PDA

View Full Version : Funny Look At SABR Geeks


spiffie
06-05-2007, 01:15 PM
Apologies if this was posted before, I didn't see it anywhere.

An English-Sabrmetric Dictionary (http://mvn.com/mlb-stats/2007/05/21/a-sabermetricenglish-dictionary/)

A sample:
Pythagorean win percentage: An formula which tells you that despite the fact that another team is piled on top of each other after winning the World Series, your team was actually better this year. And we can prove it. (Gee, that makes me feel better.)

eriqjaffe
06-05-2007, 01:23 PM
Bad link.

Try http://mvn.com/mlb-stats/2007/05/21/a-sabermetricenglish-dictionary/ instead.

WLL1855
06-07-2007, 07:38 PM
On a side note, just try reading some of this crap.

http://www.baseballprospectus.com/article.php?articleid=2518

Who dreams this stuff up?

Daver
06-07-2007, 07:42 PM
On a side note, just try reading some of this crap.

http://www.baseballprospectus.com/article.php?articleid=2518

Who dreams this stuff up?

Propellerheads.

eriqjaffe
06-08-2007, 12:16 AM
My brother does this stuff:

http://www.hardballtimes.com/main/article/managers-and-leveraging-sp-lev-part-7/

He even got featured in the Wall Street Jounral (http://online.wsj.com/public/article/SB116138190296599390-SHHWwYDudmQP3epK5EceIoTxWSA_20061027.html?mod=mktw ) last October.

All in all, I'd have to say he's the most knowledgeable Cub fan I know - in a book-smart way, that is.

whitesoxfan
06-08-2007, 01:26 AM
Propellerheads.

That sums it up nicely.

These SABR freaks don't have a clue.

chaerulez
06-08-2007, 02:58 AM
I hope these people don't think Joe Morgan is a good announcer. He can sometimes make Santo seem smart.

StillMissOzzie
06-08-2007, 03:04 AM
My brother does this stuff:

All in all, I'd have to say he's the most knowledgeable Cub fan I know - in a book-smart way, that is.

With all due respect to your doctorate student brother...that, sir, is a tallest midget contest if ever I have seen one.

SMO
:o:

SBSoxFan
06-08-2007, 11:13 AM
On a side note, just try reading some of this crap.

http://www.baseballprospectus.com/article.php?articleid=2518

Who dreams this stuff up?

It's just math. More precisely it's just statistics. And some of it is even hindsight math, which is somewhat limited.

The other part of it is to eliminate as many variables as possible so it's easier to use math --- hence, things like TTO's are born (so, if a ball is put in the field of play it didn't really happen? :?:). Something Einstein said really fits in this case: "As far as the laws of mathematics refer to reality, they are not certain, as far as they are certain, they do not refer to reality."

McCuddy
06-09-2007, 06:27 AM
That sums it up nicely.

These SABR freaks don't have a clue.

Yeah, God forbid someone who likes baseball cracks a book.

IndianWhiteSox
06-09-2007, 09:22 AM
Yeah, God forbid someone who likes baseball cracks a book.

There's a difference between reading a book and a total ignoramus who writes inane BS about how to run a baseball team.

FarWestChicago
06-09-2007, 01:01 PM
What exactly are they doing wrong that warrants such scorn? They use science and objective reasoning to create better, more accurate statistics than the ones we have now. How terrible, grab the torches and pitchforks.Oh please, are you seriously uneducated? They do not use "science", nor are the remotely "objective". They violate every principle of the scientific method in everything they do. There is no such thing as randomization, double blinding or the null hypothesis. Statheads misuses numbers to aggrandize themselves. Period. When one of them steps up and actually makes a statistic that proves themselves WRONG, we will have a breakthrough. Until then, enjoy the snakeoil you are buying. :cool:

comerica
06-11-2007, 11:41 PM
They violate every principle of the scientific method in everything they do. There is no such thing as randomization, double blinding or the null hypothesis.


they're doing statistical analysis not experiments, the scientific method has absolutely nothing to do with this discussion.

Hatred of sabermetrics is almost always a result of the person not being able to understand the concepts.

MRM
06-12-2007, 01:37 AM
On a side note, just try reading some of this crap.

http://www.baseballprospectus.com/article.php?articleid=2518

Who dreams this stuff up?

Non-athletic types who couldn't play tiddly winks with any sort of dexterity who need to feel like they contribute to the game in some way by making it far more complicated than it actually is.

MRM
06-12-2007, 01:54 AM
Hatred of sabermetrics is almost always a result of the person not being able to understand the concepts.

ROFL. So it's a matter of the dumb guy hating the "smart" guy for being smart, eh?

I fully understand the concepts. The problem is the concepts rarely have anything remotely useful in them. This is why making fun of sabermetrics is so easy.

It's said that baseball is a game of statistics. H, R, HR, RBI, etc. but those aren't true statistics, they are simply cumulative totals with no statistical value whatsoever. Sabermetrics tries to take those values and give them statistical significance in a vain attempt to "explain" the game to those of us who with less "insight".

The 2005 sabermetric "world series" certainly turned out significantly different than the one actually played on the field :D:. Sabermetrics doesn't allow for a human element and baseball depends on a HUGE human element. I'm still waiting on the sabermetrics theory concerning the song "Don't stop believing" and it's potential effect on team unity!

DSpivack
06-12-2007, 02:01 AM
ROFL. So it's a matter of the dumb guy hating the "smart" guy for being smart, eh?

I fully understand the concepts. The problem is the concepts rarely have anything remotely useful in them. This is why making fun of sabermetrics is so easy.

It's said that baseball is a game of statistics. H, R, HR, RBI, etc. but those aren't true statistics, they are simply cumulative totals with no statistical value whatsoever. Sabermetrics tries to take those values and give them statistical significance in a vain attempt to "explain" the game to those of us who with less "insight".

The 2005 sabermetric "world series" certainly turned out significantly different than the one actually played on the field :D:. Sabermetrics doesn't allow for a human element and baseball depends on a HUGE human element. I'm still waiting on the sabermetrics theory concerning the song "Don't stop believing" and it's potential effect on team unity!

A poster here, and I apologize as I don't remember who, recently said that it's one thing to use statistics to reflect what has happened, it is quite another to use them and try to extropolate out what will happen. Predicting the future is a rather futile and useless task. And yet what were once obscure stats have now become relatively commonplace; the save, obp %, slugging %, OPS, and two of my favorites, ERA+ and OPS+. Some stats are useless, some are completely made-up, sometimes they're used to show why the A's should have won every world series ever. Hence, they are only as good as the humans that use them. They can be helpful in looking at what has happened, but they are by no means a replacement for actually watching a baseball game. I like some statistics that sabermetrics has come up with, but if what sabermetricians are trying to do is look at the game solely through those lenses, then I don't like them nor do I agree with their approach.

MRM
06-12-2007, 02:18 AM
A poster here, and I apologize as I don't remember who, recently said that it's one thing to use statistics to reflect what has happened, it is quite another to use them and try to extropolate out what will happen. Predicting the future is a rather futile and useless task. And yet what were once obscure stats have now become relatively commonplace; the save, obp %, slugging %, OPS, and two of my favorites, ERA+ and OPS+. Some stats are useless, some are completely made-up, sometimes they're used to show why the A's should have won every world series ever. Hence, they are only as good as the humans that use them. They can be helpful in looking at what has happened, but they are by no means a replacement
for actually watching a baseball game. I like some statistics that sabermetrics has come up with, but if what sabermetricians are trying to do is look at the game solely through those lenses, then I don't like them nor do I agree with their approach.

Most statistical analysis of any sport is an attempt to distort reality. Rarely does it validate actual results. Sabermetrics mostly attempts to explain why actual results are wrong.

Bill Parcells was fond of saying "you are what your record says you are". Even though he was a stat-head himself, big Bill was completely aware that past stats are never predictive of future results, particularly in team sports where the variables are immense and not entirely tied to measurables such as numbers. Sabermetricians have yet to learn that simple law of sports.

ma-gaga
06-12-2007, 10:20 AM
sometimes they're used to show why the A's should have won every world series ever.

I'm calling you out. I'd like to see this stat.

:gulp:

spiffie
06-12-2007, 10:30 AM
I'm calling you out. I'd like to see this stat.

:gulp:
It's the BBINW%, which objectively shows who the best team really is. BBINW of course means Billy Beane Is Never Wrong.

I have no problem with statistical analysis. Used judiciously in combination with scouting and evaluating the human element of the game it can help to make decisions. The problem is that there seems to be an element among the SABR world, the truly hard-cores, who almost are angry that the game has to be played. The ones who would seemingly would rather decide the champion each year by who has the highest 3rd order win pct. And much like zealots of any sort, they have the impact of turning people against everything they preach about, even the stuff that might make some sense.

voodoochile
06-12-2007, 10:56 AM
The Sox appear to be headed toward proving or disproving the VORP stat this season as pretty soon they will be nothing but Replacement Players.

comerica
06-12-2007, 12:49 PM
The Sox appear to be headed toward proving or disproving the VORP stat this season as pretty soon they will be nothing but Replacement Players.
:D: that was good.

DSpivack
06-12-2007, 02:24 PM
The Sox appear to be headed toward proving or disproving the VORP stat this season as pretty soon they will be nothing but Replacement Players.

:rolling:

FarWestChicago
06-12-2007, 09:24 PM
they're doing statistical analysis not experiments, the scientific method has absolutely nothing to do with this discussion.

Hatred of sabermetrics is almost always a result of the person not being able to understand the concepts.I didn't bring up "science". SoxItToEM did. Look at the quote in my post (he has deleted his original post). I was responding to him. Now I'll respond to you. Inventing equations and adding factors until the results fit your a priori notions is not analysis, it's manipulation. I don't hate sabremetrics. I find much of it interesting. But it is not an accurate description of reality.

comerica
06-12-2007, 10:56 PM
well I retract that statement as it pertains to you then. It's just that I've seen people badly misunderstand some of it on this site before. A conversation about the lack of relevance of a pitchers record in evaluating him comes to mind.

chaerulez
06-13-2007, 04:49 AM
well I retract that statement as it pertains to you then. It's just that I've seen people badly misunderstand some of it on this site before. A conversation about the lack of relevance of a pitchers record in evaluating him comes to mind.

I don't think any intelligent baseball fan, regardless if they consider themselves into sabrmetrics or not would disagree that win-loss record is generally not a good way to determine how good a starting pitcher is.

MRM
06-13-2007, 05:24 AM
I don't think any intelligent baseball fan, regardless if they consider themselves into sabrmetrics or not would disagree that win-loss record is generally not a good way to determine how good a starting pitcher is.

Win-loss record is a HORRIBLE way to judge how good a pitcher is. It might be the absolute worst number to look at.

voodoochile
06-13-2007, 09:59 AM
Win-loss record is a HORRIBLE way to judge how good a pitcher is. It might be the absolute worst number to look at.

Well then bring back Hawk as GM. He's the one who traded Bobby Bonilla for a guy with great stats and a horrible W/L pct. I forget the pitcher's name but he never did live up to his stat hype.

Who knew Hawk was BB before the world knew who BB was?

Wins may be a team stat first, but to completely exclude them from an evalution of a pitcher is simply stupid. If nothing else they show a pitcher's "clutch" factor - how they perform in game changing innings. Of course there isn't much room for those kinds of evaluations in statistical analysis. It's one of the key reasons I don't buy a lot of these new stats. Baseball unlike most other major sports is a team game that is decided by a series of individual plays and when there are that many people doing their own thing it becomes very difficult to predict with any accuracy. I honestly believe you'd have to get deep into some very heavy math to really start to model it. Chaos theory comes immediately to mind...

Johnny Mostil
06-13-2007, 10:27 AM
Well then bring back Hawk as GM. He's the one who traded Bobby Bonilla for a guy with great stats and a horrible W/L pct. I forget the pitcher's name but he never did live up to his stat hype.



You're recalling Jose DeLeon, I do believe. I can't remember much about him, but, as I recall (accurately?), his record was something like 2-19 the year before the Sox acquired him.

spiffie
06-13-2007, 10:28 AM
Well then bring back Hawk as GM. He's the one who traded Bobby Bonilla for a guy with great stats and a horrible W/L pct. I forget the pitcher's name but he never did live up to his stat hype.

Who knew Hawk was BB before the world knew who BB was?

Wins may be a team stat first, but to completely exclude them from an evalution of a pitcher is simply stupid. If nothing else they show a pitcher's "clutch" factor - how they perform in game changing innings. Of course there isn't much room for those kinds of evaluations in statistical analysis. It's one of the key reasons I don't buy a lot of these new stats. Baseball unlike most other major sports is a team game that is decided by a series of individual plays and when there are that many people doing their own thing it becomes very difficult to predict with any accuracy. I honestly believe you'd have to get deep into some very heavy math to really start to model it. Chaos theory comes immediately to mind...
I agree with the idea of this, but Jose DeLeon is a rather poor example to use. Yes, he had a terrible W-L record, but he never really had other good stats either.

He had a nice rookie season, starting 15 games with an ERA of 2.83 (ERA+ 131) and a 7-3 record in 1983. But then in 1984 he went 7-13 with a 3.74 ERA (96 ERA+) and in 1985 he had his nightmare year of 2-19 with a 4.70 ERA (77 ERA+). Jose DeLeon simply was never a good pitcher and his record, while obviously somewhat skewed by some bad luck in 1985, reflected that in all aspects. His peripherals were nothing impressive either, with a K/9 rate of just over 8.1 and a WHIP around 1.25 or so.

Johnny Mostil
06-13-2007, 10:33 AM
Wins may be a team stat first, but to completely exclude them from an evalution of a pitcher is simply stupid. If nothing else they show a pitcher's "clutch" factor - how they perform in game changing innings.

I've always been intrigued by Steve Carlton's '72 season: 27 wins for a team that won 59 games. How many innings did he hurl that were potentially "game changing"? (Every one, perhaps, if your team only wins 32 games without you?)

comerica
06-13-2007, 12:48 PM
I have two words for people about the worthlessness of win-loss records. Nolan Ryan.

chaerulez
06-13-2007, 02:01 PM
Well then bring back Hawk as GM. He's the one who traded Bobby Bonilla for a guy with great stats and a horrible W/L pct. I forget the pitcher's name but he never did live up to his stat hype.

Who knew Hawk was BB before the world knew who BB was?

Wins may be a team stat first, but to completely exclude them from an evalution of a pitcher is simply stupid. If nothing else they show a pitcher's "clutch" factor - how they perform in game changing innings. Of course there isn't much room for those kinds of evaluations in statistical analysis. It's one of the key reasons I don't buy a lot of these new stats. Baseball unlike most other major sports is a team game that is decided by a series of individual plays and when there are that many people doing their own thing it becomes very difficult to predict with any accuracy. I honestly believe you'd have to get deep into some very heavy math to really start to model it. Chaos theory comes immediately to mind...

Actually I once heard Hawk talk on the Score around 2004 about how he wanted Russ Ortiz because he had a good W-L record over the years. A W-L record gives a general idea of how good a pitcher is- most of the time. But it can be a silly stat. If a pitcher gives up a 1-0 lead in the 9th with 2 runs, yeah that's not very "clutch", but if they end up getting the loss it's pretty misleading. Such as is a pitcher who gets a win with a line of 5 IP and 4 ER. Don't get me wrong, most of the times an win or a loss is deserved, but sometimes they can be off. I hate Clemens, but he was a much better pitcher than his 13-8 and 7-6 records in his last two Houston seasons would indicate.

comerica
06-13-2007, 02:20 PM
as another example, look at Jeremy Bonderman this year. He's 6-0, but he really hasn't been great most of the time. He's been bailed out by the hitting several times and his record doesn't tell you about the 5 no decisions(3 or 4 of which we lost) that he had earlier in the year.

downstairs
06-13-2007, 02:50 PM
Any stathead worth his salt (many aren't) fully admits that their system is
not anywhere near perfect for predicting the future.

But what's wrong with trying to put together some numbers that may very well be better that what we have now? Don't like it, don't read it.

I'm a complete stat geek, especially the murky waters of predictive stats. Is it a religion to me? Heck no. Watching the game on the field is my religion.

As a fan and stat geek, I happen to really enjoy having some numbers I can look at that tend towards a team/player being better/worse than the typical numbers (wins, BA, homers, etc.) show.

Anyone that says "The Yankees are better because their runs scored are much more than their runs against" is a fool. However, stats like these and others do show that The Yankees should start heating up. Or better put, there's a greater chance of them heating up than a team with the same record that gives up a lot of runs and doesn't score many.

Maybe that means nothing to you. Cool. Because its not "right"... its just a different way of viewing the past and thinking about what will happen in the future (which we ALL do in some way).

jabrch
06-13-2007, 03:24 PM
A poster here, and I apologize as I don't remember who, recently said that it's one thing to use statistics to reflect what has happened, it is quite another to use them and try to extropolate out what will happen.

There are a lot of us who continue to say that. I have been saying it for a while.

You can predict what machines do - not what people do.

whitesoxfan1986
06-13-2007, 08:55 PM
Hey Daver-what is the exact definition of a propellerhead? Is it just a term for SABR geeks, or does it mean something more?

Daver
06-13-2007, 09:17 PM
Hey Daver-what is the exact definition of a propellerhead? Is it just a term for SABR geeks, or does it mean something more?

Google it, I believe the term was coined for programmers that specialized in creating new ways to run data.

MRM
06-13-2007, 11:07 PM
Well then bring back Hawk as GM. He's the one who traded Bobby Bonilla for a guy with great stats and a horrible W/L pct. I forget the pitcher's name but he never did live up to his stat hype.

Who knew Hawk was BB before the world knew who BB was?

Wins may be a team stat first, but to completely exclude them from an evalution of a pitcher is simply stupid.

W-L record is simply meaningless when evaluating a pitcher. It has so much to do with the offense and bullpen it's worthless for evaluation purposes.

Mark Buehrle 3-3 3.69
John Danks 3-6 4.38
Gil Meche 3-6 3.16
Matt Cain 2-6 3.31

Miguel Bautista 7-4 5.10
Adam Eaton 7-4 5.42
Bartolo Colon 5-2 5.70

Last Season

Aaron Cook 9-15 4.23
Mark Hendrickson 6-15 4.21
Jason Jennings 9-13 3.78

Randy Johnson 17-11 5.00
Josh Beckett 16-11 5.01

The first three played on teams with anemic offenses the last two on teams with awesome offenses.

voodoochile
06-13-2007, 11:28 PM
W-L record is simply meaningless when evaluating a pitcher. It has so much to do with the offense and bullpen it's worthless for evaluation purposes.

Mark Buehrle 3-3 3.69
John Danks 3-6 4.38
Gil Meche 3-6 3.16
Matt Cain 2-6 3.31

Miguel Bautista 7-4 5.10
Adam Eaton 7-4 5.42
Bartolo Colon 5-2 5.70

Last Season

Aaron Cook 9-15 4.23
Mark Hendrickson 6-15 4.21
Jason Jennings 9-13 3.78

Randy Johnson 17-11 5.00
Josh Beckett 16-11 5.01

The first three played on teams with anemic offenses the last two on teams with awesome offenses.

Sigh... Sample size...

Sure, over the course of a single season, it probably doesn't mean much, but over the course of several seasons, good pitchers will win more than they lose regardless of who they are playing for.

MRM
06-14-2007, 12:11 AM
Sigh... Sample size...

Sure, over the course of a single season, it probably doesn't mean much, but over the course of several seasons, good pitchers will win more than they lose regardless of who they are playing for.

Sigh...Tell it to these guys:

Pedro Ramos 1955-1970 117-160 4.08
Jack Fisher 1959-1969 86-139 4.06
Mike Morgan 1978-2002 141-186 4.23
Bob Friend 1951-1966 197-230 3.58
Ken Raffensberger 1939-1954 119-154 3.60
Rick Honeycutt 1977-1997 109-143 3.72
Charlie Hough 1970-1994 216-216 3.75

There are, of course, many many many more examples. Hell, Nolan Ryan was barely over .500 for his career. 324-292 3.19

voodoochile
06-14-2007, 01:23 AM
Sigh...Tell it to these guys:

Pedro Ramos 1955-1970 117-160 4.08
Jack Fisher 1959-1969 86-139 4.06
Mike Morgan 1978-2002 141-186 4.23
Bob Friend 1951-1966 197-230 3.58
Ken Raffensberger 1939-1954 119-154 3.60
Rick Honeycutt 1977-1997 109-143 3.72
Charlie Hough 1970-1994 216-216 3.75

There are, of course, many many many more examples. Hell, Nolan Ryan was barely over .500 for his career. 324-292 3.19

You say potato... I say potato... I don't see anything on that list to get worked up about. Solid but unspectacular. With the exception of Nolan, there isn't a #1 or really even a #2 to be found.

Heck, throwing up an ERA in the low 4's in the 60's wasn't exactly something to write home about. It was a notorious pitcher's era and team ERA's were regularly in the low 3's for winning teams. That's why they lowered the mound.

Perhaps the disagreement centers around the word "good" we probably just have different standards...