Manuel throws SABR-heads under bus

[FRENCHREDSOX]

I'm a computer science major at a tech-oriented university.

If a manager isn't using sabermetrics to help his team, he's completely shut out one entire hemisphere of data useful to predict performance in certain areas, situations, etc. It's foolish.

Very well said.

.

so a person who has played OR been around Baseball for 40 odd years,day in,day out (ie the Manager/Bench Coaches etc) is "foolish" because he doesn't use sabermetrics WHICH are notoriously:

1°Offense based
2° little or no Defense based
3° HAVE zero intangibles integrated

but uses his skills that he has honed.

Um would you allow a guy to run your University because he was a great business man in say the Oil Industry & was successful in making money or a "successful" Trader such as George Soros?

Which is EXACTLY what you are implying.........

I am not saying that sabermetrics have ZERO value but there VALUE is extremely limited & useful in only certain areas - as a Computer Science major you should realize better than anyone a theory is exactly that a THEORY,a LAW is however naturally correct.

Sabermetrics,in general,have multiple faults WHICH limit their usage & for a MLB team to use them in an "in game" situation would actually constitute foolishness as THEY only take into account CERTAIN TANGIBLE aspects & even then not all that are in "play" nor even in the actual situation .

Manuel,like all Managers' has multiple sources of INFORMATION & uses them in accordance - can Sabermetrics show if a player is hiding an injury ? Or if a player is out of sync ? Or simply if a batter is "seeing" the ball better ? No,No & NO.... thus there limits & even then most Sabermetric measures are inconsistent on year to year basis - simply put if they were CONSISTENT then they would be used as STANDARD MEASURING tools such as BA/OBP or ERA (although I find ERA+ a better equalizer measure).

[MeetDaMets]

so a person who has played OR been around Baseball for 40 odd years,day in,day out (ie the Manager/Bench Coaches etc) is "foolish" because he doesn't use sabermetrics WHICH are notoriously:

1°Offense based
2° little or no Defense based
3° HAVE zero intangibles integrated

but uses his skills that he has honed.

Um would you allow a guy to run your University because he was a great business man in say the Oil Industry & was successful in making money?

Which is EXACTLY what you are implying.........

bingo.

monday morning quarterbacking is one thing. we all do it.
but
the sheer volume of people that behave (maybe im mispercieving it?)
as if there book knowledge and little league experience makes them more "qualified" then an actual mlb manager is just baffling to me..

for those who didnt actually watch met games and just looked at overall #'s heres the deal
you all know the bullpen sucked
did you know
they DIDNT HIT LATE AND CLOSE ?
yes, even david wright.

do you know they outscored the opposition by a wide margin from 1 to 3 inning.
were about even the mid 3
and basically got their butts kicked over the last 3 innings.

do you seriously thinks thats random distribution ?
do you think 3 for 42 (over 6 weeks) with the bases loaded is anything but character ?
thats not performing in the clutch.
thats not knowing HOW to win.

having your bullpen throw 3 perfect innings after coming in with a 6-0 lead (and then tacking on 4 meaningless runs for a 10-0 win)
doesnt mean squat if
you routinely stop scoring after the 6th inning with a 1 run lead while your bullpen gives up 2 runs.

[kenny1234]

so a person who has played OR been around Baseball for 40 odd years,day in,day out (ie the Manager/Bench Coaches etc) is "foolish" because he doesn't use sabermetrics WHICH are notoriously:

1°Offense based
2° little or no Defense based
3° HAVE zero intangibles integrated

but uses his skills that he has honed.

I don't think anyone is suggesting that there is a statistical formula that will produce a winning team. If that is your argument - fine, you win. The question is whether the statistics are useful enough to contribute to a decision - and there is no question that is true. People have always used statistics in baseball to help make decisions - all sabermetrics does is attempt to measure the relevance of existing statistics and try to advance ones that are better.

bingo.

monday morning quarterbacking is one thing. we all do it.
but
the sheer volume of people that behave (maybe im mispercieving it?)
as if there book knowledge and little league experience makes them more "qualified" then an actual mlb manager is just baffling to me..

for those who didnt actually watch met games and just looked at overall #'s heres the deal
you all know the bullpen sucked
did you know
they DIDNT HIT LATE AND CLOSE ?
yes, even david wright.

do you know they outscored the opposition by a wide margin from 1 to 3 inning.
were about even the mid 3
and basically got their butts kicked over the last 3 innings.

do you seriously thinks thats random distribution ?
do you think 3 for 42 (over 6 weeks) with the bases loaded is anything but character ?
thats not performing in the clutch.
thats not knowing HOW to win.

having your bullpen throw 3 perfect innings after coming in with a 6-0 lead (and then tacking on 4 meaningless runs for a 10-0 win)
doesnt mean squat if
you routinely stop scoring after the 6th inning with a 1 run lead while your bullpen gives up 2 runs.

But how is Manuel going to change this? All teams are trying to find "clutch" players, all teams are trying to find that magical player that "knows how to win". Are these factors inherent in certain players? Is Manuel going to change the personnel to bring in "clutch" players? And if that was possible I'm pretty sure all the teams would be trying to do that. Or is he going to teach the players "how to win"? I guess this is possible - maybe that is what a good manager does. But if these are all the ideas that the Mets have - I'm guessing the Phillies might have a long run on top of the NL East.

And no, I don't think that the run scoring in the last three innings is random - the Mets bullpen sucked. If the character Manuel is looking for is better relievers then it might work.

And yes, I think 3 for 42 with the bases loaded might be random.

[FRENCHREDSOX]

I don't think anyone is suggesting that there is a statistical formula that will produce a winning team. If that is your argument - fine, you win. The question is whether the statistics are useful enough to contribute to a decision - and there is no question that is true. People have always used statistics in baseball to help make decisions - all sabermetrics does is attempt to measure the relevance of existing statistics and try to advance ones that are better.

Yes,but sabermetrics' are theoretical ANALYSES which don't hold up as yet UNLIKE other stats that is why they are MARGINAL stats - they fail to be constant over time (ie the formulae is constantly altered to "fit" the results eg Sabers of 1900/1950/2000 don't HAVE 1 constant whereas BA/OBP do have RELATIVE meaning which is the aim of statistical analysis) & as said are heavily weighted towards offensive performance & undervalue defensive qualities.

There is a major distinction between Statistics of the BA/OBP kind & the Sabermetric statistical theorems - the former are simply mathematical formulae of actual events WHICH are always constant in their result,whereas saberstats are a formulated theorem which supposedly add extra informative knowledge but are not standardised in their results - THAT is a major difference.In physics it is the difference between a LAW & a possible hypothesis.....

Also I was specifically answering this comment:

If a manager isn't using sabermetrics to help his team, he's completely shut out one entire hemisphere of data useful to predict performance in certain areas, situations, etc. It's foolish.

& I ascertained that Sabermetrics actually MAY misinform because of their weights & there lack of intangibles,& basically the data RESULTS may or may not be inconclusive (at best) or wrong (at worst) which have as great implication in actual game CHOICES made.

Sabermetrics,like all stats has it uses BUT only as a subjective/hindsight tool within certain constraints - BorgHunter was implying that it is a predictive tool which it definitely is not & WOULD fail due to its bias.


Its main presumptions are exactly that presumptions which don't hold up in the real world - a common saber stat is for example VORP which uses the base line of Replacement Player being a AAA "zoner" & does NOT even consider his defense whereas IRL no team WOULD bring up:mind play a "deficit" player thus the baseline is in context redundant.It is a SHORT term stat (ie no analysis pre-1997) & all players hit the same .235/.300/.356 and have a RA of 5.85 but those numbers are not necessarily true/constant or even correct as a BASIS.

Again I re-iterate Saberstats are useful as a complementary tool but WITHIN a defined context of relative value & definitely not as a tool for manager-ship decision making IF anything the eye & team information is better than the numbers.

For example Win Shares are,as cited by Bill James' himself :

Last year I heard at the SABR convention that you were modifying the formula for Win Shares. Is that true?
Bill James reply;
I've been working on Win Shares and Loss Shares. There are some things about it that are fouled up. . .about the way that I am figuring them. . .and I can't get the time to back off and straighten them out.

Even Sabermetricians counter EACH others' analyses & formulae,so who is right James,Woolner,Pete Palmer,McCracken or some other guy ? Because they all have different formulae & also differing results & outcomes....


So why would a manager USE them & RISK the game/series/season on an analysis that IS NOT constant nor predictive ? IMHO,it would be foolish to do so rather than the opposite,no ?

[haveacigar]

Whatever you may or may not think of SABRmetrics, I still don't see what the hell SABRmetrics has to do with the Mets collapse. It just sounds like Manuel is scapegoating something to cover his ass.

[BorgHunter]

Sabermetrics,in general,have multiple faults WHICH limit their usage & for a MLB team to use them in an "in game" situation would actually constitute foolishness as THEY only take into account CERTAIN TANGIBLE aspects & even then not all that are in "play" nor even in the actual situation .

Manuel,like all Managers' has multiple sources of INFORMATION & uses them in accordance - can Sabermetrics show if a player is hiding an injury ? Or if a player is out of sync ? Or simply if a batter is "seeing" the ball better ? No,No & NO.... thus there limits & even then most Sabermetric measures are inconsistent on year to year basis - simply put if they were CONSISTENT then they would be used as STANDARD MEASURING tools such as BA/OBP or ERA (although I find ERA+ a better equalizer measure).

If you would bother to read my post, you'd find that I said exactly this. Sabermetrics is a tool, not the end all, be all of baseball knowledge. The fact that it doesn't have ultimate predictive power by itself is not an indictment of sabermetrics as a predictive tool—it only demonstrates that you should not use statistical analysis in a vacuum. And I don't think that anyone here has claimed that you should.

Sabermetrics,like all stats has it uses BUT only as a subjective/hindsight tool within certain constraints - BorgHunter was implying that it is a predictive tool which it definitely is not & WOULD fail due to its bias.Its main presumptions are exactly that presumptions which don't hold up in the real world - a common saber stat is for example VORP which uses the base line of Replacement Player being a AAA "zoner" & does NOT even consider his defense whereas IRL no team WOULD bring up:mind play a "deficit" player thus the baseline is in context redundant.It is a SHORT term stat (ie no analysis pre-1997) & all players hit the same .235/.300/.356 and have a RA of 5.85 but those numbers are not necessarily true/constant or even correct as a BASIS.

VORP is an offensive stat, not a defensive stat. If you want defensive stats, look at Range Factor. VORP isn't intended to be used as the only thing you look at when evaluating a player. It's a measure of offensive success only—no one has claimed otherwise. It's like claiming that the amount of home runs a pitcher has hit is useless because it doesn't take into account his pitching. Of course it doesn't take into account his pitching. It was never intended to! But it certainly could help you figure out if you want to pinch hit for this pitcher, yes? (Among other stats, plus, as a you said, things not taken into account by stats.)

All in all, it seems like you're tilting at statistical windmills and inventing opinions for you to argue against. Sabermetrics do have useful predictive power, no matter what you say. But no one has ever said that they're the only damned thing that matters.

[RedsoxRockies]

Borg Hunter, I have to ask, what is a Borg?

[BorgHunter]

Borg Hunter, I have to ask, what is a Borg?

I take it you've never watched Star Trek: The Next Generation?

[FRENCHREDSOX]

Borg Collective I assume ....

[TheNamelessPoet]

there should be noone throwing anyone under a bus its mean and im sure some type of assault

[HoustonGM]

could you be a little more melodramatic ?

surely you dont believe managers have nothing else to do
then play "someone is wrong on the internet"

Manuel didn't use that strawman argument. It was brought up by someone else and I'm speaking about it in general terms, not specifically regarding the original Manuel quote.

so a person who has played OR been around Baseball for 40 odd years,day in,day out (ie the Manager/Bench Coaches etc) is "foolish" because he doesn't use sabermetrics

Anybody who doesn't acknowledge the value of statistical analysis is being foolish, yes. Likewise, anybody who doesn't acknowledge the value of analysis through non-statistical means like scouting is also being foolish.

WHICH are notoriously:

1°Offense based
2° little or no Defense based

This is completely, COMPLETELY false. Sabermetrics is just a term for "statistical analysis." Yes, the analysis of offense is more advanced and concrete than the analysis of defense, but to say that "Sabermetrics are little or no Defense based" is UTTERLY WRONG. There is A TON of work being done on defense within the statistical analysis community. Zone Rating, RZR, Range Factor, Revised Range Factor, FRAA, UZR, +/-, DRS, Fielding Win Shares, etc. etc. etc. etc.

3° HAVE zero intangibles integrated

Gee, I wonder why. It wouldn't have anything to do with intangibles being...I don't know...INTANGIBLE?

Um would you allow a guy to run your University because he was a great business man in say the Oil Industry & was successful in making money or a "successful" Trader such as George Soros?

Which is EXACTLY what you are implying.........

Um no, this analogy doesn't fit at all. The goal of a university is to educate, not make money, so no, I wouldn't have a person good at making money and that's it run by university. But that has nothing to do with baseball. Baseball is about winning, and in order to maximize winning, you need to understand the value of both statistical and non-statistical value.

I am not saying that sabermetrics have ZERO value but there VALUE is extremely limited & useful in only certain areas

No. Sabermetrics are not extremely limited. Again, sabermetrics is just simply statistical analysis. Analyzing baseball statistically IS incredibly useful and FAR from extremely limited.

Sabermetrics,in general,have multiple faults WHICH limit their usage & for a MLB team to use them in an "in game" situation would actually constitute foolishness as THEY only take into account CERTAIN TANGIBLE aspects & even then not all that are in "play" nor even in the actual situation .

Wow. So, using statistics in an in-game situation is foolish? Incredible.

Here's a hint: Nobody's saying that statistics should be the only thing used. We're saying they should be CONSIDERED and FACTORED INTO DECISIONS. Yes, there are non-statistical things that need to be considered as well, but using ONLY those is just as foolish as using ONLY stats.

And I'm repeating myself again. Sabermetrics = statistical analysis. By acknowledging the value of on-base percentage from your leadoff hitter, you're "using sabermetrics."

Manuel,like all Managers' has multiple sources of INFORMATION & uses them in accordance - can Sabermetrics show if a player is hiding an injury ? Or if a player is out of sync ? Or simply if a batter is "seeing" the ball better ? No,No & NO.... thus there limits

Um, yeah, show me again where anybody said that sabermetrics are the only thing that people should use? Once again, a strawman.

There are many things that statistics tell you that you wouldn't know otherwise, just like there are things that statistics can't tell you. This is why both statistical and non-statistical methods of analysis should be used. They COMPLIMENT EACH OTHER. By ignoring statistics, you're handicapping yourself. By ignoring the non-statistical factors, you're handicapping yourself.

& even then most Sabermetric measures are inconsistent on year to year basis

There is more to sabermetrics than advanced metrics. Sabermetrics is just statistical analysis.

- simply put if they were CONSISTENT then they would be used as STANDARD MEASURING tools such as BA/OBP or ERA (although I find ERA+ a better equalizer measure).

By using batting average, on-base percentage, and ERA, you are using sabermetrics. Sabermetrics is statistical analysis. Now, using only those things, is making incredibly poor use of statistics, but it's STILL USING STATISTICS.

Also, most sabermetric measures are more consistent from year-to-year than batting average and ERA. That's why they were created - because people realized the great variance that average and ERA exhibit and sought out better measures that are less prone to fluctutaion.

do you seriously thinks thats random distribution ?
do you think 3 for 42 (over 6 weeks) with the bases loaded is anything but character ?

Yes. 42 at bats is too small a sample size to draw any conclusions from. It's especially too small a sample size to draw a conclusion about an ENTIRE TEAM from.

I don't think anyone is suggesting that there is a statistical formula that will produce a winning team. If that is your argument - fine, you win. The question is whether the statistics are useful enough to contribute to a decision - and there is no question that is true. People have always used statistics in baseball to help make decisions - all sabermetrics does is attempt to measure the relevance of existing statistics and try to advance ones that are better.

THANK YOU. It's not that hard to understand.

Yes,but sabermetrics' are theoretical ANALYSES which don't hold up as yet UNLIKE other stats that is why they are MARGINAL stats - they fail to be constant over time (ie the formulae is constantly altered to "fit" the results eg Sabers of 1900/1950/2000 don't HAVE 1 constant whereas BA/OBP do have RELATIVE meaning which is the aim of statistical analysis) & as said are heavily weighted towards offensive performance & undervalue defensive qualities.

This is a very convoluted statement that I really cannot even begin to understand, except that it really shows an ignorance of sabermetrics. ONCE AGAIN, sabermetrics = statistical analysis. Sabermetrics included theoretical analysis AND non-theoretical analysis. Any analysis using statistics is sabermetrics.

& I ascertained that Sabermetrics actually MAY misinform because of their weights & there lack of intangibles which have as great implication in actual game CHOICES made.

which is why they shouldn't be the only thing looked at. This is not a valid reason to completely ignore statistics.

[HoustonGM]

Sabermetrics,like all stats

And herein lies what I've been trying to say this entire post. There is NO difference between "sabermetrics" and "other stats." Sabermetrics is simply the analysis of statistics. Wikipedia says it perfectly: "Sabermetrics is the analysis of baseball through objective evidence." THAT'S IT.

BUT only as a subjective/hindsight tool within certain constraints - BorgHunter was implying that it is a predictive tool which it definitely is not & WOULD fail due to its bias

Are you friggen' kidding me? Seriously? Statistics have no use as a predictive tool? Wow.

Guess what? You know PECOTA? And ZiPS? That's sabermetrics buddy.

Its main presumptions are exactly that presumptions which don't hold up in the real world

This is an incorrect generalization. Once again, this may be true for certain statistics and/or areas of statistical analysis, but it absolutely 100% cannot be applied to statistical analysis as a whole.

a common saber stat is for example VORP which uses the base line of Replacement Player being a AAA "zoner" & does NOT even consider his defense

Yeah, because, it's an offensive stat. I guess OPS sucks too because it doesn't consider defense.

whereas IRL no team WOULD bring up:mind play a "deficit" player

Really? SERIOUSLY? In real life, no team calls up and play players that have a negative effect? You cannot possibly belief this. You're smarter than that.

Jose Vidro's 65 OPS+ as a DH wasn't a "deficit player"?

.It is a SHORT term stat (ie no analysis pre-1997)

That's funny, because VORP goes back to 1957.

all players hit the same .235/.300/.356 and have a RA of 5.85 but those numbers are not necessarily true/constant or even correct as a BASIS.

That's because this is not how VORP is calculated. The replacement level is different depending on the league. It's not a constant.

Again I re-iterate Saberstats are useful as a complementary tool but WITHIN a defined context of relative value & definately not as a tool for manager-ship decision making IF anything the eye & team information is better than the numbers.

You're failing to understand that the use of sabermetrics and statistical analysis consists of more than just the advanced performance metrics.

"Eye and team information" is not better than the numbers. Both types of information have their uses. Both types of information are also useless without the other.

So why would a manager USE them & RISK the game/series/season on an analysis that IS NOT constant nor predictive ? IMHO,it would be foolish to do so rather than the opposite,no ?

Win Shares are a stat to assess the value that a player has provided. It's useful in comparing the value that players have provided. Win Shares are a long-term stat, best used to compare full seasons and/or careers.

This is not a reason to ignore statistics when making a managerial decision.

VORP is an offensive stat, not a defensive stat. If you want defensive stats, look at Range Factor. VORP isn't intended to be used as the only thing you look at when evaluating a player. It's a measure of offensive success only—no one has claimed otherwise. It's like claiming that the amount of home runs a pitcher has hit is useless because it doesn't take into account his pitching. Of course it doesn't take into account his pitching. It was never intended to!

Seriously. This VORP argument pops up every once in a while, and the "anti-VORP" people still cling to this idea that people use VORP as a be-all-and-end-all stat that is the ultimate stat and perfect for everything...when all it is is a stat to assess offensive value.

I'm still completely baffled by how people can't understand this.

All in all, it seems like you're tilting at statistical windmills and inventing opinions for you to argue against. Sabermetrics do have useful predictive power, no matter what you say. But no one has ever said that they're the only damned thing that matters.

Exactly.

There seems to be a fundamental lack of understanding of what sabermetrics is. People are focusing on advanced metrics that measure performance, and that's only PART of sabermetrics, and they're speaking like that's the only thing sabermetrics consist of, when in reality, sabermetrics is a vast science, and encompasses ALL analysis of statistics.

[FRENCHREDSOX]

If you would bother to read my post, you'd find that I said exactly this.

I did read your post entirely & you clearly said that a manager would be follish NOT to use SABRmetric formulae/data & predictive analysis.I was simply answering you on THIS particular point

Sabermetrics is a tool, not the end all, be all of baseball knowledge. The fact that it doesn't have ultimate predictive power by itself is not an indictment of sabermetrics as a predictive tool—it only demonstrates that you should not use statistical analysis in a vacuum. And I don't think that anyone here has claimed that you should.

VORP is an offensive stat, not a defensive stat. If you want defensive stats, look at Range Factor. VORP isn't intended to be used as the only thing you look at when evaluating a player. It's a measure of offensive success only—no one has claimed otherwise. It's like claiming that the amount of home runs a pitcher has hit is useless because it doesn't take into account his pitching. Of course it doesn't take into account his pitching. It was never intended to! But it certainly could help you figure out if you want to pinch hit for this pitcher, yes? (Among other stats, plus, as a you said, things not taken into account by stats.)

All in all, it seems like you're tilting at statistical windmills and inventing opinions for you to argue against. Sabermetrics do have useful predictive power, no matter what you say. But no one has ever said that they're the only damned thing that matters.

Again I said,I think - that certain Sabermetrics as they stand are a "false" metric because of their presumptions & bias.

You believe they have a "predictive" value & that is your right but again I simply ask which SABR stats (as many contradict one another's results & even BASE line formulae) & why,unlike "classical" statistical analysis do their results not hold up over time leading to their creators continually re-altering their theses.

I do find,some Saberstats AS USEFUL COMPLEMETARY tools (such as DIPS,Pythagorean Expectation,FRAR or OPS) whereas others,which are often cited such as VORP,LIPS or Win Shares,can & are "poorly" constructed formulae.

That was all I was saying - no disrespect intended & it is likely Manuel was saying something similar concerning intangibles THAT all STATISTICAL analyses cannot,for whatever their worth,integrate into the decision making process.

[michaelg123789]

Interesting

[boomboom]

Mikey, We need to get BMO back on track...especially IL30 which only has 22 owners.