Bill James Postseason Predictions revisited...
Bob-Nob
Bill James's prediction system had gone since 1983. (To that point,
James's system had gone 82-31 in predicting winners). Ron Johnson
replied that he had checked at one point, but that he didn't have his
work handy.
I decided to do it myself...
World Series: 10-9
Right Wrong
1984, 1986, 1989 1985, 1987, 1988
1991, 1995, 1997, 1998, 1999 1990, 1992, 1993, 1996
2000, 2001 2002, 2003
ALCS: 10-9
Right Wrong
1984, 1988, 1989 1985, 1986, 1987
1990, 1992, 1995, 1996, 1998 1991, 1993, 1997, 1999
2002, 2003 2000, 2001
NLCS: 10-9
Right Wrong
1984, 1989 1985, 1986, 1987, 1988
1990, 1992, 1995, 1996, 1999 1991, 1993, 1997, 1998
2000, 2001, 2002 2003
ALDS: 9-8-1
Right Wrong
1995: BOTH
1996: Yankees vs. Rangers Orioles vs. Indians
1997: Orioles vs. Mariners Indians vs. Yankees
1998: Yankees vs. Rangers Indians vs. Red Sox
1999: Yankees vs. Rangers Red Sox vs. Indians
2000: Mariners vs. White Sox Yankees vs. A's
2001: Mariners vs. Indians Yankees vs. A's
2002: Twins vs. A's, (*)
2003: Yankees vs. Twins Red Sox vs. A's
NLDS: 9-8-1
Right Wrong
1995: BOTH
1996: Cardinals vs. Padres Braves vs. Dodgers
1997: Braves vs. Astros Marlins vs. Giants
1998: Braves vs. Cubs, (*)
1999: Braves vs. Astros Mets vs. Diamondbacks
2000: Cardinals vs. Braves Mets vs. Giants
2001: BOTH
2002: BOTH
2003: BOTH
(*) indicates the system did not predict a winner -- the two teams had
an equal number of points by the system.
I find it a disappointing result, albeit a predictable one, given
how Bill James came up with the system. One could do better simply
by picking the team with the better record or picking the team with the
better Pythagorean Projection (as I will show in another post). Still,
at least it's now recorded for google-posterity.
Catch you later.
--Robert Machemer
--
Robert Paul Aubrey Machemer | "For each time he falls, he shall
Amherst College, Math & Classics | rise again, and woe to the wicked!"
IF1, IF3, IF9: best films, cast | --Don Quixote (Man of La Mancha)
(What are YOU doing this weekend? See IF12 on May 23rd, 2004)
Cody M. Stumpo
: Three weeks ago or so, I asked if anyone had checked how well
: Bill James's prediction system had gone since 1983. (To that point,
: James's system had gone 82-31 in predicting winners). Ron Johnson
: replied that he had checked at one point, but that he didn't have his
: work handy.
: I decided to do it myself...
: World Series: 10-9
: ALCS: 10-9
: NLCS: 10-9
: ALDS: 9-8-1
: NLDS: 9-8-1
: I find it a disappointing result, albeit a predictable one, given
: how Bill James came up with the system. One could do better simply
: by picking the team with the better record or picking the team with the
: better Pythagorean Projection (as I will show in another post). Still,
: at least it's now recorded for google-posterity.
Not to mention you would do equally well flipping a coin. In fact, you'd
have a very large chance of doing better. I don't know what his system
is, but he may have overfit. (you would sell less books flipping a coin
though)
---------------------------*************************------------------------
co...@soda.CSUA.berkeley.edu http://www.csua.berkeley.edu/~cody/
Holding a titular Archbishopric since 1999.
David J. Grabiner
> Bob-Nob <rpmac...@nospamte.amherst.edu> wrote:
> : Three weeks ago or so, I asked if anyone had checked how well
> : Bill James's prediction system had gone since 1983. (To that point,
> : James's system had gone 82-31 in predicting winners). Ron Johnson
> : replied that he had checked at one point, but that he didn't have his
> : work handy.
>
> : I decided to do it myself...
>
> : World Series: 10-9
> : ALCS: 10-9
> : NLCS: 10-9
> : ALDS: 9-8-1
> : NLDS: 9-8-1
>
> : I find it a disappointing result, albeit a predictable one, given
> : how Bill James came up with the system. One could do better simply
> : by picking the team with the better record or picking the team with the
> : better Pythagorean Projection (as I will show in another post). Still,
> : at least it's now recorded for google-posterity.
>
> Not to mention you would do equally well flipping a coin. In fact, you'd
> have a very large chance of doing better. I don't know what his system
> is, but he may have overfit. (you would sell less books flipping a coin
> though)
James's system was quite simple. Look at each team statistic, and if
the teams which led in that statistic won the series X times and lost Y
times then award X-Y points to the team which does better in that
statistic. For example, since the teams which threw more shutouts were
45-26 when the system was developed, the team with more shutouts gets
+19. Since the teams which allowed fewer walks were 37-44, the team
with fewer walks allowed gets -7.
Thus it isn't fair to say that the system was 82-31 before 1983; the
system was developed based on data through 1973, including the
historical record of which teams won.
--
David Grabiner, grab...@alumni.princeton.edu, http://remarque.org/~grabiner
Baseball labor negotiations FAQ: http://remarque.org/~grabiner/laborfaq.html
Shop at the Mobius Strip Mall: Always on the same side of the street!
Klein Glassworks, Torus Coffee and Donuts, Projective Airlines, etc.
Bob-Nob
>> Bob-Nob <rpmac...@nospamte.amherst.edu> wrote:
>> : Three weeks ago or so, I asked if anyone had checked how well
>> : Bill James's prediction system had gone since 1983. (To that point,
>> : James's system had gone 82-31 in predicting winners).
<snip>
>> : I decided to do it myself...
>> : World Series: 10-9
>> : ALCS: 10-9
>> : NLCS: 10-9
>> : ALDS: 9-8-1
>> : NLDS: 9-8-1
<snip>
> it isn't fair to say that the system was 82-31 before 1983; the
> system was developed based on data through 1973, including the
> historical record of which teams won.
I understand that it was developed based on data through 1973.
Regardless, I asked if anyone had studied how it had done since
1983 (which was the last time, to my knowledge, that James had
published the results of the system). I didn't find anything, so
I decided to study it and post the results. I don't know how that
isn't fair. It is what it is -- the system was 82-31, by James's
account, through 1983, and is 48-43-2 since.
Personally, I had hoped the system still worked, despite my
reservations about the way it was derived. And maybe it still does
(albeit, while lately going through a rough run of luck), but I'm not
encouraged by the results.
Catch you later.
--Robert Machemer
PS: In case anyone is curious, here are the results of the system
from 1974-1983:
ALDS: 2-0
NLDS: 2-0
ALCS: 7-3
NLCS: 6-5
WS: 6-5
Terrell Miller
news:4161...@amhnt2.amherst.edu...
> Three weeks ago or so, I asked if anyone had checked how well
> Bill James's prediction system had gone since 1983. (To that point,
> James's system had gone 82-31 in predicting winners). Ron Johnson
> replied that he had checked at one point, but that he didn't have his
> work handy.
<SNIP>
> I find it a disappointing result, albeit a predictable one, given
> how Bill James came up with the system. One could do better simply
> by picking the team with the better record or picking the team with the
> better Pythagorean Projection (as I will show in another post). Still,
> at least it's now recorded for google-posterity.
here's a dirty little secret: the touts only rarely do a better job at
predicting future results than the average fan does. That's why I stopped
reading BP except for the occasional time-killer.
--
Terrell Miller
mill...@bellsouth.net
" A strong conviction that something must be done is the parent of many
bad measures."
-- Daniel Webster
Hank Gillette
"Terrell Miller" <mill...@bellsouth.net> wrote:
> here's a dirty little secret: the touts only rarely do a better job at
> predicting future results than the average fan does. That's why I stopped
> reading BP except for the occasional time-killer.
"It's tough to make predictions, especially about the future."
(attributed to almost everyone, including Casey Stengel, Yogi Berra,
Will Rogers, Winston Churchill, and Mark Twain).
--
Hank Gillette
"The real world is President Bush's Achilles' heel. He can't keep his distance
from it forever." -- Bob Herbert