tony spencer
- Start date
pawras
Stallion
I agree class/form is my mainstay but I use trainer and jockey stats to add to the jigsaw
simon boardman
Mare
Chesham
Sire
2.2 The Lost Civilisation of Woolworths
The background here is that an amateur meddler spent some time looking at maps of the locations of prehistoric sites in southern England. He decided he could see patterns in the locations and then compounded this error by extrapolating from his incorrect conclusion to assign causes for the patterns. Not ruling out extraterrestrial intervention. Rather than this bit of nonsense staying of his own personal interest a number of organisations, including the 'Daily Mail', decided it was worth printing. Another example of not trusting 'Authorities' to act in a sensible manner when it comes to information presentation. The first link in the box goes to the 'Daily Mail' article.
The second link goes to a 'Press Release' which a mathematician called Matt Parker had put together. He had studied the locations of the, now defunct, Woolworths stores using the same bogus methods. He had managed to come to similar ludicrous conclusions as the amateur but knew he had. The piece is quite funny in a dry way and quietly makes the point of how anyone that had propagated the original article was very mistaken. The piece includes the following classic line :-
"..Well, that or the fact that in any sufficiently large set of random data
it is possible to find meaningless patterns of any required accuracy."
The original 'Pointless Punditry' document listed a number of problematic issues with the way people think and how this impacts on Racing analysis and betting. We cherry-pick data which suits the theory we want to be right, we see patterns in random data which mean nothing and then are desperate to assign causes to the pattern. In the example above the random data is geographical but it can be in any form. A cloud which looks like Mother Theresa is a random set of data interpreted in the same meaningless way. In looking at sets of numbers it it easy to see meaningless patterns using the same approach. This is where it matters in racing because there is a lot of data available and a lot of ways of looking at it to convince youself of invalid conclusions. From stables which are 'in' or 'out of form' guff through to whatever your pet winner selecting theory is.
|
The background here is that an amateur meddler spent some time looking at maps of the locations of prehistoric sites in southern England. He decided he could see patterns in the locations and then compounded this error by extrapolating from his incorrect conclusion to assign causes for the patterns. Not ruling out extraterrestrial intervention. Rather than this bit of nonsense staying of his own personal interest a number of organisations, including the 'Daily Mail', decided it was worth printing. Another example of not trusting 'Authorities' to act in a sensible manner when it comes to information presentation. The first link in the box goes to the 'Daily Mail' article.
The second link goes to a 'Press Release' which a mathematician called Matt Parker had put together. He had studied the locations of the, now defunct, Woolworths stores using the same bogus methods. He had managed to come to similar ludicrous conclusions as the amateur but knew he had. The piece is quite funny in a dry way and quietly makes the point of how anyone that had propagated the original article was very mistaken. The piece includes the following classic line :-
"..Well, that or the fact that in any sufficiently large set of random data
it is possible to find meaningless patterns of any required accuracy."
The original 'Pointless Punditry' document listed a number of problematic issues with the way people think and how this impacts on Racing analysis and betting. We cherry-pick data which suits the theory we want to be right, we see patterns in random data which mean nothing and then are desperate to assign causes to the pattern. In the example above the random data is geographical but it can be in any form. A cloud which looks like Mother Theresa is a random set of data interpreted in the same meaningless way. In looking at sets of numbers it it easy to see meaningless patterns using the same approach. This is where it matters in racing because there is a lot of data available and a lot of ways of looking at it to convince youself of invalid conclusions. From stables which are 'in' or 'out of form' guff through to whatever your pet winner selecting theory is.
Last edited:
gerry
Mare
I do think trainer form is massive to be honest its like anything in life if it not working something is wrong trainers must feel same no winners in last 14 days must feel like nightmare to them and the thing is no matter if your top class trainer or class 6 trainer you still need winners in the class so i think two things matter getting winners and going by odds of your runners how many should you have had lately.
pawras
Stallion
Thanks , I should have been clear that my data is up to 06/07/26 , so the 14 day range would go back to 23/06. So that discounts the two winners on the 8th.
But the code has only picked up the one on the 3rd of July and missed the one on the 26th June, when I've looked I'm missing the Newmarket results for the 26th of June !!!
Dunno how that has gone awol but adding those into my db now. I need to write a script to find any other weird holes.
Last edited:
pawras
Stallion
the difference between correlation and causality
I'd say it's safe to assume that a positive correlation of previously winning at today's distance is a valid pattern to take note, but previous wins being on a thurs is not.
Last edited:
pawras
Stallion
which site is that?
Last edited:
Chesham
Sire
Timeform