Acro Image

Aerobatics Server

ACRO E-mail Archive Thread: TBLP, with emphasis on the P

[International Aerobatic Club] [Communications] [Aerobatics Images]

Disclaimer: These aerobatics pages are developed by individual IAC members and do not represent official IAC policy or opinion.

[Usage Statistics]


ACRO E-mail Archive Thread: TBLP, with emphasis on the P



                


Thread: TBLP, with emphasis on the P

Message: Re: TBLP, with emphasis on the P

Follow-Up To: ACRO Email list (for List Members only)

From: ACCassidy at aol.com

Date: Wed, 22 Jan 1997 05:17:42 UTC


Message:

  Judging by recent posts, I feel there is a need briefly to clarify the
currently CIVA-approved system as of last November's meeting.  I was there,
had long discussions with Herr Bauer and Senor Penteado who have combined
forces for the latest improvements.  And I understand the mathematics.

As Rudy is not posting here perhaps I can help out a little.  For the ladies
reading, please excuse my giving judges a masculine pronoun, but it does
simplify my slow typing a bit.

Firstly, we must now refer to TBLP.  The P is officially for "Progressive"
but I prefer to think of it as Penteado as Rudy has worked so hard to improve
the system and the other letters all stand for names.

The progressive label means that judges scores are no longer "just in" or
"just out" of the window (acceptable range for confidence against bias either
way).  Instead if a judge is somewhere close to the bias confidence margin
his score is still included but given less "weight" as his impartiality is
less sure.  The "weight" given is proportional to our confidence in his
impartiality.  His score will only be completely discarded if he is more than
just a bit out.

This new refinement is what will eliminate some of the oddball effects that,
while mathematically valid under the TBL system, gave rise to some anomalies
both at WAC '96 and at previous contest alluded to by Gypsy444.

Be advised with certainty, however, that the move to performing this analysis
on each figure has not yet been taken in the International forum that is
CIVA.  I know Rudy is working on it and it may come next year or after that.
 But as of 1997, TBLP only looks at each judge's total score for each pilot.

I think my summary of the benefits of TBLP, not obtainable by the
highest/lowest system, are as follows:

Firstly, it gives each judge the opportunity of an equal say in the overall
results.  His opinion will only be given less credence if he appears
significantly biased as compared to the general consensus.  This potential
equality should be of vital importance in encouraging volunteers to step
forward - their view of things will always be given a fair hearing.

If despite his best efforts to judge fairly, he does show bias, albeit
sub-consciously, then to be fair to the pilots it is necessary to reduce the
effect of his undue preferences.  TBLP does this in the smoothest way so far
devised and will let him down gently rather that booting him out sans
ceremony.

I know it is easy to muddy the waters with statistics, but I hope I have not
fallen into that trap.  If anything I have said is unclear I apologise - but
I speak English not American.

Lastly, I agree with Damon Wack that there were some uncannily accurate, top
of the range, scores at WAC, particularly from the French and Russian judges
in Programme 1.  If these we cunningly contstructed, as it seemed at the
time, I have amazing respect for their prowess ( a bit like watching David
Copperfield levitate a steam train).  But rest assured that if we change to a
less statistically sound system, the likelihood is that such super-judges
would find even more effective ways to exercise their Cray-like brains.

Fly safely, judge fairly, always do your best

Alan Cassidy
British Alternate CIVA Delegate
(Un)limited Pilot
Part-time mathematician


                


© Dr. Günther Eichhorn
Retired
Email Guenther Eichhorn