Baseball Question

[Clarkmeister]

For all of you who have read all Bill James' writings. Has he (or anyone ele) ever quantified how much of an individual players value comes from defense? That is, is offense 75% of the game? 80%? 90%? Maybe I'm not asking the question in the right way, but I hope you can see what I'm getting at.

Say one player was the best defensive player in the game and another was average. How much better does the average fielder need to be offensively? That's kind of the end result of where I'd be looking to go with the answer to my initial question.

[adios]

Not a well read Bill James guy, but if memory serves me correctly James does an analysis on how much a great fielder means in terms of runs than a very good fielder means in terms of runs over a season. It was a surprise to me that it meant so few. My take was that hitting is way, way more important.

[Clarkmeister]

So I read an article that says for win share calculations, individual defense is 1/3rd of the defensive calculation, which is half the total calculation. So does that mean that individual defense constitutes 1/4th of a player's value? (Half is defense and half is offense. However, only 1/3rd of defense is in control of a player, and all of the offense being in his control, leaving 1/4 of the controllable individual worth being attributable to their defense?)

[Clarkmeister]

I know that in Moneyball, Beane says they calculated the run impact of losing Damon in CF to be 15 runs. I'm unsure of how to relate that to offense though. Wouldn't we need some sort of par value for offense?

[adios]

[ QUOTE ]
I know that in Moneyball, Beane says they calculated the run impact of losing Damon in CF to be 15 runs.

[/ QUOTE ]

That seems kind of high but maybe not. Can't remember. I believe in the James book I read a little of, he goes on to relate the runs to probably victories.

[ QUOTE ]
I'm unsure of how to relate that to offense though. Wouldn't we need some sort of par value for offense?

[/ QUOTE ]

I'm sure he does this but I can't remember exactly what he does. I'm obviously not the right guy to ask /images/graemlins/smile.gif. Andy where are you?

[Kurn, son of Mogh]

Not very scientific, but I always thought that when the red Sox had Darren Lewis in CF, his defense was woth about .25 runs/game, or 40 per year. Probably an overstatement, but after his first year here I based it on how many times he'd make a catch and I'd think "wow, last year, that's a double."

[Kurn, son of Mogh]

How much better does the average fielder need to be offensively?

As an example, I don't think an average fielding CF could be good enough offensively for me to take him over Andruw Jones.

[Clarkmeister]

[ QUOTE ]
How much better does the average fielder need to be offensively?

As an example, I don't think an average fielding CF could be good enough offensively for me to take him over Andruw Jones.

[/ QUOTE ]

Nah. What if you had to put either Bonds or Jones in CF and couldn't have both? You'd suck up Bonds' defense and put Bonds out there, right?

[Michael Davis]

What if you could have either Frank Thomas in his prime or Jones in CF?

-Michael

[nolanfan34]

I think James, along with the rest of the sabermatrician folks out there, are still struggling with measuring defensive worth. I believe that Baseball Prospectus includes some defensive value in their VORP (Value Over Replacement Player) rankings, although I can't say I completely understand how they can possibly measure it.

I do know that the Mariners have probably lost a few games due to the fact that Randy Winn is in CF instead of Mike Cameron. Cameron supposedly saved about 30 runs over the course of the season last year, according to BP's metrics. I do believe there's some validity to those numbers, because between Cameron, and the addition of Kaz Matsui at SS, the Mets pitchers certainly are seeing lower ERA's compared to last season.

I highly recommend Baseball Prospectus for their baseball coverage. www.baseballprospectus.com. (http://www.baseballprospectus.com.) No, I'm not a shill, just merely a subscriber who knows that the baseball fans here might enjoy checking out that site. Will Carroll's reporting on injuries is the nuts.

[andyfox]

http://www.baseballgraphs.com/winshares/alwinpos.html#

[andyfox]

I posted the 2003 American League Win Shares. I've tried to read Win Shares and find it almost incomprehensible. So I honestly don't know if it's all James says it is. While I know Derek Jeter always shows up badly in the sabermetric fielding measurements, the Win Shares system shows Jason Giambi earning more defensive Win Shares in 2003 than Jeter. Any system that comes up with that has to be suspect.

Fielding has always been a tough nut to crack. Branch Rickey once said nothing can be done with it. At one time, Larry Bowa was regarded as a much better fielder than Davey Concepcion because he made 20 fewer errors a year. But then the fielding range stats showed Concepcion was getting to 60 more balls a year, so he was better. (I made these numbers up, don't know the exact figures.) But fielding range presented some problems: for example, a 42-year old Napoleon Lajoie was showing up as a phenomenal fielder (and thus showed up as, I think, the 3rd greatest player ever in Total Baseball; he was a fine player, a great player, but he wasn't even the 3rd greatest second baseman ever).

Suppose one plays shortstop for the Atlanta Braves with Maddux and Glavine pitching approximately 30% of all innings pitched. Now that shortstop gets traded to the Arizona Diamondbacks with Johnson and Schilling pitching 30% of the innings. There are now a lot more strikeouts and flyballs so it would stand to reason that you'd get fewer assists. The fielding range statistics would show you had much worse range.

Supposedly, James's defensive Win Shares system accounts for this type of thing.

[HDPM]

I think this is true because of defensive stats. Maybe I am wrong and I am not a baseball or Bill James junkie. I thought the statistical analysis in Moneyball was great, and the point about not being able to understand by watching was excellent. They show how you can't tell the difference between a .275 hitter and a .300 hitter by watching really. But how can you analyze defense without watching? How do the stats really show you some things. Like a guy who knows the hitters and what his pitchers are throwing and positions himself well. Or throwing to the correct place from the outfield. I know the Bill James crowd has tried to break it down in terms of range and some other things, and I suppose it could be done. But is it? And the stats can't show you the times something is scored as a hit where an outfielder ran in circles and was lost out there. Maybe an excellent outfielder gets it or holds the hitter to a single or double. The bad outfielder turns it into a triple. Hard to quantify that isn't it. Once it can be quantified then it has to be analyzed and weighed, a tough thing. Somebody tell me of this is being done because I dunno as I said.

[DougBrennan]

I am at work, so I cannot put my hands on it, but in one of James' Baseball Abstracts (a yearly publication he put out from the early or mid-80's to aerly 90's) he had an article examining that very question, although he phrased it in the opposite manner. That is to say, he posed the question as how much below average hitting should you be willing to accept in order to play an above average defensive player.

He used Ozzie Smith as his subject, and disected his numbers every which way, and came to the conclusion that Smith's added value defensively over the average replacement-level shortstop was not as large as one would intuit.
In other words, James would not be willing to take much of a loss in offense to play a clearly superior defensive player.

Two thoughts I must note here: 1) As Nolan mentions in his response, James, like other statisticions, states that his defensive measurements are open to question and interpretation and 2) whenever James examined this type of question he was always very careful to pont out that you must compare a player with whom he could be replaced, as opposed to comparing the player to a void, i.e. how many hits does Ozzie Smith prevent compared to say Devei Cruz (a shortstop on the Giants of slightly below-average but still major league skills), not how many hits does Ozzie Smith prevent compared to no shortsop in the field at all.

Well, I probably got too long winded and obscurred the point, but in general, Bill James conclusion was that Ozzie Smith did not prevent as many runs as one would think.

I am a big Bill James proponent, BUT: I have recently had the opportunity to compare J.T. Snow's first base play (superb) and Pedro Feliz's play (learning the position) at first, and I'd rather have JT and his .260 average with no power penciled into the lineup. Feliz hits about .270 to .285 and homers about 1 in 20 at bats. Statistically, James may be correct, but I cringe more watching Feliz in the field than I do watching Snow at the plate.

As others have pointed out, there is the question of degree. You play Barry Bonds even if he has to sit in a chair.

Doug

[Zeno]

[ QUOTE ]
You play Barry Bonds even if he has to sit in a chair.

[/ QUOTE ]

You play Barry Bonds because he is on steroids. Kind of a different way to say the same thing, I guess.


The problem of measuring 'defensive acumen' in a player is, I think, immense. The parameters are too vague and subject to whims and biases that influence outcome. I submit that a statistical 'Standard Error' (there is a certain amount of error in every measurement) of all the measurements that you compile would show that the stats are meaningless in terms of real numbers. This would be especially true the more parameters you engage in or try to use to refine your measure of defensive acumen. And using all the standard algebra and math on the things you measure or try to weight automatically compounds the error of each original measurement. Thus, the standard error is probably a larger real number than what you are trying to measure. Poof go the numbers.

Perhaps others have already said this in a different way.

-Zeno

Edit: as an added mixer - most of the measure is probably so ambiguous that it may be immpossible to compute what the Standard Error would be. And if you can't do that......

[andyfox]

Seems to be very position dependent, which makes logical sense. For example, at catcher, the total was 113.96 offensive Win Shares vs. 105.31 defensive. So 48% of the average catcher's value was defensive. But at first base, the total for offense was 200.98 vs. just 31.19 for defense. So just 13% of the average first baseman's value was defensive. (I added these hastily, the numbers might be a bit off, but the point holds.)

Defense can be very important at the key defensive positions. Note that at shortstop, even though Derek Jeter was a far better offensive performer than Jose Valentin, his defensive shortcomings cause him to rank below Valentin.

I think the Win Shares systems measures Win Shares earned, not efficiency. Thus, I note that Nick Johnson, clearly a far better fielder than Jason Giambi, earned fewer defensive Win Shares. This is either because he played far fewer innings at first base, or because the system sucks.

[Kurn, son of Mogh]

No way, unless somehow you could make Bonds 12 years younger. /images/graemlins/cool.gif

[Kurn, son of Mogh]

They show how you can't tell the difference between a .275 hitter and a .300 hitter by watching really

Of course you can't. Assuming 600 at bats, the difference between .275 and .300 is a bit more than one hit every two weeks.

[Josh W]

[ QUOTE ]
While I know Derek Jeter always shows up badly in the sabermetric fielding measurements, the Win Shares system shows Jason Giambi earning more defensive Win Shares in 2003 than Jeter. Any system that comes up with that has to be suspect.

[/ QUOTE ]

Not necessarily. Jeter is widely regarded (by those who have really looked at the numbers) as one of the worst defensive SS in the AL.

Rob Neyer has written quite a bit about this. Jeter averaged about .5 putouts&assists/game LESS than his backups during his career. (this may be the best way to look at defense, as it takes out factors such as pitching staff, turf, etc).

While Giambi is a hippo at first, there are many such hippo's at 1B (there aren't many david seguis, john oleruds, and such), and there are many many fine defensive SSs.

That said...I've detested Jeter for a long while. And my detestation (perhaps a new word) is declining. he dives facefirst into stands. nomar sits and watches. I think any sports fan has to appreciate jete...

Josh

[andyfox]

The problem with any system, especially one rating fielding (and to some extent as well, baserunning) is that it can't measure brains nor heart. You are considering this when you talk about Derek diving into the stands to save a game and Nomar watching. Jeter's fielding may indeed be subpar, yet it's interesting that the most successful team in baseball chose to let him stay at shortstop when they acquired the number one player in the league. I once saw Jeter, when Andy Pettitte got a comebacker with one out and the based loaded, point to home when Pettitte wheeled to throw to second base. Get the first out at home: Pettitte, by then a veteran, should have known this. He forgot in the moment; but not Jeter. (They got a home to first double-play.) There was "the play," the underhand flip to home from the first base line that saved the pennant. The dive into the stands the other day.

In Jeter's eight years his team has won six pennants and four World Series. One can certainly argue, and the numbers back it up, that he has more than made up for his defensive shortcomings with his hitting; but I also think he makes the big plays at big times. There may be no such thing as clutch hitting (at least James claims this), but there is such a thing as clutch fielding. It takes brains and heart.

[Kurn, son of Mogh]

There was "the play,"

The writers may nickname him Derek "past the diving" Jeter, but that play against the A's was, well, to borrow a name from another sport, a Larry Bird play.

[B Dids]

OK-

Jeter didn't dive into the stands to "save the game".

Jeter caught the ball, and after he'd MADE the catch he dove into the stands. Player with better range and a better first step (most SS's in baseball) get to the ball and don't have to face plant after they catch it.

And while we're at it- the stupid play against the A's is SOP for the Yankess, that wasn't Jeter being a genius in any way shape or form. That's how he's supposed to back up a play like that.

I believe the Yanks have Alex at 3rd because while he's a better SS than Jeter is (by a long shot, Alex is words ahead of Jeter, he's so big and quick that he gets to balls on two steps that other people would give for), Jeter'd be an even worse 3rd basemen, as his slow reflexes would be even more exposed in that position.

Quantifying defense has proved very, very difficult for baseball analysts. None of the systems have proved meaningful enough to place a ton of stock in them.

[HDPM]

Hard to measure brains/heart of course, but one thing you can do is teach strategy and fundamentals. I don't care how much the players make or if they're big leaguers. There is NO downside to teaching strategy and fundamentals and stressing situational awareness to players. In a "moneyball" context the returns are excellent because any player regardless of salary or skill can learn them. I am not saying you do an old school thing and go for "heady" "scrappers" and such over better players. But why not drill the good players on some basics. How much would a team improve with better situational awareness, fewer mental errors in defense and baserunning? Maybe it can't be quantified, but I think it is silly for teams to ignore. Then if you get beat, let it be because the other teams were more talented. Why squander what you have?

[andyfox]

1) Diving into the stands: I agree, the play was not as great as ESPN seemed to think it was. Jeter's range on a pop-up, where there is time to get to the ball, is excellent, because of his speed. I don't know why his range on ground balls is poor, your reflexes/first step slowness seems likely.

2) The "stupid" play against the A's was, according to all the Yankees, improvised by Jeter. The shortstop does not go to the first base line on a throw home from right field. Spencer missed the cut-off man and Jeter alertly picked up the ball and flipped it underhanded, backhanded to Posada. If Giambi slides, he's safe. That was the stupid play, not sliding. As it was, he might well have been safe, a very close call. But it was a great play by Jeter. I've never seen a shortstop pick up a throw home from right field.

I think the Yankees have kept Jeter as shortstop because, well, he's Derek Jeter. The captain. Did they think the quicker-reflexed A-Rod would be better at third? Maybe. But they could have switched Jeter to second base. I think it's more of a psychological situation than anything else.

I still think Jeter is a money player, espeically in the field.

[B Dids]

Andy- what I've read is that the Yankee's say that play wasn't improvised. I could be wrong, but that's my recollection. I'll leave it at that.

Jeter's a good player, but to me he's overrated inasmuch as he's not at the same level that Alex is or Nomar was.

But this is because Alex Rodriguez may be the best player ever to play the game. (IMO)

[Kurn, son of Mogh]

they could have switched Jeter to second

Questionable. The footwork needed to play 2nd at this level is not something most lifelong shortstops can pick up in one spring. Pokey Reese might be the only player I've ever seen that can play both positions so well.

[B Dids]

Moreover- 2nd base is a postion that gets beat up (with more frequency than the SS). No way do you want to take his bat out of the line-up because he got hurt when somebody broke up a double play.