Smartgambler
Pro-Punter

Go Back   OZmium Sports Betting and Horse Racing Forums > Public Forums > General Topics
User Name
Password
Register FAQ Search Today's Posts Mark all topics as read

To advertise on these
forums, e-mail us.

Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 11th July 2020, 09:38 AM
chook chook is offline
Member
 
Join Date: Jan 2019
Posts: 21
Default archie is problem

I saw archie mentioned in Gritz' post, keep it simple. and i was looking for a test which
would give me an idea of how credible my data can be.

here is some real data
18 race selections all were favorites:
1.9,1.7,2.65,1.85,3.0,1.8,1.7,4.0,1.8,2.4,2.2,1.8, 1.6,1.7,1.9,1.5,1.6,1.6
they popped up as most likely to win on 18 different race days. AKA specials or outstanding prospects.

10 winners:
1.7,2.65,3.0,1.8,2.2,1.8,1.7,1.9,1.5,1.6
the other 8:
1.9,1.85,1.7,4.0,1.8,2.4,1.6,1.6

nr = 18
ew = 9.398115
nw = 10

archie = ( nr * ( nw - ew )**2 ) / ( ew * ( nr - ew ))

archie :: 0.081, that indicates lots of luck. But i believe that the data is good, and chance played a very small part in the outcome. They were all very expected to win. So i have to question the calculation. I believe that the best estimate of the next outstanding prospect winning is 11/19.
Reply With Quote
  #2  
Old 11th July 2020, 12:02 PM
entropy entropy is offline
Member
 
Join Date: Feb 2019
Posts: 19
Default

Quote:
Originally Posted by chook
I saw archie mentioned in Gritz' post, keep it simple. and i was looking for a test which
would give me an idea of how credible my data can be.

here is some real data
18 race selections all were favorites:
1.9,1.7,2.65,1.85,3.0,1.8,1.7,4.0,1.8,2.4,2.2,1.8, 1.6,1.7,1.9,1.5,1.6,1.6
they popped up as most likely to win on 18 different race days. AKA specials or outstanding prospects.

10 winners:
1.7,2.65,3.0,1.8,2.2,1.8,1.7,1.9,1.5,1.6
the other 8:
1.9,1.85,1.7,4.0,1.8,2.4,1.6,1.6

nr = 18
ew = 9.398115
nw = 10

archie = ( nr * ( nw - ew )**2 ) / ( ew * ( nr - ew ))

archie :: 0.081, that indicates lots of luck. But i believe that the data is good, and chance played a very small part in the outcome. They were all very expected to win. So i have to question the calculation. I believe that the best estimate of the next outstanding prospect winning is 11/19.


Hi chook!

Your Archie Score workout is correct!

My $0.02 :

Your sample of 18 favs is very small, I have seen 18 favs in a row get beaten, so hard to draw any firm conclusions.
Off the top of my head I think you need at least a sample of 80.
Can you get more data?

Were the list of 18 selections cherry-picked/backfitted after the races were run or were they actual bets made before the race started and after the selection process occurred?
Reply With Quote
  #3  
Old 11th July 2020, 03:08 PM
chook chook is offline
Member
 
Join Date: Jan 2019
Posts: 21
Default

Thanks for your reply.
yes, 18 in a row can happen, but here we are dealing with a lot of odds on bets at Sale. it is not every odds on horse, just the odds on horse that comes up as being the very best winning chance on the entire day's racing across eastern Australia, and at that track.
I can not add to that data, it data comes from the real world. I understand 18 is a bit low. i will see if i can locate some rider data. i came across a couple of strange results there too, and the number of races was over 50.
in this instance, a $4. shot lost and a $3 one got the chockies so it feels goodish. My big quibble is centered around what it tells you if number of wins = expected.
i will do some more searching to see if i can come up with bigger sample.
Reply With Quote
  #4  
Old 11th July 2020, 03:58 PM
chook chook is offline
Member
 
Join Date: Jan 2019
Posts: 21
Default

more data. rider on horse with ability ($ per race). ability is ranked and permits ties.
l meech (A=1) nr = 92, nw = 42 ew = 37.013159249376 profit 10.1% archie:1.12
g ryan (A=2) nr = 94, nw = 39 ew = 35.835059957987 profit 19.8% archie:0.45
greg is riding country and provincial NSW linda VIC metrop and country.
oddson: she wins 15/22 he wins 11/20. explains the profit. but the expected and actual number of wins are close. his horses are at slightly longer odds.
but that statistic is wildly different.
Reply With Quote
  #5  
Old 12th July 2020, 08:21 AM
UselessBettor UselessBettor is offline
Member
 
Join Date: Sep 2011
Posts: 1,602
Default

chook,

A good explanation of the archie score is here: https://www.thestakingmachine.com/a...hance-or-fluke/

Anything that have less than 500 bets is likely to have archie scores that are very volatile.

The archie score is basically an analysis of how your selections have performed against the expected outcomes.

Essentially what it is doing is finding where your distribution of prices falls compared to the expected distribution of prices and hopefully your distribution is more on the profit side.

I have lots of systems which have 1000+ selections and have archie scores which show chance is a major factor. That is because they look for long shots and while they may be in profit the results are due to just a few long shots winning and bringing it into profit. You don't want to betting any system which relies on only a few winners to make that profit.

Personally if you are using archie I would be looking at those scores of 10 or higher only for betting with an automated system.
Reply With Quote
  #6  
Old 12th July 2020, 04:42 PM
Gritz Gritz is offline
Member
 
Join Date: Apr 2019
Posts: 49
Default

Hi Chook,


This may be of interest to you... A "Kelly Criteria" applied to Pre-post favs in the Herald Sun. One with all Favs and a second one with 1 filter applied.


The sheet PrePost Favs is in on OneDrive.. No macros. There is a bonus data table for more than 5 years.


Hope you like!
__________________
One Drive


“Tell me and I forget. Teach me and I remember. Involve me and I learn.”
“If the plan doesn’t work, change the plan, but never the goal.”
“Let us learn to appreciate there will be times when the trees will be bare, and look forward to the time when we may pick the fruit.”
Reply With Quote
  #7  
Old 14th July 2020, 05:20 AM
chook chook is offline
Member
 
Join Date: Jan 2019
Posts: 21
Angry

i do not have thousands or hundreds in my shortlist.Yesterday (13th July) there were 8 from that list and 4 which won.The most samples numbered 67, the least 11 and i have to work with that. The winners numbered 67,31,35,62 the losers 56,47,48,11. i doubt if a worhtwhile archie score can be calculated on the days card where 12 or 14 features are combined and sorted to give the shortlist. i only have data from 2018 in these tables, dating from when the good old Queenslaad TAB shuttered and we moved on to xml or json. life. i will stop looking at archie
Thanks
Reply With Quote
  #8  
Old 14th July 2020, 05:29 AM
chook chook is offline
Member
 
Join Date: Jan 2019
Posts: 21
Default

gritz: thanks. with my sample size there is be big variation. I use the tab prices after scratching time. Somewhere i also have nz tab data. I should post the code on gitlab. but the nz code is not as robust, it mostly works, and i live with the failures
Reply With Quote
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is On
Forum Jump



All times are GMT +10. The time now is 06:20 AM.


Powered by: vBulletin Version 3.0.3
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
©2008 OZmium Pty. Ltd. All rights reserved . ACN 091184655