SAD Eights 2013

PostSAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At11:01, 30th Apr 2013
238 weeks ago
SAD web-pages for Eights now up and awaiting your scores ...

www.atm.ox.ac.uk/rowing/sad.html
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At16:44, 6th May 2013
237 weeks ago
1st iteration now complete.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At10:11, 13th May 2013
236 weeks ago
2nd iteration now complete.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At09:56, 16th May 2013
235 weeks ago
I could really use more scores for women's crews.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At12:35, 18th May 2013
235 weeks ago
Iteration#3 now complete.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At12:06, 22nd May 2013
235 weeks ago
4th and final pre-race iteration now complete.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At09:51, 23rd May 2013
234 weeks ago
SAD scores now modified by SAD BaStaRD to incorporate results from the first day's racing.

Note that the tables now show the Thursday starting positions and predicted gains are for the remaining three days of racing.

Currently the most accurate contributor is 'AB' with an average error of 0.193 (ie mean absolute difference between their scores and the current SAD scores for the same crews). And the only person doing worse than Fantasy Bumps is 'AD', whoever that is.

Early days ...
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At10:01, 24th May 2013
234 weeks ago
6th iteration now complete.

As racing progresses, the original contributor scores have decreasingly less weight and the results more weight. On the other hand the contributor weights are adjusted according to how accurate they seem to be. So it's not always the case that all scores will converge towards actual performance after each iteration.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At11:36, 25th May 2013
234 weeks ago
Iteration#7 now complete, with predictions for the finishing order given today's starting positions.
 
PostRE: SAD Eights 2013
AuthorAnu Dudhia
dudhia@atm
Posted At09:49, 6th Jun 2013
232 weeks ago
Final SAD scores now on www.atm.ox.ac.uk/rowing/sad.html

These are the best estimates the SAD BaStaRD program could come up with for each crew's speed based on the original contributor estimates and what actually happened.

Mostly it does a pretty good job of predicting the actual outcome (shown by the yellow highlighted crews mostly occupying the most probable positions). Difficulties occur when crews 'bumped back', eg Exeter - St Anne's in mid Men's Div 2, Keble & Merton in W.Div 1, Wolfson & Somerville top of W.Div 2, Mansfield and Linacre bottom of W.Div 2. In which case you can interpret these predictions as what 'should' have happened given just the average amount of randomness rather than exceptional randomness.

Of the contributors, 'BK' proved the most accurate, followed closely by 'RC' and 'AS'. These people, or people claiming to be these people, will no doubt be hiring out their services as market analysts to those planning on trading in the next Fantasy Bumps. I, obviously, won't be.