PDA

View Full Version : HE SAID WHAT? (Sheehan's excuse for why we are an 82 win team)


jabrch
10-29-2005, 11:06 AM
http://www.baseballprospectus.com/article.php?articleid=4571

Germany demilitarized, start another World War, and then demilitarized again. The computer was invented - digital electronics itself. Man went to the moon, and now doesn’t do that anymore.

Interestingly, the White Sox have won more WS in the last 10 years than all of the Sabremetricly managed teams with under a $125mm annual salary budget have COMBINED. That's a bit more relevant than your story about Germany.

To celebrate Chicago’s World Series sweep, we’re looking at our projections that showed the Sox as a .500 team at the beginning of the season. How’d we go wrong?

I'd say it is because you never actually WATCH baseball games and can't tell what is going on in the field of play, but I'll wait for your poor analysis.

A.J. Pierzynski (file:///dt/pierzaj01.shtml)

Pierzynski underperformed his VORP projection

Nearly every pitcher on this staff pitched better this year than ever before. Tell me where that's factored into your VORP? Not gonna give AJ any credit for that? Yet somehow we should go ga-ga over Varitek while his pitchers underperformed their career averages. Consistency is not your strong suit, huh? I've told you 1000+ times Joe, there are too many variables to the game for you to try and stuff it into one convenient number - yet you keep trying (and failing)

Paul Konerko (file:///dt/konerpa01.shtml)
PA HR RBI AVG/OBP/SLG VORP
Actual 664 40 100 .283/.375/.534 56.3
PECOTA 641 31 103 .276/.353/.494 30.8

The first of two serious underestimations on offense, Konerko had a big walk year, contributing roughly three wins more than PECOTA expected of him. Even more dramatic is the difference between Konerko’s projected and actual defensive numbers

"PECOTA pegged Konerko"
First of all, PECOTA didn't PEG anything. PEG is an action verb. PECOTA took no action, did nothing, and is nothing other than a poorly defined set of assumptions built into a childishly simple model. Your analysts PEGGED Konerko - VERY POORLY. Don't go blaming this mythical PECOTA as if it were a person and it erred. It is your weak system.

Second, PECOTA is a wild ass guess with a very low rate of being accurate for any player who doesn't perform within 10% of their 3 year trend line. That's not quality analysis, that's basic stats 101.
Tadahito Iguchi (file:///dt/iguchta01.shtml)

Actual 582 15 71 .278/.335/.438 30.9
PECOTA 591 11 61 .277/.344/.408 32.3

Iguchi’s defensive projection was less studly--PECOTA projected Iguchi as an above-average defender, with a Rate of 102.6. Reality presented Iguchi as one of the worst defensive second basemen in the game, with a Rate of 89.

Well this is just stupid. Again, your model is screwed up if the output of it is a statement that Iguchi is one of the worst defensive 2B in the game. Everyone who actually watches him play says otherwise. Your twisted version of reality must never have seen Iguchi turn the double play, or make plays going to his left. The Sox IF defense was so solid up the middle and Iguchi was a key part to it. You really need to get out some - and actually watch a baseball game so you can understand what your models are telling you.
Joe Crede (file:///dt/credejo01.shtml)

Actual 471 22 62 .252/.301/.453 13.8
PECOTA 585 25 77 .266/.321/.459 13.5

For all that’s being made of Crede’s “breakout season," PECOTA expected more of the third baseman. In a continuing theme, Crede’s defensive performance was far better than we’ve come to expect of him.


I haven't seen people talking about this as a breakout season for Joe. It was mentioned as a breakout POST season. And if you didn't expect top tier defense from Joe Crede, it is only because you never saw him play. I'm not sure what your defenisve models say about his D before this year, but those who saw him play regularly had already said he was in the class with guys like Chavez.

Juan Uribe (file:///dt/uribeju01.shtml)

Actual 540 16 71 .251/.294/.411 12.9
PECOTA 595 19 73 .269/.316/.444 22.0

Your VORP has a 12.9 for Juan, and a 30.9 for Iguchi. That's a fundamental flaw in the system. Any system that had Manos being worth more than Magglio is clearly not a valid analytics tool to begin with…but this is just more evidence that you seem to throw numbers into a random number generator and let the outcome be interpreted without any care for the associated processing.

Scott Podsednik (file:///dt/podsesc01.shtml)

Actual 568 0 25 .289/.347/.348 13.4
PECOTA 689 14 65 .278/.342/.416 19.3

What were the odds that Podsednik would underperform his projection and make it to the All-Star Game in the same season? I guess chicks don’t dig the longball as much as we think.



What are the odds that VORP is a dysfunctional "statistic" that has absolutely no relevance to winning baseball? Your insistence on measuring a leadoff hitter based on his SLG% only goes to indicate how little you truly understand baseball. If my boss were to come up to me and blame me (the CIO) for low sales, I'd laugh about it - since I am not running the sales team. Using VORP to convict Pods is similar in nature - as I don't give a damn what his SLG% is as long as he is on base and wreaking havoc on the basepaths.

Aaron Rowand (file:///dt/rowanaa01.shtml)

Again, the defense was better than expected.



You never saw Aaron play CF.


We’ll present the following as a package:

Mark Buehrle (file:///dt/buehrma01.shtml)
Jon Garland (file:///dt/garlajo01.shtml)
Freddy Garcia (file:///dt/garcifr02.shtml)
Jose Contreras (file:///dt/contrjo01.shtml)

Oh, this is how you confuse a 100 win team for an 82 win team.


Yes - you take wild-ass guesses about how pitchers will perform, and make those guesses through your own biased view of the game. That’s how you confuse a 99 win team for an 82 win team.

All that repeated "underestimated his defense" comes to roost in these numbers, since the White Sox pitchers allowed batters to put a fair number of balls in play.

Ozzie, Kenny and a lot of others said, from the beginning of the spring, that this team was built on defense and pitching - you just failed to consider this is a valid approach since it doesn't fit into your models. FYI, allowing balls to be put in play is not a bad thing - it depends on how and where they are hit.

The reason you were dead wrong to have predicted this team to be an 82 win team, Joe, isn't that players outfielded your expectations and PK/Dye had better seasons than you guessed. That's stupid. That doesn't explain 17 wins. The reason your models led your analysts to predict 82 wins, and we won 110 is because your models are fundamentally flawed. There is no way to take a game that is as complex and multidimensional as baseball is and reduce it to a point where a few statistics can predict what will happen either in any given game, or over the course of a 162 game season. If there was, you'd be making money in Vegas, not making money by suckering dopes into paying to read your dribble.

jabrch
10-29-2005, 11:07 AM
FYI - some of that was edited to keep in under the text limit and all in one post. But I didn't substantially change what Dumb Joe said, and I linked to his original piece for the sake of accuracy.

Tragg
10-29-2005, 11:14 AM
ONe thing I've never understood about BP or Sabremetrics, etc.

They consistently argue that a high strikeout total for a hitter is irrelevant.

They also consistenly argue that a pitchers' strikeout rate is one of the most telling signs of pitchers' effectiveness.

To me that's a remarkable inconsistency. If it doesn't matter if a hitter strikes out for his outs, why does it matter if a pitcher gets his outs from strikeouts?

Anyway, anyone with a modicum of baseball knowledge knew that we had a better staff to start the year; no back-end weaknesses and a front 2 of MB and Garcia, while not in the Johnson/Schilling top tier starter, were solid #2 starters (and better #1 starters than 50% of the teams in MLB have). And that was at the start of the season when Johnson/Schilling were elite and based on MB/FG historical stats.

jabrch
10-29-2005, 11:20 AM
They consistently argue that a high strikeout total for a hitter is irrelevant.

They also consistenly argue that a pitchers' strikeout rate is one of the most telling signs of pitchers' effectiveness.

To me that's a remarkable inconsistency. If it doesn't matter if a hitter strikes out for his outs, why does it matter if a pitcher gets his outs from strikeouts?


I know - they argue that 163 Ks from Danny Haren is what makes him so good, but 169 Ks from Richie Sexson doesn't mean anything...

It's bad statistical modeling - nothing more.

TornLabrum
10-29-2005, 11:22 AM
I would think a high strikeout rate would be bad for a pitcher since it means he could very well be throwing more pitches than necessary. Wouldn't the ideal pitcher be one who is effective in getting guys to ground out or pop out early in an at bat, one who looks deceptively hittable? Of course, I'm only thinking about wear and tear and the abuse factor....

jabrch
10-29-2005, 11:28 AM
I would think a high strikeout rate would be bad for a pitcher since it means he could very well be throwing more pitches than necessary. Wouldn't the ideal pitcher be one who is effective in getting guys to ground out or pop out early in an at bat, one who looks deceptively hittable? Of course, I'm only thinking about wear and tear and the abuse factor....

The fact is, and I know you know this, there are more than one ways to be an effective pitcher. Some guys can do it with Ks. Some guys are more effective with fewer Ks, Some guys get worse when they get Ks. Others are worse when the ball is put into play.

Thus, no single stat can be used to evaluate every player - since there are multiple ways to be an effective pitcher. Garland, more than most, doesn't need Ks to be good. He was 30th in the AL in Ks. He let people put the ball in play. And he was clearly better than the 30th best pitcher in the AL. (in the top 5 - I'd say behind Santana, Colon and a few others)

BP and the Sabredopes unequivocating belief in Ks for pitchers vs Ks for hitters is just another area where their models lack legitimate analytical rigour.

fquaye149
10-29-2005, 12:14 PM
I would think a high strikeout rate would be bad for a pitcher since it means he could very well be throwing more pitches than necessary. Wouldn't the ideal pitcher be one who is effective in getting guys to ground out or pop out early in an at bat, one who looks deceptively hittable? Of course, I'm only thinking about wear and tear and the abuse factor....

The old saying is that the only true outcomes in baseball are the walk the strikeout and the home run. That is, those are only outcomes that don't involve unpredictable factors like fielding, weather, etc. etc. That's kind of the argument. A pitcher who strikes out a lot of hitter isn't relying on bad bounces not to happen, good fielders, the batter to get fooled, etc.

I don't know how true this is. I would guess it underestimates a pitcher's pitcher's ability to confuse a hitter or saw them off or whatever. But nevertheless this is the claim by sabrmetricians who place high value on the strikeout.

I guess the reason they don't condemn the strikeout for hitters is that there's no real correlation between high strikeout hitters and poor production (at least when you look at the numbers at the end of a season). Actually a lot of strikeouts is a good thing to them, because it means they're taking a lot of pitches. Look at Dunn - he walks a lot and he strikes out a lot as well - he waits for pitches he can hit at the expense of getting behind in the count. Also he strikes out looking a lot trying to take a walk.

These are just their claims. I don't know if it's adequate or not. I know there's a lot about baseball I don't understand. Right Hawk?

:hawk:
"heck you could play baseball for infinity years and not even know what the infield fly rule is!"

jabrch
10-29-2005, 12:20 PM
One may be an effective pitcher by keeping the ball down, and in the zone. You won't throw many pitches, and you will induce lots of ground balls. Now what happens after that, is of course out of your control. But I'll tell you this much - a guy who has a strong IF defense behind him, and gives up lots of 1 and 2 pitch GBs will get a lot of outs. Sure - he will have some seeing eye grounders that will get through the holes. But if your plan is to yield ground balls, keep your walks down, and give up the occasional HR, you can have a pretty good season that way - just ask John Garland. I'd rather have that than a pure power pitcher who throws 90 pitches through 5 innings, gives up runs on a few walks and a blast, etc. like some of the other guys who throw in this town.

Buehrle and Garland both showed that you can get guys out very effectively with good IF defense and strong pitching inside the strike zone.

SOXintheBURGH
10-29-2005, 12:23 PM
I never grow tired of these... Thanks jabrch!:kneeslap::drunken:

jabrch
10-29-2005, 12:32 PM
I never grow tired of these... Thanks jabrch!:kneeslap::drunken:

Later this offseason I plan on going through my copy of BP2005. I'll have 567 pages of this stuff to read through. I already have picked out some really ripe ones just waiting... Here's a preview for you...

Their pandering to Barry Bonds starts out like this...
"Do steroids, call things considered, improve a baseball player's perfromance? For all the hand-wringing, moralizing, and hoigh-horsing the issue has triggered, the net effect of steroids on baseball players remains unknown."

it ends with...

"Bonds will remain the best hitter in the game in 05, though injuries and age will start cutting into his production. We'll leave the moral outrage to someone else."


Yes - that's right - the "Moral Outrage" function in Excel is broken - so you can't calculate it enough to realize that Barroids is not the greatest hitter of all time, rather the greatest CHEATER of all time. And since that doesn't play into your models, you go and make all the justifications for this spoiled rotten lying and cheating piece of dirt you want.

Hendu
10-29-2005, 12:48 PM
ONe thing I've never understood about BP or Sabremetrics, etc.

They consistently argue that a high strikeout total for a hitter is irrelevant.

They also consistenly argue that a pitchers' strikeout rate is one of the most telling signs of pitchers' effectiveness.

To me that's a remarkable inconsistency. If it doesn't matter if a hitter strikes out for his outs, why does it matter if a pitcher gets his outs from strikeouts?

Anyway, anyone with a modicum of baseball knowledge knew that we had a better staff to start the year; no back-end weaknesses and a front 2 of MB and Garcia, while not in the Johnson/Schilling top tier starter, were solid #2 starters (and better #1 starters than 50% of the teams in MLB have). And that was at the start of the season when Johnson/Schilling were elite and based on MB/FG historical stats.

I can unterstand the love of strikeouts for a pitcher to some extent. After all, it is the only guaranteed out in baseball (unless Josh Paul is catching :D:). No contact, no chance for bloop hits, infield hits, etc.

However, I think something that's just as important as K's is getting ahead in the count. That's why our pitching staff was successful. No need to make a perfect pitch into a teacup when the batter is constantly 0-1 or 0-2.

jdm2662
10-29-2005, 06:49 PM
Strike outs for a pitcher is so overrated. Yes, they are great for fantasy baseball teams. However, an out is an out however you get it. A lot of strikeouts can be a bad thing since they may require more pitches thrown.

Banix12
10-29-2005, 06:57 PM
Strike outs for a pitcher is so overrated. Yes, they great for fantasy baseball teams. However, an out is an out however you get it. A lot of strikeouts can be a bad thing since they may require more pitches thrown.

I think you just "pegged" where Joe Sheehan and analysts like him get most of their baseball knowledge.

Daver
10-29-2005, 07:06 PM
I would think a high strikeout rate would be bad for a pitcher since it means he could very well be throwing more pitches than necessary. Wouldn't the ideal pitcher be one who is effective in getting guys to ground out or pop out early in an at bat, one who looks deceptively hittable? Of course, I'm only thinking about wear and tear and the abuse factor....

Change the rules so that the batter is out on a HBP and you could see some 27 pitch ballgames.........




:)

MRKARNO
10-29-2005, 07:21 PM
One of the biggest problems that I have with BP's conception of sabermetric theory, (which, by the way is only a school of sabermetric thought. Bill James openly dissagrees with a lot of the stuff they post over there) is that it is entirely reliant on the Defense Independent Pitching Statistics (DIPS) which state that pitchers have no control over a ball in play and it evens out over the long haul, for all pitchers. The problem is that not all balls in play are equal. Jon Garland is going to induce more popups than Carlos Silva and Silva is going to produce 200000x as many groundballs as Garland. Popups are a lot easier to handle than groundballs, but double plays can result from groundballs. I think the research on balls in play that BP has done is lacking and unconvincing and that they have not fully explored the topic enough to make the types of statements and assumptions that they operate under.

Look at Johan Santana. Look at Wakefield. Look at Garland. All of them have managed to keep their BABIP pretty constant over the past three years and DIPS theory says that it is nearly impossible to do that. I think you can apply the rule to a large number of pitchers, but certainly not all of them, as they like to do.

jabrch
10-30-2005, 10:21 AM
One of the biggest problems that I have with BP's conception of sabermetric theory, (which, by the way is only a school of sabermetric thought. Bill James openly dissagrees with a lot of the stuff they post over there) is that it is entirely reliant on the Defense Independent Pitching Statistics (DIPS) which state that pitchers have no control over a ball in play and it evens out over the long haul, for all pitchers. The problem is that not all balls in play are equal. Jon Garland is going to induce more popups than Carlos Silva and Silva is going to produce 200000x as many groundballs as Garland. Popups are a lot easier to handle than groundballs, but double plays can result from groundballs. I think the research on balls in play that BP has done is lacking and unconvincing and that they have not fully explored the topic enough to make the types of statements and assumptions that they operate under.

Look at Johan Santana. Look at Wakefield. Look at Garland. All of them have managed to keep their BABIP pretty constant over the past three years and DIPS theory says that it is nearly impossible to do that. I think you can apply the rule to a large number of pitchers, but certainly not all of them, as they like to do.

Exactly!

I pitcher who has stuff that tends to induce ground balls, and keeps the ball in the park likely will perform better than a pitcher who either goes K or HR. It's simple stuff really. If some of these fools watched a little bit of baseball, they might frigure it out.

voodoochile
10-30-2005, 12:09 PM
I would think a high strikeout rate would be bad for a pitcher since it means he could very well be throwing more pitches than necessary. Wouldn't the ideal pitcher be one who is effective in getting guys to ground out or pop out early in an at bat, one who looks deceptively hittable? Of course, I'm only thinking about wear and tear and the abuse factor....

Many statisticians claim that pitchers have no control over where a batted ball goes, so a guy who doesn't gives up less contact outs is more effective since the guy who gives up a lot of contact outs is simply getting lucky.

Overall, I completely agree with jabarch on this analysis. I don't know why they claim Iguchi is bad defensively. Maybe it is in part that Uribe gets to so many balls up the middle (thus allowing Iguchi to shade to his left) and thus reduces the statistical value of his defense, while making the whole infield defense that much better.

I think the vorp stuff is hysterical in general. It has some meaning, but not on the level they have it because if you actually fielded any teams AAA team for an entire season (the mythical "replacement players" they are referring to) that team would be lucky to win 40 games and the number might end up closer to 30 once the team cracked like a dropped egg and gave up in mid-May with 35 losses under their belt already and a pitching staff in mental tatters.

Great point about the 3-year averages too, jabarch. It's how they project Pods to have 14 HR (because he had like 5 one year and like 23 the next or something equally ridiculous). It's not a valid factor with the younger players and the Sox had a few on this team. It ignores trendlines (PK's HR's have been increasing steadily, Pods was playing in a better field for him and made an effort to shorten his swing to be more effective by reaching base more, Crede is finally finishing growing up, etc.) A guy who is 30 and has 7 consistent seasons under his belt (see Frank CA 1998) would be someone you could make accurate projections about (though you would have failed with Frank due to his injury), but trying to project for younger players who are still feeling their way toward consistency is ridiculous and obviously their defensive stats are for crap...

cheeses_h_rice
10-30-2005, 01:42 PM
I can't stand propellerheads.

Tragg
10-30-2005, 02:14 PM
Just for the record, Sheehan's system actually said that we were a 72 win team at the start of the season. We became an 82 win team AFTER they re-ran the numbers after the 1st 2 months of the season.
By this time, I'm sure we've progressed to an 86 win team. If we could just improve our margin of victory, then the computers would luv us.

MRKARNO
10-30-2005, 02:44 PM
Exactly!

I pitcher who has stuff that tends to induce ground balls, and keeps the ball in the park likely will perform better than a pitcher who either goes K or HR. It's simple stuff really. If some of these fools watched a little bit of baseball, they might frigure it out.

As far as I'm concerned, their reliance on DIPS is a cop-out. They like to say that baseball is more complicated than "pitch the ball, hit the ball, catch the ball," and it is, but it's even more complicated than they give credit. Most of their statistics are context-irrelevent and in judging actual performance, it is impossible to view the game this way.

Also, a lot more work needs to be done about what happens when the ball does go in play. Good statistical evaluations about defense and pitcher's ability to make it easy on a defense just are not out there right now or are at least not being fully explored. More work should be done about popup rates, line drive rates, etc., as well as just about everything related to defense.

Statistical analysis is definitely a valid way to look at the game, and I stand by that, but a lot of the conclusions that BP in particular have come to are very questionable.

Stroker Ace
10-30-2005, 03:24 PM
Since when did baseball turn into a math class?

Flight #24
10-30-2005, 03:52 PM
What's hilarious is that it's all based on the premise that 1-run games are almost entirely won by luck. Therefore a team that plays close games must by definition be around .500 or worse.

Which of course, to anyone who watches games is crap. Teams that execute and can scratch out runs offensively, and who have good 'pens consistently win those games. And if you add in top-flight starting pitching, another of their theories goes out the window - that scoring first is meaningless. When you have a dominant starter on the board, giving them an early lead lets them settle down and pitch easier.
Even the uberlord of stat geeks, Bill James, is on record as saying some teams consistently win close games, and those that do tend to play effective smallball. Also that some managers seem to win those types of games more consistently than others. Ozzie's 2-year record indicates that he may well be one of those managers.

MRKARNO
10-30-2005, 04:37 PM
Since when did baseball turn into a math class?

Since Branch Rickey essentially invented statistical analysis, way back when:

http://www.baseballthinkfactory.org/btf/pages/essays/rickey/goodby_to_old_idea.htm

That ****ing idiot propellorhead.

jabrch
10-30-2005, 05:53 PM
Statistical analysis is definitely a valid way to look at the game, and I stand by that, but a lot of the conclusions that BP in particular have come to are very questionable.

I'll say it is a valid way to look at the game within the context of reason. However, none of the BP models even remotely propose to bring in the context of reason. None of the Propellerheads with a podium profess this either. They have their few monolithic stats, and they try to use those to extrapolate out the game down to statistics - thus being able to call us a 72, 82 and 86 win team despite the fact that we won 110, and then still justify that.

You can not isolate out enough variables in sport, with any sort of reliability, to use statistics as PREDICTORY measures. They are great for the purpose of historical evaluation. Few love evaluation of HISTORICAL stats more than I. But to use statistics as predictory measures, in baseball in particular, continues to be an unreliable discipline, as shown by the powers that be in Propellerhead Land.

jabrch
10-30-2005, 05:56 PM
Since Branch Rickey essentially invented statistical analysis, way back when:

http://www.baseballthinkfactory.org/btf/pages/essays/rickey/goodby_to_old_idea.htm

That ****ing idiot propellorhead.


That Rickey article mostly professes statistics as measures of what a team did - and not predictors of what an individual, or a collection of individuals will do.

Mendoza Line
10-30-2005, 07:17 PM
Stats are overwhelmingly useless when defining baseball, unless you're playing it on your Xbox. Someone remind these guys that it's a game played by human beings, not machines. You just can't measure heart with numbers.

MRKARNO
10-30-2005, 07:33 PM
Stats are overwhelmingly useless when defining baseball, unless you're playing it on your Xbox. Someone remind these guys that it's a game played by human beings, not machines. You just can't measure heart with numbers.

Statements like these are as ridiculous as those made by the Joe Sheehan's of the world. People who deny the use of statistics in baseball are blind to the reality that statistics often do reflect what is going on in the baseball diamond. On the other side of things, those who deny the use of scouting in baseball are equally ridiculous, for reasons that many here can state pretty well. The question is, "How do you best reconcile the statistics and the scouting?" not "Should I use only scouting or only statistics?" There is a place for both in baseball and I think most GMs know that.

FedEx227
10-30-2005, 10:01 PM
BP is ridiculous. While I respect some of the Moneyball guys with their utilization of statistics, mainly OBP which I feel is a great statistic and Beane's hatred for batters who strikeout early in counts (Tejada/Soriano/Stash)... but BP is just stupid, they are so hard-headed over there

They have no baseball-knowledge over there. I wrote a paper of the history of baseball statistics and went there to ask the guys some questions, the responses I got were awful "I'm not too sure about where they started, Alan Schwarz has a good book about it" or "I'm not sure of anything really past 1996".

I love these breakdowns jabrch, me and one of my friends tend to do the same thing from time to time look at Baseball Prospectus and laugh at its stupidity.

Its good to know, though, that we would've won the world series much easier if Willie Harris' TRIOTILOIL7 was .00065753 points lower than his APOAPO!

1951Campbell
10-30-2005, 10:39 PM
A guy who is 30 and has 7 consistent seasons under his belt (see Frank CA 1998) would be someone you could make accurate projections about (though you would have failed with Frank due to his injury), but trying to project for younger players who are still feeling their way toward consistency is ridiculous and obviously their defensive stats are for crap...

No doubt. I'm sure if BP was around in the 70's, they'd "project" Mark Fidrych to win about 400 games, even though he clearly had a horseshoe up his bum before he got injured.

Chips
10-30-2005, 10:51 PM
Since Branch Rickey essentially invented statistical analysis, way back when:

http://www.baseballthinkfactory.org/btf/pages/essays/rickey/goodby_to_old_idea.htm

That ****ing idiot propellorhead.



http://www.baseballthinkfactory.org/btf/pages/essays/rickey/Image27.gif

http://www.baseballthinkfactory.org/btf/pages/essays/rickey/Image28.gif

What does F stand for?

MRKARNO
10-30-2005, 11:17 PM
What does F stand for?


Honestly, I'm not sure. The only thing I know that F has stood for is "Final" and Flyouts and only in the context of G/F ratios.

Chips
10-30-2005, 11:25 PM
Honestly, I'm not sure. The only thing I know that F has stood for is "Final" and Flyouts and only in the context of G/F ratios.

The only other F that I could think of was Faced batters, but that doesn't fit into the equation nicely. I think something is missing or I just don't know what I am doing.

ma-gaga
10-30-2005, 11:32 PM
Did a google search. 'F' is fielding.


http://fairmodel.econ.yale.edu/rayfair/pdf/2005C.PDF

MRKARNO
10-30-2005, 11:39 PM
Did a google search. 'F' is fielding.


http://fairmodel.econ.yale.edu/rayfair/pdf/2005C.PDF

And this paper states something to the effect of "he does not use F in this article and it is not shown how he measures 'F,' so just ignore it."

Chips
10-30-2005, 11:55 PM
I just entered all of stats for 2005 into that equation and our winning percentage would be .473. I'm pretty sure we did better than that. I doubled checked my math, so unless I used the wrong numbers somewhere, statistics like that don't mean ****.

MRKARNO
10-31-2005, 12:36 AM
I just entered all of stats for 2005 into that equation and our winning percentage would be .473. I'm pretty sure we did better than that. I doubled checked my math, so unless I used the wrong numbers somewhere, statistics like that don't mean ****.

I believe G is an index and not an actual predictor of win percentage, because I just plugged in the numbers and got something which had no resemblence to a win percentage, and after perusing the article, it seems that it is more of an efficiency rating than anything else.

voodoochile
10-31-2005, 01:05 AM
As far as I'm concerned, their reliance on DIPS is a cop-out. They like to say that baseball is more complicated than "pitch the ball, hit the ball, catch the ball," and it is, but it's even more complicated than they give credit. Most of their statistics are context-irrelevent and in judging actual performance, it is impossible to view the game this way.

Also, a lot more work needs to be done about what happens when the ball does go in play. Good statistical evaluations about defense and pitcher's ability to make it easy on a defense just are not out there right now or are at least not being fully explored. More work should be done about popup rates, line drive rates, etc., as well as just about everything related to defense.

Statistical analysis is definitely a valid way to look at the game, and I stand by that, but a lot of the conclusions that BP in particular have come to are very questionable.

I was thinking about some of this on the way to work today after posting in this thread earlier.

Defensive stats are inherently bad because they are all simply based on what other people at that given position do during the year.

For example: If the 30 starting CF for all MLB teams combine to put up 10000 outs during the season (roughly 2 per game per player) then a player who averages 2.5 outs per game is a great defender, while a guy who puts up 1.5 per game is a poor defender. Of course those stats are NOT isolated. It is all based on other factors, weather, park size, style of pitchers (a guy who has 4 ground ball pitchers on his team is going to have a lot less chances than a guy who has 4 fly ball pitchers on his team), opponents batters (think the Yankees get more flyballs playing the Red Sox than the Sox do playing the Royals?), etc. So it's a gross oversimplification to compare two starting CF's and say "player A records more putouts than player B so he must be a better defender." In addition of course how good the guys playing the corner spots are has a huge influence on these numbers too. I'd bet good money that Toriiii Hunter does not get as many chances as some other CF do because the guys who play the corner spots with him are excellent defenders too.

It's just one example of how defensive stats are misunderstood and very hard to quantify. In the end, the scouting approach is as effective or more so because a good scout can see the whole picture. In short they are acting as a sort of computer who can take all of the other factors into consideration.

Flight #24
10-31-2005, 10:07 AM
http://www.baseballthinkfactory.org/btf/pages/essays/rickey/Image27.gif




http://www.baseballthinkfactory.org/btf/pages/essays/rickey/Image28.gif













What does F stand for?



F= Fudge Factor. Defined as follows:
If Billy Beane is the GM, F=-30
If a Beane-ite is the GM, F=-20
For most GMs, F=0
For Kenny Williams, F=20

:wink:

maurice
10-31-2005, 06:27 PM
Just for the record, Sheehan's system actually said that we were a 72 win team at the start of the season.

Exactly. He also had us pegged for 4th place and was convinced that we wouldn't even make the playoffs, even after the hot start. Instead, we won 99 games + 11 playoff games and the WS. "Luck" might account for a few wins here and there, but it's not the difference between a 4th place team and WS champions.

The REAL reason that Sheehan was off by 38 wins is that Sheehan is incompetent.

SoxSpeed22
10-31-2005, 09:04 PM
I'm surprised FWC hasn't gotten into this yet. Sheehan can keep trying, but we won, period.
:dumbass:
or
http://home.comcast.net/~sjredux/Aku.jpg"FOOL!"

MRKARNO
10-31-2005, 09:13 PM
I'm surprised FWC hasn't gotten into this yet. Sheehan can keep trying, but we won, period.


And as one of Sheehan's collegues put it, "Flags fly forever."

ma-gaga
11-01-2005, 07:04 PM
And this paper states something to the effect of "he does not use F in this article and it is not shown how he measures 'F,' so just ignore it."

Yeah... I just did a google search for "branch rickey formula "F"" and came up with that link. I didn't check the formula, I just skimmed until I found F somewhat defined.

Defensive stats are inherently bad because they are all simply based on what other people at that given position do during the year.
...
It's just one example of how defensive stats are misunderstood and very hard to quantify. In the end, the scouting approach is as effective or more so because a good scout can see the whole picture.

I don't argue this at all. Defensive stats listed on ESPN, or BP are worthless. I think you need a fielding chart, and scouting to get a much better sense of the capabilities of a fielder.

I think some people are getting "smart" and doing a better job trying to mesh the subjective and the objective. In case you've never seen it before: Tango Tiger's Fielding Report. (http://www.tangotiger.net/scouting/index.html) For the fans, by the fans. Go ahead and rate out the W.Sox players if you think you are a good enough judge of baseball talent. I like this one. It's the right idea IMO. You just need baseball professionals (scouts) doing this kind of stuff, rather than 100's of biased fans, who's opinions are skewed by the local television or newspaper columnists.

Then again, Tango flies a bit under the radar. :cool:

MRKARNO
11-01-2005, 08:27 PM
I would dissagree with the statement that it's impossible to capture defense with statistics, but I think the types of statistics that would be needed are going to be a lot different than the statistics that we think of today because they'll need to be more situation-based and require a lot more visuals, such as field charts. But in capturing these stats, there is going to have to be some more of a scout person to take note of these as opposed to an official scorekeeper. What needs to occur is, instead of scouts going to look at players and saying "I've seen him and he's got a great arm and great range," they're going to have to go out and say, he ran from point A to point B and that's a heckuva long distance to run to make that play and it would have been a dunk-in single for anyone else. Numbers are going to play a role in the sense of the distances that fielders can cover and their reaction times.

FarWestChicago
11-01-2005, 08:49 PM
I would dissagree with the statement that it's impossible to capture defense with statistics, but I think the types of statistics that would be needed are going to be a lot different than the statistics that we think of today because they'll need to be more situation-based and require a lot more visuals, such as field charts. But in capturing these stats, there is going to have to be some more of a scout person to take note of these as opposed to an official scorekeeper. What needs to occur is, instead of scouts going to look at players and saying "I've seen him and he's got a great arm and great range," they're going to have to go out and say, he ran from point A to point B and that's a heckuva long distance to run to make that play and it would have been a dunk-in single for anyone else. Numbers are going to play a role in the sense of the distances that fielders can cover and their reaction times.Isn't that really just using numbers to use numbers. You already have the scout making the assessment. :smile:

Daver
11-01-2005, 08:58 PM
Isn't that really just using numbers to use numbers. You already have the scout making the assessment. :smile:

Once a propellorhead.......

SABRSox
11-01-2005, 09:03 PM
A plea:

Baseball Prospectus does not encompass the entirety of the Sabermetrics world. Can we please stop with the BP=SABR assumptions? While it happens to be one of its most popular publications, the fact remains that there are some people who are very interested in sabermetrics and think that BP is more often than not completely wrong in their analysis.

(I personally believe the site is skewing way too far towards the fantasy baseball player, and this is evident in a seemingly total disregard of improving and furthering study in defensive statistics.)

It's a shame that Joe Sheehan has become the mouthpiece of SABR. I think he hurts SABR more than he helps. In his wrap-up of the White Sox season, he said "We can learn more." While I agree that this is the creed of sabermetrics, I have my doubts that he'll actually follow through. He's the Mariotti of SABR, as far as I'm concerned, and everytime another one of these threads open and it's a SABR bash fest, I cringe. I think of it like people who think the Ligues represent the sum of White Sox fans. It's just not the truth.

FarWestChicago
11-01-2005, 09:11 PM
It's a shame that Joe Sheehan has become the mouthpiece of SABR. I think he hurts SABR more than he helps. In his wrap-up of the White Sox season, he said "We can learn more." While I agree that this is the creed of sabermetrics, I have my doubts that he'll actually follow through. He's the Mariotti of SABR, as far as I'm concerned, and everytime another one of these threads open and it's a SABR bash fest, I cringe. I think of it like people who think the Ligues represent the sum of White Sox fans. It's just not the truth.I've never seen you or ma-gaga act like BP morons. And it isn't just BP. It's the FOBB's. The Beane worship really bothers a lot of us. But, you and ma have always seemed like reasonable fellows.

MRKARNO
11-01-2005, 10:55 PM
Isn't that really just using numbers to use numbers. You already have the scout making the assessment. :smile:

I've been preaching that a healthy balance of scouting and statistics is necessary to run a successful ML organization. The more successful scouts are going to have to have a better grasp on at least basic numbers and more detailed information than they may have been in the past. The only numbers which seem to matter to them now are the times to first base and the speed gun reading. The successful scouts of the future are going to need to provide more objective information from their observations than in the past and come to conclusions based on data that they themselves have collected.

Additionally, on the statistical analysis side of things, there is going to have to be more observational analysis involved with the gathering of the relevent stats of the future. There is only so much that the box score can tell without actually looking at what's happening on the field.

The two "sides," if they are to be effective in the future, will come together and become more like each other. The folks at BP can't let go of this and I would agree with SABRSox's general view of BP. There are also the folks who have totally fawned over Depodesta, Ricciardi and Beane without taking a critical look at what they've actually done. I don't understand why some in the sabermetric community are so quick to praise the actions of these three besides them being one of their own, or at least that's their perception. I've been called a FOBB around here, but I have never been one to irationally praise the dealings of these three, as many in some baseball circles have been eager to do.

Back to the main point, on the other hand, scouts who resist the fact that their opinion might actually need to be backed up with examples and facts probably shouldn't be holding on to their job very long. Just as in anything, you can't make statements without support and my perception is, scouts would sit at a game, watch the game and their reports would go something like "Good arm, average range, lots of power, good bat speed," etc. Obviously, some of this can't be objectively quantified, but you can do some objective quantification in some areas.

Mr. White Sox
11-01-2005, 11:40 PM
Just to clarify to everyone, Joe Sheehan didn't write that piece on the White Sox; Derek Jacques (derekbaseball@gmail.com) did.

Anyway...

The problem that I see with BP is that they take their correct predictions and tout them as though they are the masters of the universe. If they're wrong and can get away with it, they won't mention their idiocy. The White Sox won the world series, so they had to call themselves out on their prediction. I didn't check, but how many other predictions did those guys get right/wrong?

I'd love to see a SABR site with legitimate analysis, writers who understand baseball and don't belittle certain teams or GMs, and statistical projections that aren't moronic. In fact, I'm sure that most SABR sites are as described, but BP is the exception.

jabrch
11-02-2005, 12:27 AM
The Beane worship really bothers a lot of us. But, you and ma have always seemed like reasonable fellows.

I totally agree with West. It's the WORSHIP of BB and of the use of stats that drives me nuts. It's when people attempt to apply numbers over logic that I get riled up. At the end of the day, I don't want a calculator trying to tell me what I saw. I watch enough baseball to know what is a good team, what is a bad team, what makes a good player, and what makes a bad one. I don't need a calculator to tell me that. And frankly, I trust my evaluation much more than I do numbers put into a formula that was manually created to a certain outcome. Core statistics do a good job telling you what happened. However overly regressed statistics, or numbers ran through artificially created hoops so many times that they no longer look close to the initial raw data, do not paint an accurate picture of what happened, and they certainly don't tell you what is going to happen.