Re: Suggesting a Poll

Angus Rodgers <twirlip@xxxxxxxxxxx> wrote:
Bill Dubuque <wgd@xxxxxxxxxxxxxxxxxxxx> wrote:

I doubt Google Groups star ratings are useful since many experts read
sci.math not through Google Groups but rather with real newsreaders.
But someone with some spare time might investigate how Google Groups
star ratings compare to the CRAP scores tabulated below.

Last year [1] I gathered some crude statistics that may be useful
in this regard. I noticed that there is a fairly strong inverse
correlation between the mathematical quality of one's sci.math
posts and the percentage of one's posts that are "crank bashing".
Below are the statistics, gathered crudely by simply computing the
percentage of posts in the Google Groups sci.math archive that
involve telltale crank-bashing keywords (crank, loon, crackpot...).
This CRAP (CRAnk-bashing Percentage) score is tabulated below.
A '*' denotes a known/suspected professional mathematician/PhD.
Notice almost all the professionals/PhDs have less than 0.4% CRAP,
whereas the high-CRAP offenders exceed such by an order of magnitude.
If nothing else, this might prove useful for constructing killfiles,
"scoring" posts in various newsreaders (e.g. Gnus), etc.

[1] http://google.com/groups?selm=y8zwsjtbxmf.fsf%40nestle.csail.mit.edu

6.5% 72/1100 Tonico, tonio, Tonicopm (Jose Antonio Perez)
4.4 98/2204 tommy1729, amy666
4.1 202/4910 uncle al
3.8 54/1420 galathaea

But galathaea is a crank un-basher, not a crank basher (or indeed a crank).

3.7 67/1790 Larry Hammick
3.2 214/6670 *Chris Hillman
3.0 126/4240 MoeBlee
2.9 178/6220 *Proginoskes (Christopher Heckman)
2.9 110/3790 Denis Feldmann
2.7 35/1300 Angus Rodgers

The ignominy! I must look into how you compute this, because I
almost never attack anyone for being a crank. The only crank
I recall really bashing was Mehran Basti, and I used much ruder
words than "crank" when I was bashing him! I also don't get
bashed for being a crank (although I did for a while when I
first came back to sci.math in 2007).

Does your program distinguish between /uses/ of the telltale
words and /quotations/ from posts containing them?! If not,
that's a pretty basic error.

As I said, it's based on a crude Google Groups search for telltale
words that appear in cranky threads. If you happen to participate
much in those threads then you will probably have posts that get
scored as containing those words (even if they're quoted in other
posts). This is probably as should be, since one shouldn't be
encouraging those off-topic posts by participating. _Generally_
I find that it provides a good crude indicator of mathematical
quality. Of course there will be a few exceptions since there
are some mathematicians who seem to enjoy crank bashing. But
this is by far the exception rather than the rule.

2.7 278/10300 Lester Zick
2.1 42/1960 Will Twentyman
2.1 41/1890 Aatu Koskensilta

Huh? Never was there less of a crank-basher (or crank).

2.0 99/4910 Mensanator
1.8 85/4710 Ross A. Finlayson
1.5 104/6920 Archimedes Plutonium, Ludwig Plutonium

I'm worse than Archimedes Plutonium? You cut me to the quick.

(Where's James Harris in the list, by the way?)

I didn't have the time to search for all his aliases.

1.5 13/866 Tim Smith
1.5 15/986 *Chan-Ho Suh
1.4 48/3390 Phil Carmody
1.4 12/877 *Edgar E. Escultura
1.3 46/3630 The Ghost In The Machine
1.2 19/1530 Michael Press
1.2 65/5280 *Pertti Lounesto
0.8 14/2600 *Robert Silverman, Bob Silverman, Pubkeybreaker
0.8 250/29500 Virgil
0.8 59/7790 *** T. Winter
0.8 24/3050 *Daryl McCullough
0.8 74/9580 Bill Taylor
0.8 22/2870 *Torkel Franzen
0.7 6/822 *Alexander Abian
0.7 23/3070 Dann Corbit
0.6 18/2880 *Dave L. Renfro
0.6 37/5880 quasi
0.6 12/2093 John Ramsden, Owlhoot
0.6 43/7160 *Robin Chapman
0.6 171/29500 *David C. Ullrich
0.5 31/6690 *John Baez
0.5 29/5980 *Gerry Myerson
0.5 16/3030 *Doug Norris
0.4 1/283 *hagen <knaf@xxxxxxxxxxx>
0.4 8/2010 *Keith Ramsay
0.4 11/2590 *Matthew P Wiener
0.3 51/16400 *Arturo Magidin
0.3 5/1570 *Derek Holt <mareg@xxxxxxxxxxxxxxxxxxxxxxxx>
0.3 7/2280 *Ilias Kastanas
0.3 26/9710 *Peter Nyikos
0.3 3/944 *David Eppstein
0.3 3/1080 *Tal Kubo
0.2 4/1760 *David Bernier
0.2 4/1890 *Bill Dubuque
0.2 9/4160 The World Wide Wade
0.2 9/3690 *Dave Rusin
0.2 11/7050 Dave Seaman
0.2 7/3450 *Lee Rudolph
0.2 6/3310 *Ronald Bruck
0.2 4/2120 *Timothy Murphy
0.2 2/1160 *Ed Hook
0.2 4/1900 *Toby Bartels
0.2 2/1230 *Timothy Y. Chow
0.2 3/1490 *Paul Sperry
0.1 3/5920 *G. A. Edgar
0.1 1/1097 *Mariano Suarez-Alvar
0.1 1/952 *Michael Barr
0.1 5/4800 *Zdislav V. Kovarik
0.1 1/1160 *John Conway
0.1 12/1880 *James Dolan
0.1 1/1210 *Noam Elkies
0.0 0/1180 *Axel Vogt
0.0 0/4860 David Cantrell
0.0 0/1230 *Jannick Asmus
0.0 0/2150 *Peter L. Montgomery
0.0 1/2210 *[Mr.] Lynn Kurtz
0.0 5/13500 *Herman Rubin
0.0 0/1900 Rob Johnson
0.0 1/3070 *William C Waterhouse
0.0 0/2210 *Greg Kuperberg
0.0 0/1090 *Miguel A. Lerma
0.0 0/2190 *D. J. Bernstein
0.0 0/1220 *Edwin Clark

OK, there's a pattern here, I'll grant you that: all good quality
at the bottom of the list, but what's happened at the top bothers
me - there are anomalies. (I know that that's special pleading,
but there are.)

That's the point. The pattern is obvious to any long-time reader
of sci.math. I was surprised that it correlates so strongly with
mathematical quality.

Perhaps, if you can spare the time, you might e-mail me with some
indication of how I came to figure so embarrassingly high in the
list of offenders.

Most likely because you participate a lot in such threads.

I hate witch hunts of any kind, especially when I am fingered as
one of the witches.

It's certainly no witch hunt. It's one of the variables that
I use to compute a score for posts in my newsreader (gnus).
.

Loading