PDA

View Full Version : Lies, lies and damn statisticsa



AussieMongrel
-1st June 2012, 08:56
Last night my poll on British Fencing had 12 respondents 70% of which said British fencing fell short in Ethics in its application of policy and procedure in regard to Olympic selection. Plenty of blue water there to help you make your opinion.

BF might prefer you to take into consideration the statistics as they were at 9.00 am which show 24 respondents and 63% believe that BF lacked ethics in this matter, but that is not relevant because we are only taking the figures from 12.30 pm last night as specified by me. Of*course I am biased because I have a very low opinion of the BF and it's decision making process, but don't let that get in the way of forming your opinion, just look at the statistics that I showed you from 12.30 am last night they are conclusive!

Statistics: it's all about who is applying them and the picture that they want to draw.

Question: Did any of the morons in the room making the decision question the statistics and ask for alternate data sets?

Vallangton
-1st June 2012, 09:02
I want to know who chose those particular statistics to be used.

Was it the board, the selection committee, the national coach, and was it the same across all the weapons?

Bezza
-1st June 2012, 09:10
I feel sorry for all the fencers involved. Can't be nice to have been stuffed by the selection or have some remarks about not deserving it. I hope they do well still. I'll still be cheering team GB, although my respect for BF may have hit a surprising new low.... which I thought would be impossible to get lower. Hopeful we can all get behind our fencers come July.

hokers
-1st June 2012, 09:12
The data presented was all the results from the start of Olympic qualifying to the close of Home Nation place selection. I was also asked to put together comparison of data between Williams, Hutchison and Nicoll in their first senior season (aged around 20 depending on when their birthday falls) with a view to looking at future potential.

There's your answer. SOMEONE asked Jon to prepare that stat at least, presumably the PD. I think picking and choosing your statistics from the data is going to be demonstrably the reason behind justification of selecting the #4 ranked fencer.

I'm preparing another response looking at some other statistics from the same data etc, but it will be a few more hours.

max
-1st June 2012, 09:32
Statistics: it's all about who is applying them and the picture that they want to draw.



Was this written with any sense of irony given that you want to believe your own statistics at a particular point in time and then imply that BF would prefer a later set? Just asking, you sam. :rolleyes:

AussieMongrel
-1st June 2012, 09:36
Yes

Hansei
-1st June 2012, 10:55
I want to know who chose those particular statistics to be used.

Was it the board, the selection committee, the national coach, and was it the same across all the weapons?

I would add the additional question of whether this was communicated to the athletes involved. Were they aware of how their performances were going to be measured and compared?

hokers
-1st June 2012, 11:22
The fencers will be sent a copy of the results and rankings of all the fencers at their weapon, which will be used by the selectors to decide who should be nominated. The names of the Olympic selectors will also be included. The candidates will have [5] working days from receipt of this information to question the accuracy of any of the data, or challenge the impartiality of any of the selectors. Any challenge to the impartiality of the selectors will be referred to the Appeal body (which shall be independent of the selection panel and free from bias) for an immediate ruling. They may also supply the selectors with any other information that they feel may be relevant to their selection, e.g. periods of injury
during the past 12 months etc. The selectors will not be obliged to consider any further information from the fencers after this 5 day period.


Results and rankings only, no mention of any statistical analysis being required to be revealed etc.
Would have to see the minutes of the meetings to know what they were, but not published by BF.

hokers
-1st June 2012, 14:06
So I'm not finished by some distance, as I'm struggling to get older poule results data from some comps, but I have a few selected stats that might be of interest.

Poule victories in last 6 comps before 9th May
Jojo 30 (D6)
Louise 16 (D9) (Missed Bologna due to injury)
Chrystall 16 (D17) (3 short poules)
Sophie 13 (D21) (2 short poules)

For comparison, due to short poules.
Jojo 30/36 = 83%
Louise = 16/25 = 64%
Chrystall 16/33 = 48%
Sophie 13/34 = 38%

Data
Bologne Moscou Antalya Londres Orléans Catania
04.05.12 16.03.12 09.03.12 24.02.12 10.02.12 10.10.11
Fencer
Louise 0-0 3-3 3-2 3-2 5-1 5-1
Jojo 6-0 4-2 6-0 4-2 4-2 6-0
Chrystall 3-3 3-3 2-3 4-1 1-5 3-2
Sophie 2-3 3-3 1-4 4-2 0-6 3-3



Scoring GP/A-Grade/Championship comps (made L64 or better) in the full 1/4/2010-9/5/2012 selection period (excludes Tianjin 2012)
Jojo: 20/23 = 87%
Louise: 15/19 = 79%
Chrystall: 18/22 = 82%
Sophie 8*/17 = 47%

*(discounted newcastle Satellite for comparison)
Source: http://fie.ch/Competitions/Ranking.aspx

Counter-arguments for some of these. Poules are a good measure of how you fight against international fencers, but they are only to 5 and get easier if your ranking is higher.
Some of Sophie's comps here were while she was still a Junior (not that I think that really matters, but that is the case nonetheless)

The overall point I'm trying to make here is that you can present a wide range of stats that can give very different impressions based on the same data. If stats were to be used in the selection process, the choice of stats should have been made available to the eligible fencers in advance to check/rebut.

coach carson
-1st June 2012, 14:09
Why do these lies look better than the other ones?

hokers
-1st June 2012, 14:18
Why do these lies look better than the other ones?

Cause that's the point I'm making, see the edit, it's all about the presentation.

"JOJO 1.85 TIMES AS LIKELY TO SCORE POINTS AS SOPHIE! EXTRA EXTRA!"
"JOJO TWICE AS LIKELY TO WIN A POULE FIGHT! READ ALL ABOUT IT!"

vs
"ALL 4 FENCERS ELIGIBLE FOR NOMINATION AND WITHIN 11 RANKING PLACES OF EACH OTHER! Nothing to see here..."

Gavinoo
-1st June 2012, 14:44
In the Olympics fencers go straight into a DE, so looking at poule stats wouldn't be ideal. Looking at ranking would as you get drawn depending on your world ranking, I think.:eek:

hokers
-1st June 2012, 14:52
In the Olympics fencers go straight into a DE, so looking at poule stats wouldn't be ideal. Looking at ranking would as you get drawn depending on your world ranking, I think.:eek:

I know, I acknowledged there were counter arguments. It is still competitive international fencing though. I might say that given how few DE bouts are actually fought in the whole 2 year period, it's a reasonable indicator of how adaptable they are to different styles, opponents etc as well.

coach carson
-2nd June 2012, 06:11
The simple issue is how we managed to design or use a process with the result that we take our weakest contender over our strongest contender.

Purple Fencer
-23rd June 2012, 05:06
As we say in the Colonies....I have no dog in this fight, but I thought I'd point out that the apparent Mark Twain quote used for the thread title should actually read "Lies, damned lies, and statistics," although the FUL quote of "There are three kinds of lies: lies, damned lies, and statistics." is sometimes attributed to Benjamin Disraeli.

Thank you....carry on....