Chicago Fanatics Message Board
https://mail.chicagofanatics.com/

When advanced Metrics go wrong
https://mail.chicagofanatics.com/viewtopic.php?f=24&t=87339
Page 1 of 2

Author:  conns7901 [ Tue Jun 10, 2014 11:27 pm ]
Post subject:  When advanced Metrics go wrong

Scott in Davenport tweeted this on Bob Welch.
Quote:

Scott Lindholm @ScottLindholm
#MLB @dan_bernstein hipped me to this--two different Bob Welch seasons, traditional metrics vs. advanced


Image

Author:  Nas [ Tue Jun 10, 2014 11:35 pm ]
Post subject:  Re: When advanced Metrics go wrong

That's awful.

Author:  Hussra [ Tue Jun 10, 2014 11:44 pm ]
Post subject:  Re: When advanced Metrics go wrong

conns7901 wrote:
Scott in Davenport tweeted this on Bob Welch.
Quote:

Scott Lindholm @ScottLindholm
#MLB @dan_bernstein hipped me to this--two different Bob Welch seasons, traditional metrics vs. advanced


Image



1990 A's were 4th in MLB in runs scored.

1986 Dodgers were 4th from last in MLB in runs scored.

Image

Author:  Rod [ Wed Jun 11, 2014 5:55 am ]
Post subject:  Re: When advanced Metrics go wrong

Hussra wrote:
conns7901 wrote:
Scott in Davenport tweeted this on Bob Welch.
Quote:

Scott Lindholm @ScottLindholm
#MLB @dan_bernstein hipped me to this--two different Bob Welch seasons, traditional metrics vs. advanced


Image



1990 A's were 4th in MLB in runs scored.

1986 Dodgers were 4th from last in MLB in runs scored.

Image


I don't care if the A's scored 200 runs per game. If you really think that Dodger season was better, you're baseball retarded. It's like arguing that black is white. And getting arrogant about it.

Author:  Bucky Chris [ Wed Jun 11, 2014 6:31 am ]
Post subject:  Re: When advanced Metrics go wrong

JORR, you are literally arrogant about everything though.

Author:  Rod [ Wed Jun 11, 2014 6:39 am ]
Post subject:  Re: When advanced Metrics go wrong

Bucky Chris wrote:
JORR, you are literally arrogant about everything though.


Well, I guess that's for others to say. I'm sorry you feel that way. I'm sure there are others who disagree. I try to stay away from topics I know nothing about.

Anyway, back to the seasons of Bob Welch, you can see that the real issue here with "wins"* above replacement is that pitchers are getting too much credit for strikeouts and for limiting home runs. Those are the bases for any claims that the Dodger season was superior.



* I find it endlessly fascinating that the same people who get so angry about the semantics of a pitcher's "win" have no problem embracing the concept of a "win" or a partial "win" even in instances where the team lost.

Author:  Hussra [ Wed Jun 11, 2014 6:54 am ]
Post subject:  Re: When advanced Metrics go wrong

Hmm, the difference btw Fangraphs and Baseball Reference on WAR might support Bob Ryan's War on WAR:


Bob Welch WAR (pitching value):

FANGRAPHS
========

1986 5.3
1987 4.2
1990 1.6

BASEBALL REFERENCE
=============

1986 4.0
1987 7.1 (8th in cy young voting that year)
1990 3.0 (Won Cy Young, All-Star team, MVP-9)

Bob Ryan's War on WAR:
viewtopic.php?f=92&t=87233

Author:  rogers park bryan [ Wed Jun 11, 2014 7:36 am ]
Post subject:  Re: When advanced Metrics go wrong

Its important to remember that advanced statistics are built upon predicting future performance, not telling you what already happened.

Author:  good dolphin [ Wed Jun 11, 2014 7:48 am ]
Post subject:  Re: When advanced Metrics go wrong

Walk, strikeout and home run ratios are now considered "advanced"? That has been around and used throughout my life.

I think I have asked this before and I certainly cannot be the first to have asked this but if a strikeout is equal to any other out from a batting perspective the equation must balance that the same is true from a pitching perspective.

Author:  Rod [ Wed Jun 11, 2014 7:51 am ]
Post subject:  Re: When advanced Metrics go wrong

good dolphin wrote:
Walk, strikeout and home run ratios are now considered "advanced"? That has been around and used throughout my life.

I think I have asked this before and I certainly cannot be the first to have asked this but if a strikeout is equal to any other out from a batting perspective the equation must balance that the same is true from a pitching perspective.


What's advanced is the concept that anything besides strikeouts, walks, and homers is out of the pitcher's hands and merely a function of luck.

Author:  badrogue17 [ Wed Jun 11, 2014 8:01 am ]
Post subject:  Re: When advanced Metrics go wrong

Joe Orr Road Rod wrote:

I don't care if the A's scored 200 runs per game. If you really think that Dodger season was better, you're baseball retarded. It's like arguing that black is white. And getting arrogant about it.

That's pretty much it in a nutshell.

Author:  312player [ Wed Jun 11, 2014 8:54 am ]
Post subject:  Re: When advanced Metrics go wrong

I have never liked war ,this is a great example of how flawed it is.

Author:  Douchebag [ Wed Jun 11, 2014 8:55 am ]
Post subject:  Re: When advanced Metrics go wrong

312player wrote:
I have never liked war

What is it good for?

Author:  312player [ Wed Jun 11, 2014 8:58 am ]
Post subject:  Re: When advanced Metrics go wrong

:lol:

Author:  Brick [ Wed Jun 11, 2014 8:59 am ]
Post subject:  Re: When advanced Metrics go wrong

conns7901 wrote:
Scott in Davenport tweeted this on Bob Welch.
Quote:

Scott Lindholm @ScottLindholm
#MLB @dan_bernstein hipped me to this--two different Bob Welch seasons, traditional metrics vs. advanced


Image
Sabrmetrics is a myth.

Author:  rogers park bryan [ Wed Jun 11, 2014 9:04 am ]
Post subject:  Re: When advanced Metrics go wrong

good dolphin wrote:

I think I have asked this before and I certainly cannot be the first to have asked this but if a strikeout is equal to any other out from a batting perspective the equation must balance that the same is true from a pitching perspective.

It's about the single player and what is and isnt in their control. It's designed to predict future performance, not tell you what happened in the past.

Same thing with the Runs Scored/Runs against. A team might win a bunch of one run games and that's fine. They won those games. They count. But the team that is outscoring opponents like crazy is more likely to win more in the future.

But yes, like all stats they are flawed


The HUGE problem with WAR is the balance between Off and Def which is why you have guys like Zobrist and Lawrie on top ten lists

Author:  good dolphin [ Wed Jun 11, 2014 9:24 am ]
Post subject:  Re: When advanced Metrics go wrong

rogers park bryan wrote:
good dolphin wrote:

I think I have asked this before and I certainly cannot be the first to have asked this but if a strikeout is equal to any other out from a batting perspective the equation must balance that the same is true from a pitching perspective.

It's about the single player and what is and isnt in their control. It's designed to predict future performance, not tell you what happened in the past.

Same thing with the Runs Scored/Runs against. A team might win a bunch of one run games and that's fine. They won those games. They count. But the team that is outscoring opponents like crazy is more likely to win more in the future.

But yes, like all stats they are flawed


The HUGE problem with WAR is the balance between Off and Def which is why you have guys like Zobrist and Lawrie on top ten lists


I'm just focused on this one, compartmentalized problem. It MUST balance mathematically and logically.

Author:  rogers park bryan [ Wed Jun 11, 2014 9:30 am ]
Post subject:  Re: When advanced Metrics go wrong

good dolphin wrote:
rogers park bryan wrote:
good dolphin wrote:

I think I have asked this before and I certainly cannot be the first to have asked this but if a strikeout is equal to any other out from a batting perspective the equation must balance that the same is true from a pitching perspective.

It's about the single player and what is and isnt in their control. It's designed to predict future performance, not tell you what happened in the past.

Same thing with the Runs Scored/Runs against. A team might win a bunch of one run games and that's fine. They won those games. They count. But the team that is outscoring opponents like crazy is more likely to win more in the future.

But yes, like all stats they are flawed


The HUGE problem with WAR is the balance between Off and Def which is why you have guys like Zobrist and Lawrie on top ten lists


I'm just focused on this one, compartmentalized problem. It MUST balance mathematically and logically.

You're wrong. The numbers are what they are.

It's two completely different situations. One is a hitter trying to get on base. One is a pitcher trying to record an out. Certain things indicate future success at each activity.

It's not equal on both sides when using numbers to predict the future.

If you're talking about what happened, then yes the strikeout is just another out. That's the history. But these are predictive numbers.

4 strikeouts yesterday is a bad game but it's not any worse than 4 flyouts when predicting future performance

Buehrle might throw a perfect game with 0 Strikeouts. Thats fine and its a great game. A guy who gives up 2 soft hits and strikes out 14 is more likely to be better going forward.

Author:  conns7901 [ Wed Jun 11, 2014 9:40 am ]
Post subject:  Re: When advanced Metrics go wrong

Maybe I am not looking at the WAR stat correctly as I am not that into Sabremetrics:

My first thought when looking at that stat last night is the Average replacement player would of been expected to only get two wins for an entire season on a 73 win team.

Then on the 90 A's team he would of expected to get 25 wins. Both scenarios seem ridiculous.

Author:  Hatchetman [ Wed Jun 11, 2014 10:06 am ]
Post subject:  Re: When advanced Metrics go wrong

Welch forgot how to be a winner in 86. then he remembered again. Then forgot again. :lol:

Author:  Rod [ Wed Jun 11, 2014 10:22 am ]
Post subject:  Re: When advanced Metrics go wrong

Hatchetman wrote:
Welch forgot how to be a winner in 86. then he remembered again. Then forgot again. :lol:



Why is that any funnier than thinking he forgot how to strike guys out? He had a bad year in '86. It resulted in a losing record. The runs his team scored are just part of his circumstances. Every pitcher has different circumstances. It would be silly to say that Juan Marichal isn't as good as he seems because he had Willie patrolling center field while poor Fergie had Adolfo Phillips. Every guy has to deal with different shit.

Like RPB said, the SABRmetric stats are predictive. But what they really come down to is measuring tools rather than game management. Sure, a guy who has the ability to strike a bunch of guys out has an advantage over a guy who can't. But there's a whole lot more to it. I can give you all kinds of state of the art equipment and I doubt you can build a house as well as Bob Vila can with just a shovel, hammer, nails, and a handsaw. The SABR guy considers you "better" based upon your fucking bulldozer.

Author:  Don Tiny [ Wed Jun 11, 2014 10:26 am ]
Post subject:  Re: When advanced Metrics go wrong

Joe Orr Road Rod wrote:
If you really think that Dodger season was better, you're baseball retarded.


Are you able to locate one person in the world that has made such a statement? Not strawman conjecture, I mean someone who has actually tried to argue that point?

Author:  Rod [ Wed Jun 11, 2014 10:29 am ]
Post subject:  Re: When advanced Metrics go wrong

Don Tiny wrote:
Joe Orr Road Rod wrote:
If you really think that Dodger season was better, you're baseball retarded.


Are you able to locate one person in the world that has made such a statement? Not strawman conjecture, I mean someone who has actually tried to argue that point?


Isn't that the reason for Scott in Davenport bringing it up? To show that his record is "meaningless" and his actual WAR (which is what counts) was far superior in the 1986 season?

Author:  Hatchetman [ Wed Jun 11, 2014 10:33 am ]
Post subject:  Re: When advanced Metrics go wrong

you don't forget how to strike guys out. you get old or you get hurt. you don't forget how to win either.

Author:  Don Tiny [ Wed Jun 11, 2014 10:37 am ]
Post subject:  Re: When advanced Metrics go wrong

Joe Orr Road Rod wrote:
Don Tiny wrote:
Joe Orr Road Rod wrote:
If you really think that Dodger season was better, you're baseball retarded.


Are you able to locate one person in the world that has made such a statement? Not strawman conjecture, I mean someone who has actually tried to argue that point?


Isn't that the reason for Scott in Davenport bringing it up? To show that his record is "meaningless" and his actual WAR (which is what counts) was far superior in the 1986 season?


I presumed he brought it up to showcase a rather abberant anomaly in regards to the WAR statistic itself. I don't have the context for it, so it's only a guess.

Personally, I also read it as something interesting shown by the anomaly ... that Welch's shit season with the Dodgers apparently happened during a shitty pitching season overall as his efforts, dreadful as they were, would seem to not be counted amongst the absolute worst that season (and that season alone). Conversely, I took it to mean that his inarguably sensational season w/ the A's came during a season of, apparently, a numberous array of sensational pitching performances.

Author:  Rod [ Wed Jun 11, 2014 10:38 am ]
Post subject:  Re: When advanced Metrics go wrong

Hatchetman wrote:
you don't forget how to strike guys out. you get old or you get hurt. you don't forget how to win either.



Or you figure out how to manage a game. Part of the reason that guys are only getting six innings is that they are going so long in counts. This is due in large part to an offensive philosophy that stresses seeing pitches as a value that trumps all else. It's also resulting in a huge amount of strikeouts. I don't think Welch lacked his wipeout pitch in his best season. He figured out how to get outs in different ways without running up his pitch count.

Even before pitches were counted, pitch counts were significant. Not necessarily due to "Pitcher Abuse Points" but because the more pitches a batter sees the more likely he is to get a hit.

Author:  Hatchetman [ Wed Jun 11, 2014 10:40 am ]
Post subject:  Re: When advanced Metrics go wrong

well I disagree that pitcher wins mean much, but if those greats of yore pitched 200 innings instead of 300, they'd have put up unbelievable numbers. Fergie'd have probably K'd 12 per 9.

Author:  Rod [ Wed Jun 11, 2014 10:42 am ]
Post subject:  Re: When advanced Metrics go wrong

Don Tiny wrote:
Joe Orr Road Rod wrote:
Don Tiny wrote:
Joe Orr Road Rod wrote:
If you really think that Dodger season was better, you're baseball retarded.


Are you able to locate one person in the world that has made such a statement? Not strawman conjecture, I mean someone who has actually tried to argue that point?


Isn't that the reason for Scott in Davenport bringing it up? To show that his record is "meaningless" and his actual WAR (which is what counts) was far superior in the 1986 season?


I presumed he brought it up to showcase a rather abberant anomaly in regards to the WAR statistic itself. I don't have the context for it, so it's only a guess.

Personally, I also read it as something interesting shown by the anomaly ... that Welch's shit season with the Dodgers apparently happened during a shitty pitching season overall as his efforts, dreadful as they were, would seem to not be counted amongst the absolute worst that season (and that season alone). Conversely, I took it to mean that his inarguably sensational season w/ the A's came during a season of, apparently, a numberous array of sensational pitching performances.


There are so many different versions of WAR, I'm not sure if any or all of them confine the average or model for the "replacement player" to a single season. If they do, people certainly shouldn't be using them across different seasons to declare one guy better than another.

Anyway, with my knowledge of bernstein's war on pitchers' W/L record, I think it is safe to assume that this was nothing more than an attempt by Scott to curry favor with his master.

Author:  Don Tiny [ Wed Jun 11, 2014 10:53 am ]
Post subject:  Re: When advanced Metrics go wrong

Joe Orr Road Rod wrote:
There are so many different versions of WAR, I'm not sure if any or all of them confine the average or model for the "replacement player" to a single season. If they do, people certainly shouldn't be using them across different seasons to declare one guy better than another.

Anyway, with my knowledge of bernstein's war on pitchers' W/L record, I think it is safe to assume that this was nothing more than an attempt by Scott to curry favor with his master.


I think we basically agree on WAR ... I would say it's use is limited to 2, maybe 3 seasons minus the current and maybe as a non-significant part of predicting the next year's effort after the current one is at least 2/3 done. Anyone who would tout it as a "superstat" that can stand on its own without context or without any other comparative stat - old or new - is clearly a snake-oil salesman.

That's not some sideways swipe at Dan ... I don't think he's ever remotely tried to indicate WAR is be-all, end-all (I think that's a case of some people having their imaginary radios on and confusing being an arrogant ass with being a run-of-the-mill dope).

That said, I cannot envision a scenario wherein I would put more importance on W/L than WAR anymore, but it's not like I won't at least look at W/L as part and parcel of some larger package. But I cannot fathom how anyone - and I'm not indicating that you do - but I cannot fathom how anyone can every pretend to think W/L means a tenth of what it did, say, 40 years ago. The function, handling, and expectations of a starting pitcher - or really any pitcher - is simply not the same as it was then, so to expect a stat almost as old as the game itself that has not evolved along with the position to remain as important and informative as it once was is ludicrous.

But again, I think we agree enough that we're not standing at different podiums on the issue, just different shades of it .... easily discernable shades to be sure, but just different, not dissimilar.

Author:  Brick [ Wed Jun 11, 2014 10:55 am ]
Post subject:  Re: When advanced Metrics go wrong

I laugh whenever I read "SABRmetrics is only meant to be predictive".

If that is the case, why is it ever brought up in any baseball discussion about current or past performance?

I'm starting to think my "SABRmetrics is a myth" stance is correct.

Page 1 of 2 All times are UTC - 6 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/