STATISTICAL ARGUMENT, MR. PEASE
Jason Burke
Now, the original post to this thread was discussing the STATS Zone
Rating, I am not familiar enough with the DA to comment on that, but in
the post it said that it was an equivalent measurement. Well, if you
look at the Zone rating for the different positions, you come up with
some interesting findings that I would dispute with anybody.
At Second Base.
I see Roberto Alomar regularly make plays on the Shortstop side of
second all the way over to behind the firstbaseman, but if you look at
the Zone Rating, our own Felix Fermin gets to an amazing 91% of the balls
in his zone, while Alomar gets to something like 85%, only problem for
Roberto is that the average for a second baseman is 86%, so not only is
does Fermin have more range than Alomar, but Alomar does not even have
average range. Now Fermin only played like 30 games at second base, so
after I saw this stat, I looked at other postitions.
At Shortstop.
Who would you guys say has the best range at shortstop? Vizquel, Barry
Larkin, Ozzie Smith, Shawn Dunston, or Felix Fermin? Why are we bringing
up Alex, Felix Fermin has a Zone rating of like 92.8%. Barry Larkin
hangs in there with 90.5%, but our man Omar is below average at 88%.(The
average for shortstops is 89%). I am sorry, but I have seen all of these
guys a bunch and Felix Fermin doesn't get to half of the balls that guys
like Vizquel or Larkin get to. Larkin and Vizquel regularly make sweet
stops behind second base, when playing normal depth, while Fermin cannot
even make those plays when he is at double play depth.
A quick look at third base shows that we have been bitching about Edgar's
fielding a little too much. Edgar has a Zone Range of 82% which is a
little below average(the average was something like 83 or 84), but he
dominates Mike Blowers, who has a pathetic Zone Rating of 76% which was
the lowest rating that I found in the Majors. Blowers may be bad, he may
even have the worst range of any everyday third baseman, but can anyone
sit there and honestly tell me that Blowers covers more ground at third
than Edgar?
In rightfield.
Has anyone watched the Yankees new acquisition Ruben Sierra misplay
flyballs all over his new homepark this last week, or seen his misplay a
ton of balls for my former hometown Oakland Athletics. For anyone not
familiar with Sierra, he has excellent speed, but if there is a guy that
gets a worse read off the bat I can't think of him. The only reason I
mention Ruben is because he is being ripped on by guys like Peter Gammons
for how bad of a fielder he is. The only problem is that Ruben,
according to the STAT Zone Rating is one of the better rightfielders in
the game. He dominates his teammate Paul O'Neill, who is rated as an
above average to excellent outfielder in most scouting reports, he
domiates Tim Salmon, and he destroys Jay Buhner. Now, I would definitely
say that all of these guys (O'Neill, Salmon, and even Jay) have a little
better range than Sierra, but for those of you that watch the Mariners
and few other teams, the Zone Range that I just couldn't believe, is that
Ken Griffey Jr. should not be playing centerfield. It should be Jay
Buhner! For all of you people that bitch about Buhner's range, his 77%
is better than Griffey's 76%. That's been our problem, we should have
Jay in center and Junior in right!
But, the rating that just made me laugh at this statistic was the
leftfielders. Barry Bonds(before this season) was considered one of the
best fielders in the game. His Zone Rating was like 78%. So, going on
the premise that I agree with this statistic, Bonds has range that is
quite a bit below average. I couldn't think of many leftfielders to
compare him to because I was reading the book at like three in the
morning, but as I took one last look at the book, I came across the
biggest non-firstbaseman fat ass in the game, your friend and mine Kevin
Mitchell. Now according to the Zone Rating, Kevin Mitchell had a rating
of 85%. Now I have watched Mitchell play left a bunch of times, and I
have watched Bonds play left a bunch of times. Kevin Mitchell is lucky
if he can field any ball within a four foot circle around where he is
positioned, yet according to this he has better range than Barry Bonds?
Now, I won't at this time argue about whether or not Bonds has sweet
range, or whether or not Griffey has sweet range, I will just say that
any system that has Fermin in front of Vizquel and Barry Larkin, has
Edgar in front of Blowers, has Bret Boone in front of Roberto Alomar, has
Jay Buhner in front of Griffey, and has Kevin Mitchell in front of Barry
Bonds is a stat that I don't find very accurate. If someone thinks that
I have a bias toward Griffey, they are probably correct. I enjoy
watching him play and since I have watched the Mariners more than any
other team the past couple of years and since they have been my favorite
team since their inception, I can understand someone believing that I
have a bias toward Griffey. But looking at these stats just shows me
that this rating system has got to be flawed somehow. I played baseball
from when I was 6 untill I was 18, and I probably have watched 80-100
games a year since I was 9 years old, and I think that I have a pretty
good idea of who is a good fielder and who is not. But even if I might
not know who has more range at third base of maybe in rightfield, I have
to question that I cannot judge between two players that I have seen play
50-100 times in person. With my own eyes I would say that Vizquel has
twice as much range as Fermin, and I think that if you ask any Mariner or
Indian season ticket holder, you will get the same response. With my own
eyes I would say that Barry Bonds gets to twice as many balls as Kevin
Mitchell, and I think that if you ask anyone who has watched those two on
a regular basis, they would agree with me. This stat has to be done
based on who the player is. I don't care what anyone says, but Felix
Fermin does not get to more balls than Roberto Alomar. Felix might
however get to a higher percentage of balls in his smaller range than
Alomar gets to in his larger range, but to believe this stat I ask only
two things. That Roberto Alomar has a better range at 2B than Fermin,
and that Barry Bonds has a better range than Kevin Mitchell. This
statistic does not meet these criteria I am afraid.
Jas
Will Shortt
HALLELUJAH!! Thank you Mr. Burke. Finally, the definitive post to rebuke
the stats-over-reality ravings of Mr. Pease.
Statistics are only the result of formulae which are meant to approximate
reality. The more complex the formula, the more likely intervening
variables will jam the works skewing the results. The stats that make
baseball great - Ave, RBI's, ERA's, W-L, etc. - are all relatively
simple, but still can be misleading. Example: Blowers and Buhner's nearly
identical stats - yet having very different years.
The point, however, is that these simple stats help make baseball great.
I've just moved to Eugene, OR and have had the pleasure to see the Class
A Emeralds here twice now. Of course, I knew none of the players, but
armed with all the stats on the program, I was able within ten minutes to
get a feel for the players and thus enjoyed the game much more. (BTW,
ex-husky CF Joe Trippy is hitting .305 and leads the Em's in RBI's).
However, the elaborate formula Mr. Pease presents serves only to
obfuscate rather than illuminate. We all had gut feelings and anecdotal
evidence that something was fishy with Griffey's readout. Certainly
everyone in baseball couldn't have been wrong about the Kid. Certainly
his record streak a couple years back of opportunities without an error
couldn't have been done by a sub-par CF. Could we have only been seeing
his good games and missing the reality of Griffey's slow legs and limited
range?
Confusion, emotional reaction ensued.
And then Jason stepped to the plate.
WOW! Whattapost!! Knocked that formula deep, deep, deep into center.
Pease going backbackback . . . goodbye statistic!!! (Much rejoicing)
The Fermin over Vizquel thing was the clincher. Hilarious. And Kevin
Mitchell?? get outta here! Thank you Mr. Burke for taking time to find
the holes and systematically present them.
And Mr. Pease, your initial post sparked the most spirited debate I've
seen in this group. For that, I thank you. But remember, while we all love
stats, they are NOT the numerical equivalent to reality. At best they are
an approximation which helps us gauge reality. They are not a reality
substitute, and when they fly in the face of the observations of millions
of serious fans of the game, they must be suspect.
But if you really love stats, you gotta love the kingdome, right? no
intervening weather, light , or field condition variables - just the
perfect little petri dish for statistical observation. What's the fuss
all about anyway?
Will "got the faith and it's burnin' a hole in my pocket" Shortt
GO M's!
Michael Cox
<Pine.A32.3.91j.95081...@homer09.u.washington.edu>, Will
Shortt <sho...@u.washington.edu> wrote:
>
> HALLELUJAH!! Thank you Mr. Burke. Finally, the definitive post to rebuke
> the stats-over-reality ravings of Mr. Pease.
And (although continued argument is certain to ensue regardless) it's
about time. Maybe Mr. Cameron will stop (and I say that as someone who's
on *his* side.)
> However, the elaborate formula Mr. Pease presents serves only to
> obfuscate rather than illuminate. We all had gut feelings and anecdotal
> evidence that something was fishy with Griffey's readout. Certainly
> everyone in baseball couldn't have been wrong about the Kid. Certainly
> his record streak a couple years back of opportunities without an error
> couldn't have been done by a sub-par CF. Could we have only been seeing
> his good games and missing the reality of Griffey's slow legs and limited
> range?
Well, wait. There WAS that one game when he didn't run for a ball and a
few people booed...
> WOW! Whattapost!! Knocked that formula deep, deep, deep into center.
> Pease going backbackback . . . goodbye statistic!!! (Much rejoicing)
And then they ate Mr. Sieferman's minstrels...
> The Fermin over Vizquel thing was the clincher. Hilarious. And Kevin
> Mitchell?? get outta here! Thank you Mr. Burke for taking time to find
> the holes and systematically present them.
I dunno, maybe Mitchell's lard-butt act was all a clever ruse. He only
*looked* slow. What about all those balls he bare-handed? Whaddya mean
there was only one? I'm sure I've seen it a thousand times!...all
replays?? D'ohh!
MC
---------------------------------------------------------------------
"I came to kick @ss and chew | The opinions expressed herein,
bubblegum, and I'm all out of | blah, blah, blah, you know
bubblegum!!" | the drill...
-"Rowdy" Roddy Piper | mich...@accessone.com
Richard Cameron
has no range is put to death
In <Pine.A32.3.91j.95081...@homer23.u.washington.edu>
(Deletia of lots more GREAT Points)
Scott Fischthal
>
>Now, the original post to this thread was discussing the STATS Zone
>Rating, I am not familiar enough with the DA to comment on that, but in
>the post it said that it was an equivalent measurement. Well, if you
>look at the Zone rating for the different positions, you come up with
>some interesting findings that I would dispute with anybody.
>
>At Second Base.
>
>I see Roberto Alomar regularly make plays on the Shortstop side of
>second all the way over to behind the firstbaseman, but if you look at
>the Zone Rating, our own Felix Fermin gets to an amazing 91% of the balls
As with any other stat, you have to look at these numbers over the course of
more than one year, especially when that one year is '94, which was a short
season. At some of the positions, the differences in DA were not statistically
significant. Anyway, I haven't studied ZRs very carefully, but look at the DAs:
Felix Fermin's DAs:
Yr Fermin League Rank
94 .573 .639 11/11
93 .571 .641 14/14
92 N/A .656
91 .616 .661 13/14
90 .653 .653 10/14
89 .684 .665 7/16
Putrid, basically.
>in his zone, while Alomar gets to something like 85%, only problem for
>Roberto is that the average for a second baseman is 86%, so not only is
>does Fermin have more range than Alomar, but Alomar does not even have
>average range. Now Fermin only played like 30 games at second base, so
At this point in his career, Alomar does not even have average range. His range
disappeared after his winter ball injury a couple of years ago.
>after I saw this stat, I looked at other postitions.
>
>At Shortstop.
>
>Who would you guys say has the best range at shortstop? Vizquel, Barry
>Larkin, Ozzie Smith, Shawn Dunston, or Felix Fermin? Why are we bringing
>up Alex, Felix Fermin has a Zone rating of like 92.8%. Barry Larkin
Does Fermin actually show up that high in ZR? If so, there may be some serious
problems with ZR. In particular, we know that ZR tends to rate surehandedness
above range compared to DAs, which are devastating to players with little range
but reward high-range players very well.
>hangs in there with 90.5%, but our man Omar is below average at 88%.(The
>average for shortstops is 89%). I am sorry, but I have seen all of these
>guys a bunch and Felix Fermin doesn't get to half of the balls that guys
>like Vizquel or Larkin get to. Larkin and Vizquel regularly make sweet
>stops behind second base, when playing normal depth, while Fermin cannot
DAs rate Valentin, Guillen, Ripken, Blauser (who everyone agrees has improved
immensely with the glove), Clayton and Larkin at the top. Vizquel didn't make
the list because he didn't play enough games at SS last year (70 was the cutoff,
I believe), but is usually above average. Fermin, as shown above, is the worst
defensive SS in the league, according to DAs.
>even make those plays when he is at double play depth.
>
>In rightfield.
>
>Has anyone watched the Yankees new acquisition Ruben Sierra misplay
>flyballs all over his new homepark this last week, or seen his misplay a
>ton of balls for my former hometown Oakland Athletics. For anyone not
>familiar with Sierra, he has excellent speed, but if there is a guy that
>gets a worse read off the bat I can't think of him. The only reason I
>mention Ruben is because he is being ripped on by guys like Peter Gammons
>for how bad of a fielder he is. The only problem is that Ruben,
>according to the STAT Zone Rating is one of the better rightfielders in
>the game. He dominates his teammate Paul O'Neill, who is rated as an
It looks like it's time to run that DA/ZR correlation study. In '94, Sierra had
just about the worst DA for an OF since 1988 (.480). League average was .560;
Sierra is consistently horrible. Meanwhile, O'Neill is one of the best at .585.
Buhner, the statue with a strong arm, languished near the cellar at .502.
>above average to excellent outfielder in most scouting reports, he
>domiates Tim Salmon, and he destroys Jay Buhner. Now, I would definitely
>say that all of these guys (O'Neill, Salmon, and even Jay) have a little
>better range than Sierra, but for those of you that watch the Mariners
>and few other teams, the Zone Range that I just couldn't believe, is that
>Ken Griffey Jr. should not be playing centerfield. It should be Jay
>Buhner! For all of you people that bitch about Buhner's range, his 77%
>is better than Griffey's 76%. That's been our problem, we should have
>Jay in center and Junior in right!
You can't compare numbers between fielders at different positions; the amount of
ground a centerfielder is assigned is much larger than the amount of ground a
rightfielder is assigned, for example. Anyway, Griffey's DA last year was
higher than Buhner's (.516 to .502).
>
>But, the rating that just made me laugh at this statistic was the
>leftfielders. Barry Bonds(before this season) was considered one of the
>best fielders in the game. His Zone Rating was like 78%. So, going on
>the premise that I agree with this statistic, Bonds has range that is
>quite a bit below average. I couldn't think of many leftfielders to
Barry Bonds' DAs:
Yr Bonds League Rank
94 .551 .563 7/8 (Difference between #7 and #2 not stat. sig.)
93 .584 .566 5/15
92 .598 .582 4/8 (Pittsburgh team: 3/12)
91 .600 .580 4/11
90 .623 .567 1/9 (by a huge margin)
89 .621 .574 2/9
88 .590 .546 1/11
Note also that Bonds, year in, year out, allows fewer extra base hits than
anyone else in the league, a valuable asset (check on the baseball archives at
www.baseball.org for more data). Last year was an aberration, and even then the
difference between him and #1 was smaller than the difference between him and #2
the years he led the league in DA.
>any system that has Fermin in front of Vizquel and Barry Larkin, has
DA doesn't; it rates Fermin as incredibly horrible.
>Edgar in front of Blowers, has Bret Boone in front of Roberto Alomar, has
No data on Martinez/Blowers for the same year. Boone rated behind Alomar the
year they were both in the AL (.658-.646).
>Jay Buhner in front of Griffey, and has Kevin Mitchell in front of Barry
Again, you can't compare ZRs for one position with those for another.
>Bonds is a stat that I don't find very accurate. If someone thinks that
Bonds rated ahead of Mitchell every year except last, and last year they wre
about even -- except many more of the hits in Mitchell's zone were extra base
hits. Of course, last year was weird in other ways too. Didn't Felix "Noodle
Bat" Fermin hit .300 last year? Does that tell us that batting averages are
worthless, or that it was just one of those things?
But yes, your eyes can deceive you about range. You lack a reference point to
tell how fast a player is really moving, so it's awfully tough to tell that
Griffey is really too slow to play CF. But that's a different argument. I
can't explain why ZRs gave such bizarre results, not having studied them
closely. I know there are some flaws in DAs (e.g. the 3B zone is too small),
but as you can see from the above, they seem to work much better than ZRs from
an "eyeball" perspective.
There are very few players whose DA's are at odds with their reputations
(Griffey and Alomar being two, both of whom people are studying very carefully
to determine why). Ozzie Smith had the best DA in the NL from 1988-92 (they
started being run in '88), with his DA dropping each year until he fell into
second in '93 and down to average in '94 -- perfectly consistent with what one
would expect from a now-40 year old SS who in his prime was the best ever.
Scott Fischthal
Artificial Intelligence Technology Center
Loral Federal Systems Group
Gaithersburg, MD
Eric Sieferman
Michael Cox <mich...@accessone.com> wrote:
>In article
><Pine.A32.3.91j.95081...@homer09.u.washington.edu>, Will
>Shortt <sho...@u.washington.edu> wrote:
>
>>
>> HALLELUJAH!! Thank you Mr. Burke. Finally, the definitive post to rebuke
>> the stats-over-reality ravings of Mr. Pease.
>
>> WOW! Whattapost!! Knocked that formula deep, deep, deep into center.
>> Pease going backbackback . . . goodbye statistic!!! (Much rejoicing)
>
>And then they ate Mr. Sieferman's minstrels...
>
>> The Fermin over Vizquel thing was the clincher. Hilarious. And Kevin
>> Mitchell?? get outta here! Thank you Mr. Burke for taking time to find
>> the holes and systematically present them.
The anti-stat argument presented here reminds me of the Argument From
Amazement which we see so often on the religious-oriented newsgroups:
you couldn't possibly believe that, could you??!!
There is a reason why the two zone-rating approaches to evaluating defense
is better than the standard examination of simply errors and assists.
They measure the relative ability of defensive players to cause opponent's
outs, which is the best thing that a defensive player can do. It gives a
higher rating to players who do more good things rather than (as 'errors'
do) rewarding players who simply don't do as many bad things.
These measures are not without problems. For instance, there is some
indication that the ratings for outfielders are slightly affected by the
performance of other outfielders. Nonetheless, as long as there is a
sufficiently large sample size (like several years) and no demonstrated
bias, they are the best means for evaluating relative defensive prowess
currently available.
These statistics show that Ken Griffey Jr and Roberto Alomar are not close
to the best fielders at their positions. This causes some people,
expecially fans and mediots in Seattle and Toronto, to feel extreme
consternation. This is not really much of a problem in the case of the
fans: since their role in the process is that of spectator, they can use
or dismiss any statistic that makes them happy. The anti-analytical
attitude of the media (or, worse, baseball management) is a more serious
problem. Sadly, most members of the media are baseball's equivalent of
creationists.
What's puzzling to me is the willingness of sports fans to swallow
'conventional wisdom' while maintaining a critical attitude in other areas
of life. The methodolody of statistical analysis applied to baseball is
no different than elsewhere: development of measurements, collection of
data, hypothesis testing, etc. If it doesn't add to your enjoyment of
the game, then don't use it. But don't pretend that alternative methods
of evaluation, like "gee, Griffey looks pretty spectacular to me" or "Joe
Carter's game winning home run in the World Series proves he's a clutch
hitter" have any value other than inoffensive gossip.
Jason Burke
> 'conventional wisdom' while maintaining a critical attitude in other areas
> of life. The methodolody of statistical analysis applied to baseball is
> no different than elsewhere: development of measurements, collection of
> data, hypothesis testing, etc. If it doesn't add to your enjoyment of
> the game, then don't use it. But don't pretend that alternative methods
> of evaluation, like "gee, Griffey looks pretty spectacular to me" or "Joe
> Carter's game winning home run in the World Series proves he's a clutch
> hitter" have any value other than inoffensive gossip.
I'm sorry, but I have to disagree about people paying any attention to
statistics in real life. Statistics say that smoking is one of the
leading causes of death in our country, yet millions of people continue
to smoke regularly. Statistics say that alcohol impairs your reaction
time, yet everyday thousands, if not hundreds of thousands, of people
drive while under the influence of alcohol. Statistics say that if you
drive 55 you get better gas mileage than if you drive faster, yet if you
drive 55 on the freeway, your so slow compared to the other cars on the
road that you are a hazard. I think that sports fans pay more attention
to statistics when they are related to sports then they do in any other
facet of their lives. I just feel that using a statistic to judge
defense is impossible. How can you modify for a team that moves around
their outfield vs a team that plays straight up. How to modify for
players that play 100 games on turf vs a player that plays 50 games on
turf. How do you modify for a player that plays shallow vs a player that
plays deeper. How do you modify for stadium differences. How do you
modify for wind. How do you modify for an excellently groomed infield vs
a poorly groomed infield. I think that you cannot quantitatively
determine defensive prowess just as you cannot determine who is the best
hitter in the league. Is there a stat that shows who the best hitter
is? No. You can say that someone that hits for a high average is a
better hitter, or someone that hits for power is a better hitter. Just
as you cannot factor in how someone is protected in a lineup, how many
fastballs vs off speed pitches a batter gets, how often he hits with
runners on base, you can't modify for defensive things either. Vizquel
moved from turf to grass, yet his range % dropped. Why? His fielding
percentage was about the same, so the field didn't cause him any more
errors, and as I recall three of the errors he made were on pop-ups
anyway. Wouldn't it follow that if Vizquel committed a similar number of
errors his range factor would be higher on slow grass than it would be on
fast astroturf? and using that line of reasoning, shouldn't Fermin's
range factor drop as he moves from a slower surface to a faster surface?
and I don't care if someone wants to call it a 1-year abberation, but
Kevin Mitchell does not have more range than Barry Bonds any season
unless Bonds has leg problems, and he didn't last year. Could it be that
do to the wind in Candlestick Park that Bonds had a lot of trouble with
balls last year? It might or might not be, but since this possibility is
not factored into the equation you must assume that Mitchell gets a
better jump on the ball than Bonds, and I would have to say that Mitchell
must get the best jump of any outfielder in baseball if he is going to
get to balls that Barry Bonds doesn't get to, because he has about half
of the speed. It is just like in basketball. For all of you Sonic fans
out there, Nate McMillan has a higher 3 point field goal percentage over
the last couple of years than Vernon Maxwell. However, if you NBA scouts
throughout the league and asked who they would rather have to take the 3
point shot to win the championship, almost all of them would say
Maxwell. and just as Zone Ratings don't factor in the things I mentioned
above, 3 point percentage does not factor in how a player handles
pressure, or how often the player shoots a wide open shot vs a shot with
someone in his face. Just some thoughts.
Jas
Will Shortt
Mr. Seiferman, please, the eyes glaze.
Until you are ready to take on the Vizquel - Fermin comparison (or any of
the others raised by Mr. Burke), enough already.
You didn't even address these comparisons, choosing instead to snip the
most excessive of our rejoicings and then rant tenuously about mediots
and creationists. Weltenshaung under attack? Questioning the
unquestionable dominance of the numerical over the narrative??
Good. Live a little.
By rejecting your argument, we are not blindly "following conventional
wisdom" as you so patronizingly charge. Quite the contrary, we are
believing our own eyes.
And now you can, too.
Tonight, please join me in welcoming back . . .Ken Griffey JUNIOR!!!
And there ain't nothin' wrong with his legs.
- Will "Sure I'm a fan. That's the point." Shortt
Richard Cameron
>At this point in his career, Alomar does not even have average range.
His range
>disappeared after his winter ball injury a couple of years ago.
Ok, now this just got me. I apologize again THIS WILL BE MY LAST POST
ON THE SUBJECT OF THIS STAT.
To say that Alomar has no range, Umm, are we thinking of the same
Roberto Alomar? You know THE BEST DEFENSIVE 2Bman in the game. Ask
anyone, experts, players, owners, gm's, anyone! That is why he is
going to get a huge contract as a free agent this year.
By the way, you said this stage in his career, how old do you think he
is? He is 27 or 28! I dont consider this to be old!
Eric Sieferman
Will Shortt <sho...@u.washington.edu> wrote:
>
>Until you are ready to take on the Vizquel - Fermin comparison (or any of
>the others raised by Mr. Burke), enough already.
Since this was done more than adequately by another poster, I won't
repeat it.
>You didn't even address these comparisons, choosing instead to snip the
>most excessive of our rejoicings and then rant tenuously about mediots
>and creationists. Weltenshaung under attack? Questioning the
>unquestionable dominance of the numerical over the narrative??
Except that isn't what was happening. So far in this thread, there has
been one post that brought up thoughtful critiques of the use of zone-type
statistics, and most of those criticisms were answered. The rest of the
comments were mostly of the form "oh year?" or "everyone knows".
Personally, I don't find these arguments very convincing in any area of
life, not even in the entertainment areas.
>Good. Live a little.
>
>By rejecting your argument, we are not blindly "following conventional
>wisdom" as you so patronizingly charge.
This is probably true for you, since your posts are normally thoughtful
and frequently entertaining. It is not true for the majority of anti-stat
people who incessantly post to rec.sport.baseball.
>Quite the contrary, we are
>believing our own eyes.
The point is that there is more going on than what we see with our eyes.
The data on defense is compiled by people who watch EVERY ball hit in
EVERY major league game, and do nothing else. When I go to games, I'm
watching all sorts of things, and even if I had a photographic memory I
doubt that I give you much of a ranking of the defensive skills of every
player at any position.
Using meaningful statistics can illuminate some questions, like "Is Devon
White really a better fielder than Ken Griffey Jr" (or, in more serious
areas, "Are most welfare recipients really people who breed indiscrimately
and don't want to work?")
Eric Sieferman
Jason Burke <val...@u.washington.edu> wrote:
>> What's puzzling to me is the willingness of sports fans to swallow
>> 'conventional wisdom' while maintaining a critical attitude in other areas
>> of life. The methodolody of statistical analysis applied to baseball is
>> no different than elsewhere: development of measurements, collection of
>> data, hypothesis testing, etc. If it doesn't add to your enjoyment of
>> the game, then don't use it. But don't pretend that alternative methods
>> of evaluation, like "gee, Griffey looks pretty spectacular to me" or "Joe
>> Carter's game winning home run in the World Series proves he's a clutch
>> hitter" have any value other than inoffensive gossip.
>
>I'm sorry, but I have to disagree about people paying any attention to
>statistics in real life. Statistics say that smoking is one of the
>leading causes of death in our country, yet millions of people continue
>to smoke regularly. Statistics say that alcohol impairs your reaction
>time, yet everyday thousands, if not hundreds of thousands, of people
>drive while under the influence of alcohol. Statistics say that if you
>drive 55 you get better gas mileage than if you drive faster, yet if you
>drive 55 on the freeway, your so slow compared to the other cars on the
>road that you are a hazard.
Good points. I'm thinking of the people whom I meet on the net, who
usually have achieved a high level of education and tend to be more
mathematically literate than the rest of the population. The sort of
situation I was talking about was hearing a person mock the use of
statistics, and then turn around and claim that a player is a "clutch"
hitter because he batted .400 in one World Series.
>I think that sports fans pay more attention
>to statistics when they are related to sports then they do in any other
>facet of their lives. I just feel that using a statistic to judge
>defense is impossible. How can you modify for a team that moves around
>their outfield vs a team that plays straight up.
By looking at a large sample size. By looking at the data for players
who move from one type of team to another.
> How to modify for
>players that play 100 games on turf vs a player that plays 50 games on
>turf.
By looking at data for individual players, and comparing their performance
on turf vs grass. By comparing the data for all players on a particular
field, like the Kingdome. And, by not putting too much stock in numbers
that are tabulated for only 50 games.
>How do you modify for a player that plays shallow vs a player that
>plays deeper.
Well, positioning is an aspect of fielding. If a player consistently
plays in the wrong position, he'll not make as many outs. As someone
else pointed out, consistently bad positioning MAY be due to poor
management. After the data show that a player's defensive prowess is
less than adequate, the analysis goes beyond this to determine why. Bad
positioning is one of the possible explanations, and is usually
determined by (*gasp*) actually watching games.
> How do you modify for stadium differences.
People who do this analysis tell me that this is one of the areas of
current research. A rough idea can be obtained now by looking at
comparisons for similar stadiums.
> How do you
>modify for wind. How do you modify for an excellently groomed infield vs
>a poorly groomed infield.
It's just as windy for both teams. With the possible exception of a
place like Candlestick Park, the climate variations are probably not
important. As said before, more research is being done.
> I think that you cannot quantitatively
>determine defensive prowess just as you cannot determine who is the best
>hitter in the league. Is there a stat that shows who the best hitter
>is? No. You can say that someone that hits for a high average is a
>better hitter, or someone that hits for power is a better hitter.
There is no single statistic which says who is the best hitter or the
best fielder. This is an impossible standard. However, there are some
statistics which are better than others. Part of the research is to find
out which statistics are more meaningful in examining individual
performance. OBP is a better indicator of individual performance than is
runs-scored, because the latter is much more dependent on the actions of
other team members. The analysis of defensive zones is a better
indication of defensive skill than the conventional measures (like
errors) because it more accurately measures a player's ability to turn
hit balls into outs, which is the reason he's in the field.
> Just
>as you cannot factor in how someone is protected in a lineup,
Why not?
>how many
>fastballs vs off speed pitches a batter gets, how often he hits with
>runners on base,
The latter is measured now; I don't know if anyone compiles the former
league wide.
> you can't modify for defensive things either. Vizquel
>moved from turf to grass, yet his range % dropped. Why?
Good question. One of the problem with analyzing Vizquel's numbers is
that he didn't play very many games in Cleveland last year.
> His fielding
>percentage was about the same, so the field didn't cause him any more
>errors, and as I recall three of the errors he made were on pop-ups
>anyway. Wouldn't it follow that if Vizquel committed a similar number of
>errors his range factor would be higher on slow grass than it would be on
>fast astroturf? and using that line of reasoning, shouldn't Fermin's
>range factor drop as he moves from a slower surface to a faster surface?
>and I don't care if someone wants to call it a 1-year abberation, but
>Kevin Mitchell does not have more range than Barry Bonds any season
>unless Bonds has leg problems, and he didn't last year.
Maybe not. How do you know that?
> Could it be that
>do to the wind in Candlestick Park that Bonds had a lot of trouble with
>balls last year?
At the moment, there's no shortage of people in SF who will tell you that
it's because Bonds doesn't "hustle" and has a "bad attitude". Not me, of
course.
> It might or might not be, but since this possibility is
>not factored into the equation you must assume that Mitchell gets a
>better jump on the ball than Bonds, and I would have to say that Mitchell
>must get the best jump of any outfielder in baseball if he is going to
>get to balls that Barry Bonds doesn't get to, because he has about half
>of the speed.
Really?
> It is just like in basketball. For all of you Sonic fans
>out there, Nate McMillan has a higher 3 point field goal percentage over
>the last couple of years than Vernon Maxwell. However, if you NBA scouts
>throughout the league and asked who they would rather have to take the 3
>point shot to win the championship, almost all of them would say
>Maxwell.
I'll bet that the scouts for the Houston Rockets wouldn't say that.
> and just as Zone Ratings don't factor in the things I mentioned
>above, 3 point percentage does not factor in how a player handles
>pressure, or how often the player shoots a wide open shot vs a shot with
>someone in his face. Just some thoughts.
Good thoughts.
Nowhere did I or anyone claim that the zone rating systems are not
without flaws, or couldn't be improved. The only claim is that they are
the best measures available. It's only important because we find
baseball questions like "who's the best?" interesting. Or:
Who's the best defensive centerfielder in major league baseball?
(a) the player who looks the most impressive.
(b) the player with the most Gold Glove awards.
(c) everyone knows the answer, you doofus.
(d) the player who makes the fewest errors.
(e) the player who causes the highest percentage of outs on balls hit
into his defensive area.
(f) we'll never know, so why bother doing the analysis.
That's how it all began.
David Pease
Jason Burke <val...@u.washington.edu> wrote:
Before we go in to your numerous points, please keep in mind that
defense seems to fluctuate nearly as much as offense does, by either
ZR or DA. A player will frequently deviate by a large margin from his
normal performance. Just as the sun will shine on Eddie Sprague's bat
every once in a blue moon, so will fielders such as Tim Salmon or Kenny
Lofton have a bad year every once in a while. One year of any stat
really isn't very indicative--when you see that Harold Reynolds managed
to bat .300 one year, are you ready to indict batting average as a stat?
[deletia...]
>I see Roberto Alomar regularly make plays on the Shortstop side of
>second all the way over to behind the firstbaseman,
How many times have you seen Roberto Alomar play defense? TV really
doesn't count, since it's not a good medium through which to evaluate
defensive play.
>but if you look at
>the Zone Rating, our own Felix Fermin gets to an amazing 91% of the balls
>in his zone, while Alomar gets to something like 85%, only problem for
>Roberto is that the average for a second baseman is 86%, so not only is
>does Fermin have more range than Alomar, but Alomar does not even have
>average range.
That was a gargantuan sentence there.
I don't understand what the problem is, though. So Alomar doesn't get to
many balls, flashy as he looks. So what?
>At Shortstop.
>
>Who would you guys say has the best range at shortstop? Vizquel, Barry
>Larkin, Ozzie Smith, Shawn Dunston, or Felix Fermin? Why are we bringing
>up Alex, Felix Fermin has a Zone rating of like 92.8%.
Because Fermin can't hit.
Was this a serious question?
>Barry Larkin
>hangs in there with 90.5%, but our man Omar is below average at 88%.(The
>average for shortstops is 89%). I am sorry, but I have seen all of these
>guys a bunch and Felix Fermin doesn't get to half of the balls that guys
>like Vizquel or Larkin get to.
Sure.
To test your ability to recieve, store, and retrieve large amounts of
information, please rate the following Mariners in order of on-base
percentage, without using statistics.
Rich Amaral
Mike Blowers
Jay Buhner
Tino Martinez
Warren Newson
Alex Rodriguez
Luis Sojo
Dan Wilson
I hope you don't get any wrong, because defense is much harder to evaluate
in your mind alone than something as clear-cut as OBP is.
[deletia...]
>even have the worst range of any everyday third baseman, but can anyone
>sit there and honestly tell me that Blowers covers more ground at third
>than Edgar?
You said it, not me.
[deletia...]
>say that all of these guys (O'Neill, Salmon, and even Jay) have a little
>better range than Sierra,
You would say that Jay Buhner has better range than *anyone*?
Why?
but for those of you that watch the Mariners
>and few other teams, the Zone Range that I just couldn't believe, is that
>Ken Griffey Jr. should not be playing centerfield. It should be Jay
>Buhner! For all of you people that bitch about Buhner's range, his 77%
>is better than Griffey's 76%. That's been our problem, we should have
>Jay in center and Junior in right!
ZR in center != ZR in right.
I've deleted the rest of your post to save some bandwidth; your arguments
are fine, but what you are basically saying is "the statistic is not valid
because it doesn't produce the numbers I think it should".
Call me old-fashioned, but I don't buy that kind of argument. Perhaps
if you have a problem with the way DA or ZR is compiled, I would be more
moved.
I'm certainly ready to be proven wrong, but "from what I've seen" doesn't
cut it.
Thank you for your time
dp
--
dave pease dpe...@qualcomm.com
"Hee ith Veego. You are like the butthing of FLITH to heem!"
- Yanosh Poha, _GB_II_ Resident VigoTool
David Pease
Richard Cameron <ric...@ix.netcom.com> wrote:
>
>
>>At this point in his career, Alomar does not even have average range.
>His range
>>disappeared after his winter ball injury a couple of years ago.
>
>ON THE SUBJECT OF THIS STAT.
We can hope...
>To say that Alomar has no range, Umm, are we thinking of the same
>Roberto Alomar? You know THE BEST DEFENSIVE 2Bman in the game.
OK Richard, by now you know the drill.
How many times have you seen Alomar play? As opposed to how many times
you've seen a media type fall over themselves praising him? I'd guess
the ratio is about 10 media types for every Alomar game you've seen.
What makes you think that you are qualified to evaluate good defense
merely by what you see and remember? Human brains are notoriously bad
at cataloging hundreds of repetitions of an event without losing any
or distorting the memories.
And who is your control 2B? How do you know Alomar is even above
average? How do you know what average is?
Etc. etc. etc.
>Ask anyone, experts, players, owners, gm's, anyone! That is why he is
>going to get a huge contract as a free agent this year.
No, he's going to get a huge contract because he's the best offensive
2B in baseball, and I think anyone will tell you that.
If Alomar hit like Luis Lopez, he wouldn't be staring millions upon
millions of dollars in the face.
>By the way, you said this stage in his career, how old do you think he
>is? He is 27 or 28! I dont consider this to be old!
Of course you don't. Do you know that most players peak around the age
of 27? That Alomar will be declining, in all probability, for at least
some of the years that he gets on that multiyear contract that he's sure
to sign with someone during the off-season?
Of course you don't.
Michael Cox
(Eric Sieferman) wrote:
> The data on defense is compiled by people
Wait! Stop right there! I think you're starting to get the idea...
David Nieporent
>>At this point in his career, Alomar does not even have average range.
>>His range disappeared after his winter ball injury a couple of years ago.
>Ok, now this just got me. I apologize again THIS WILL BE MY LAST POST
>ON THE SUBJECT OF THIS STAT.
>To say that Alomar has no range, Umm, are we thinking of the same
>Roberto Alomar?
No. The people using DA are talking about the *actual* Roberto Alomar,
and you're talking about the anecdotal Alomar.
>You know THE BEST DEFENSIVE 2Bman in the game.
No.
>Ask anyone, experts, players, owners, gm's, anyone!
Why? I prefer to think for myself, rather than to ask other people what
I should think. You, evidentally, need to be told the proper things to
think.
If I'm going to ask anybody, it's going to be somebody who has seen every
game Alomar AND EVERY OTHER SECOND BASEMAN IN THE LEAGUE has played.
Because that's the only way to compare them. Unfortunately, since that
person doesn't exist, I'll use the compiled observations of people who
watch every game.
>That is why he is
>going to get a huge contract as a free agent this year.
Gee, I thought that possibly might have had something to do with his
offense. Did you ever see Jose Lind getting a huge free agent contract?
>By the way, you said this stage in his career, how old do you think he
>is? He is 27 or 28! I dont consider this to be old!
Where did he say anything about age?
"At this stage in his career" refers to his post-injury performance.
--
David M. Nieporent "Scouts work in a field in which they are never
Niep...@pluto.njcc.com judged for their decisions, sort of like the post
Deer Creek/Plainsboro, NJ office, but without the guns." ---- Chris Kahrl
ORIOLES 1995!!!!!!!! ReganWatch: Orioles Held Hostage, Day 112
Jason Burke
On 16 Aug 1995, David Pease wrote:
> In article <Pine.A32.3.91j.95081...@homer23.u.washington.edu>,
> Jason Burke <val...@u.washington.edu> wrote:
>
> Before we go in to your numerous points, please keep in mind that
> defense seems to fluctuate nearly as much as offense does, by either
> ZR or DA. A player will frequently deviate by a large margin from his
> normal performance. Just as the sun will shine on Eddie Sprague's bat
> every once in a blue moon, so will fielders such as Tim Salmon or Kenny
> Lofton have a bad year every once in a while. One year of any stat
> really isn't very indicative--when you see that Harold Reynolds managed
> to bat .300 one year, are you ready to indict batting average as a stat?
---Players may have a bad year, but how does that effect their range. A
player may steal 50 bases one year and almost never get caught, and then
only steal like 20 and get caught half of the time, but does that mean
they are all of a sudden slow? As far as batting average goes, that is a
stat which is not subject to perception(except for errors). You watch
Edgar walk up to the plate and single to center, he's 1 for 1, no
debate.
>
> >I see Roberto Alomar regularly make plays on the Shortstop side of
> >second all the way over to behind the firstbaseman,
>
> How many times have you seen Roberto Alomar play defense? TV really
> doesn't count, since it's not a good medium through which to evaluate
> defensive play.
---I would say that I have seen him play about 40 times in person and I
think that is enough to determine what he is going to get to and what he
isn't going to get to.
> >but if you look at
> >the Zone Rating, our own Felix Fermin gets to an amazing 91% of the balls
> >in his zone, while Alomar gets to something like 85%, only problem for
> >Roberto is that the average for a second baseman is 86%, so not only is
> >does Fermin have more range than Alomar, but Alomar does not even have
> >average range.
>
> That was a gargantuan sentence there.
>
> I don't understand what the problem is, though. So Alomar doesn't get to
> many balls, flashy as he looks. So what?
---I am not necessarily quiestioning whether Roberto is average or not
because as you mentioned I probably don't know what average is, I am just
saying that Fermin does not have better range than Alomar, yet the
statistic says that he does.
> >At Shortstop.
> >
> >Who would you guys say has the best range at shortstop? Vizquel, Barry
> >Larkin, Ozzie Smith, Shawn Dunston, or Felix Fermin? Why are we bringing
> >up Alex, Felix Fermin has a Zone rating of like 92.8%.
>
> Because Fermin can't hit.
---and Alex can?
> Was this a serious question?
---No
> >Barry Larkin
> >hangs in there with 90.5%, but our man Omar is below average at 88%.(The
> >average for shortstops is 89%). I am sorry, but I have seen all of these
> >guys a bunch and Felix Fermin doesn't get to half of the balls that guys
> >like Vizquel or Larkin get to.
>
> Sure.
>
> To test your ability to recieve, store, and retrieve large amounts of
> information, please rate the following Mariners in order of on-base
> percentage, without using statistics.
>
> Rich Amaral(5)-seems to have a decent eye, but not a great eye
> Mike Blowers(4)-gets pitched around alot because of who hits behind him
> Jay Buhner(3)-usually has a high number of walks
> Tino Martinez(2)-gets pitched around and intentionally walked alot
> Warren Newson(1)-he seems to walk a bunch and he hasn't played many games
> Alex Rodriguez(8)-never walks
> Luis Sojo(7)-swings at too many bad balls to have a high OBP
> Dan Wilson(6)-walks enough to be better than Sojo
>
> I hope you don't get any wrong, because defense is much harder to evaluate
> in your mind alone than something as clear-cut as OBP is.
---I will see how I do, but I still say that OBP and batting average are
OK stats because they are not based on anything other than seeing if
someone gets a hit or is walked or not.
> but for those of you that watch the Mariners
> >and few other teams, the Zone Range that I just couldn't believe, is that
> >Ken Griffey Jr. should not be playing centerfield. It should be Jay
> >Buhner! For all of you people that bitch about Buhner's range, his 77%
> >is better than Griffey's 76%. That's been our problem, we should have
> >Jay in center and Junior in right!
>
> ZR in center != ZR in right.
---I assumed that the zones were equidistant so whether you play center,
right, or left shouldn't matter.
> I've deleted the rest of your post to save some bandwidth; your arguments
> are fine, but what you are basically saying is "the statistic is not valid
> because it doesn't produce the numbers I think it should".
>
> Call me old-fashioned, but I don't buy that kind of argument. Perhaps
> if you have a problem with the way DA or ZR is compiled, I would be more
> moved.
---Oh I definitely have that. I think that things such as positioning,
weather conditions, other players around you, stadium conditions(fence
distances and turf or grass) have a huge effect that is not taken into
consideration. But my number one gripe about this measure is that it is
not consistent. As I mentioned in an earlier post, when Fermin moved
from a grass infield in Cleveland to a turf infield in Seattle and had
his range go up while Omar moved from turf to grass and had his range go
down is something that I question. It may be easier to field on turf
because you don't get bad hops, but the balls travels so much faster on
turf that it would only follow that range would be effected. I also have
a problem with the fact that range seems to very so much. As I stated
earlier it is like speed, your stolen bases might change, but your speed
stays pretty constant, same thing with range. Your errors might change,
or your assists might change, but your range should stay pretty constant
assuming that you are playing on the same type of field (turf or grass)
and you haven't had any major leg or back injuries.
> I'm certainly ready to be proven wrong, but "from what I've seen"
doesn't
> cut it.
---I am curious why you think that the people that
compile these statistics have are so much more qualified than anybody else
to watch a baseball game and make judgements. As you said yourself,
watching a game on television does not give you a good idea of what is
going on, so why do you think that when I see something that my argument
doesn't hold water and yet when someone who in essence does the same thing
is like the word of god. When I go to a game I watch how batters are
pitched to, how the defense is positioned, if guys hit cut off men and all
of the other little things that make this game so interesting to watch but
since I don't write it down and publish it in a book my perceptions cannot
be taken into account. I would have to disagree with that logic I am
afraid.
Jas
Eric Sieferman
Michael Cox <mich...@accessone.com> wrote:
>In article <40t4k7$h...@nntp4.u.washington.edu>, sief...@u.washington.edu
>(Eric Sieferman) wrote:
>
>
>> The data on defense is compiled by people
>
>Wait! Stop right there! I think you're starting to get the idea...
Is this a variant on the
since-everyone-makes-mistakes-no-knowledge-is-possible philosophy?
DA's and ZR's are not perfect. They are better than other commonly used
measures of defensive skill. Are you starting to get the idea?
David Nieporent
Will Shortt <sho...@u.washington.edu> wrote:
>Statistics are only the result of formulae which are meant to approximate
>reality.
Wrong. Statistics are the result of counting what has happened on the field.
>The more complex the formula, the more likely intervening
>variables will jam the works skewing the results. The stats that make
>baseball great - Ave, RBI's, ERA's, W-L, etc. - are all relatively
>simple, but still can be misleading. Example: Blowers and Buhner's nearly
>identical stats - yet having very different years.
Where do you get this stuff? Blowers is hitting 268/335/491. Buhner is
hitting 266/350/542. (As of last Sunday.) Buhner is hitting
significantly better, although I think that positional adjustments
probably make them equally valuable.)
>The point, however, is that these simple stats help make baseball great.
Guess what -- at one time each of the stats listed above were considered
complex and newfangled and "not reality." RBIs? "Teammates who happen
to advance a base after I hit the ball, but only if they make it to
home plate, unless I hit into a double play, or unless the scorer judges
that they scored on a play in the field rather than on my hit. Or if I
hit a home run, I get one." That's simple? Only because you're used to
it.
W-L and Saves are *really* complex. How many people post on
rec.sport.baseball "what's the save rule again?"
>I've just moved to Eugene, OR and have had the pleasure to see the Class
>A Emeralds here twice now. Of course, I knew none of the players, but
>armed with all the stats on the program, I was able within ten minutes to
>get a feel for the players and thus enjoyed the game much more. (BTW,
>ex-husky CF Joe Trippy is hitting .305 and leads the Em's in RBI's).
>However, the elaborate formula Mr. Pease presents serves only to
>obfuscate rather than illuminate.
Um, excuse me, but what person over the age of 3 is confused by
"successes/opportunities?" Because that's the "elaborate formula" we're
talking about. Exactly as complex as batting average.
>We all had gut feelings and anecdotal
>evidence that something was fishy with Griffey's readout. Certainly
>everyone in baseball couldn't have been wrong about the Kid.
Certainly they could. At one time everybody thought the earth was the
center of the solar system. After all, it *looked* as if everything went
around the earth. Guess what -- they were all wrong.
A belief doesn't become true because a lot of people hold it.
>Certainly
>his record streak a couple years back of opportunities without an error
>couldn't have been done by a sub-par CF.
Nonsense. The worst fielder in history could do this. I could easily
set the record. All this says is that he doesn't drop routine balls hit
to him. Well, neither do I. I wouldn't/don't get to anything more than a
few feet away from me, but then again, that doesn't count in the "error"
stat, even though it's exactly the same result for the team.
>Could we have only been seeing
>his good games and missing the reality of Griffey's slow legs and limited
>range?
No, it's not that you only see his good games. It's called an optical
illusion. You thought you saw something which wasn't actually there.
Also, since you rarely saw the other CFs in the league, you didn't have
anything to compare him to.
>Confusion, emotional reaction ensued.
>And then Jason stepped to the plate.
>WOW! Whattapost!! Knocked that formula deep, deep, deep into center.
>Pease going backbackback . . . goodbye statistic!!! (Much rejoicing)
>The Fermin over Vizquel thing was the clincher. Hilarious. And Kevin
>Mitchell?? get outta here! Thank you Mr. Burke for taking time to find
>the holes and systematically present them.
Hint: "It doesn't agree with my uninformed opinion" is not a "hole."
>And Mr. Pease, your initial post sparked the most spirited debate I've
>seen in this group. For that, I thank you. But remember, while we all love
>stats, they are NOT the numerical equivalent to reality. At best they are
>an approximation which helps us gauge reality. They are not a reality
>substitute, and when they fly in the face of the observations of millions
>of serious fans of the game, they must be suspect.
"Millions?" Get real. Most people don't see Ken Griffey more than a few
times a year. They simply say he's great because they hear other people
saying it, and like you, would rather parrot what others say than think
for themselves.
Unscientific observations, filtered through human memory, are hardly
"reality" either. They are not a reality substitute, and when they fly
in the face of the observations of millions of fans of the game, they
must be suspect. (But when *I* say "observations of millions of fans,"
I mean stats. Carefully collected observations, not filtered through
memory, but recorded.)
>But if you really love stats, you gotta love the kingdome, right? no
>intervening weather, light , or field condition variables - just the
>perfect little petri dish for statistical observation. What's the fuss
>all about anyway?
There are field conditions in a dome.
--
David M. Nieporent "Scouts work in a field in which they are never
Niep...@pluto.njcc.com judged for their decisions, sort of like the post
Deer Creek/Plainsboro, NJ office, but without the guns." ---- Chris Kahrl
ORIOLES 1995!!!!!!!! ReganWatch: Orioles Held Hostage, Day 113