Hell…even Hummer Winblad sounds like a scout’s dream.
Bill James started this Sabermetric movement in the early 80’s, claiming that by using complicated data you can predict every outcome. They also believe that you can predict a players future worth so when a player hits a down trend it’s time to let him go. Johnny Damon was a good example, Boston let him go because of back to back down trend years yet, Damon made several playoff runs with the Yankees and Rays, while Boston hasn’t been able to get past the first round. It takes the human factor or clutch play out of the equation.
The numbers guru’s don’t believe that clutch hitting exist, sitting a their computer crunching numbers would have had guys like Reggie Jackson on the bench. Reggie was a below avg. fielder, struck out too much yet come playoffs “Mr. October” would come to life with multiple home run games and clutch hitting.
...Personally if you look hard enough at stats you can find one to justify almost any decision. People that back sabermetrics,think they know best and if you don’t agree with them it’s because you don’t understand baseball. I played ball til I was 40, coached 20 seasons, and managed to put 18 of 20 into the playoffs, won 2 titles, none of them with sabermetrics. I think it’s time that we return baseball to the way it was meant to be played. Team chemistry, clutch hitting, heart, old fashion hard work, and putting the human factor back into America’s game.
Reader Comments and Retorts
Go to end of page
Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.
1. ??'s Biggest Fan! Posted: May 25, 2012 at 07:01 AM (#4139848)Can this statement be true? Data provides you with information on how to properly measure production and on-the-field value. You can make some educated assumptions and gamble on players with odds that are more favorable to your objectives. I doubt a smart guy like Bill James would ever declare numbers can "predict every outcome" given how variable a human baseball player is outside of his statistical history. BABIP even brings in the influence of "luck" into baseball analysis, making the author of the article's statement of predictive claim by Bill James suspect.
This statement is nonsense to the point that the writer is lucky that Bill doesn't like getting into slander lawsuits. Aside from the insult to Bill, this guy betrays a complete lack of understanding of what statistics is, much less sabermetrics. Statistics can reasonably be defined as "that branch of mathematics that deals with things that cannot, at least yet, be subjected to proof." All "sabermetrics" amounts to is the application of statistics to baseball, which is obviously useful if you're at all careful, but also, due to the definition of statistics, cannot ever "predict every outcome." - Brock Hanke.
Until he gets the part where he moans about Verlander (he says various WARs have his 2011 season between 4 and 6 which is obviously wrong, though I think he'd make the same argument with 8 anyway).
Obviously plenty for us to moan about in this, but in the hopes of steering the discussion away from the same old...what are the best strategies for explaining what sabermetrics is to people who are unfamiliar? Either long-time baseball fans or people entirely new to baseball.
I recall having a very pleasant conversation with a friend (who isn't much of a sportsman of any kind) about NHL pythag records. It seemed to make immediate intuitive sense to him. I guess it all depends on how open-minded the person you're talking to is, how diplomatically the pair of you deal with differences in perception, and the sanity of your cause. I have some conspiracy theory friends whose facebook threads I sometimes read through for fun, and one of the common sentiments expressed is the frustration they have when friends and family don't see the obvious evidence of conspiracy before their eyes.
I'm not really sure what it is I'm getting at, but I'm at heart a very anti-confrontational person. So whenever there are disagreements over things (baseball) that clearly both parties hold very dear, I assume there is simply a breakdown in communication and if the two people could just express themselves clearly the world would magically be a better place.
hubpages.myinternet.freeblogs.cashforgold.info?
I mean, this:is pretty clear evidence that we're not linking to either a discussion worth having on the internet or an established sportswriter worth engaging due to his position.
I'm not much of a statistician, although I know applications of probability somewhat well, but I don't think that is a textbook definition of statistics. I don't want to elaborate because A) I might be wrong and B) the discussion would be hopelessly dry for most folks.
Much of the allure of baseball for me is that it is a drama of how the human responds under stressful conditions. Why did Mickey Hatcher go nuts in a World Series when he was a fairly average player otherwise? Don Larsen, similarly.
Because of the dicrete nature of the data of much of baseball (there is no case of being on 2.3nd base), it does lend itself to statistical analysis and has for more than a century. How long have people been keeping batting averages? So applying statistics to baseball is nothing new. However, I don't now if anyone would be writing in BBTF if they didn't appreciate the beauty of a great fielding play, a timely hit or strikeout.
Memo to ShootersCenter: Less trendy analysis, more proofreading.
As John McEnroe told the linesman, "You can't be serious!"
Damon's 2002-05 with the Sox had OPS+ 109,94,117,110, and his .316 BA his final Boston season was the best of the 4. Back to back down trend?
The Sox lost the LCS to the Rays in 2008, and IIRC they got a bit further than that in 2007.
This isn't misunderstanding, it's flat out lying. (Unless it's trolling and I bit.)
has got to be tongue in cheek. No one can be that obtuse by accident.
I mean, Sabreists are the one telling the old school guys that strikeouts don't matter that much, not vice versa.
I do sense a slight tendency from time to time among people who are into sabermetrics: to extrapolate and project with more certainty than is really warranted by the data. E.g. "bunting in that situation is absolutely stupid," "pinch-hitting Ramirez instead of Gonzalez is totally wrong," "That guy's a great prospect and is going to be a 5-WAR player for years," "that contract is stupid/brilliant," and the like. It's the dogmatism that is born of having a little more information than the other guy.
My explanatory analogy would be to handicapping races (though admittedly that won't work for every explainee). You can bet on your gut feeling about a jockey, or a horse's behavior in the paddock, or undigested information like the bare results of the last few races. But the more information that you add into the mix in a properly balanced way, the closer you will come to good predictions. That's why the betting favorites do win races a plurality of the time. OTOH, one must be very careful to stress that it's only a plurality; and longshots do win occasionally, just as Greg Gagne once hit a home run off Roger Clemens. But when your handicapping is informed by the data, it's exceedingly better than just knowing in your bones about some sure thing.
Though having said that, I think that the belief that intuition trumps data is extremely hard to overcome. "My mind is already made up, don't confuse me with the facts."
And those are some of the nicer ways MGL puts things.
My mother is a traditional diehard baseball fan. She loves and knows the game but at the same time she buys into the mythology of players very easily. While it's not exactly advanced stuff I've had a lot of luck with OPS and then OBP vs. BA with her. The OPS I've generally explained as a school grade; .900 is an A, .800 is a B, .700 is a C, not perfect, but it seems to hit home.
On OBP it's just a matter of reminding her of my little league games. If she always shouted "a walk is as good as a hit" why is that not still true? That helped.
When UZR was a big issue in Boston a couple of years ago (the Mike Cameron off-season) I focused on the idea of UZR coming from a study of every play, not just a few highlight plays and why would you evaluate something on less data rather than more.
In general with these concepts any success I have in getting them across to people not familiar with them comes from those sorts of baby steps. I try not to be condescending but simply to give a frame of reference. That always seems to help.
I think BDC's #19 is right that sometimes there is a certainty that gets conveyed in an arrogant manner to the detriment of the "cause". That arrogance exists with the old school folks too but that doesn't make it right.
it's very difficult to explain why stat guys are using some "replacement" non-existent player to measure others against - trust me, this I know because no one has yet explained to me in a way that makes any sense to me except that using the "average" ML player gives too many guys negative numbers, and why that's bad i don't get.
this writer has a point when he/she states that he/she can't understand why justin verlander would win only 6 more games/year than some crappy AAA pitcher on the exact same team - gotta say that.
as for OBP, i don't understand why on earth that stat does not include every single way that a batter reaches first, including reaching on error, seeing as how what you want to know is how many times of the total plate appearances, the batter arrived at first base safely.
but the "stats predict everything" is obviously total complete absolute nonsense. absolutely NO stat ever predicted that 5'5" jose altuve would so much as reach the major leagues, let alone outhit and out slug dustin ackley...
but anyhow, back to the point - WHY do these things keep getting written? what exactly has changed about the game so much that these writers keep complaining about stats? it's not exactly like they are hardly ever mentioned by radio guys or tv guys or the few baseball reporters who are still left. so what actually IS the problem?
First, as somebody noted upthread, BB-Ref says that Verlander had 8 WAR. But really, in this case, I think it's pretty easy to show. In 2011, the Tigers went 25-9 in games started by Verlander. At 8 WAR, that implies they would have gone 17-17 in those games if they were started by "some crappy AAA pitcher".
As it happens, the Tigers had two regular starters who were pretty close to "crappy". Rick Porcello put up an ERA+ of 87. In the 31 games he started, the Tigers went 21-10. Brad Penny put up an ERA+ of 79. In the 31 games he started, the Tigers went 16-15.
Those numbers all seem broadly consistent - the Tigers offense was so great that even crappy pitchers could win half their games, and Verlander was maybe 6-8 wins better than that.
Sounds like he's complaining more about writers and broadcasters talking about sabermetrics than teams using them. I've never heard anyone complain when team chemistry is good, hitters hit in the clutch, hearts beat regularly, or players work hard. It's when people talk about statistics rather than talking about these things that this guy gets annoyed. I don't see why he just doesn't stick to writers and broadcasters who talk about the stuff that he wants to hear about.
We Sabrists know that strikeouts matter for pitchers, representing a repeatable skill to get hitters out without any masking effects from the fielders. We know that they don't matter much for hitters, because they aren't worse than any other outs and don't negatively correlate with hits or slugging.
That's a tough distinction for the mainstream to understand, though. "Are strikeouts important?" "Depends on what you're talking about..." "Just tell me, ARE THEY IMPORTANT?"
I'm a non-mathy person and I get most sabermetric ideas. I can't prove them and I can't argue over them with the mathematically inclined, but I can understand most of them and see them in action on the field. Any open-minded fan can too. You just can't go about pushing them by announcing "Batting average is meaningless!" with no coherent explanation, as saber types sometimes did in the '90s.
Of course, your mother is a lot more likely to cut you some slack than most people. After all, she's still letting you live in her basement.
Barrels!
Strawmen!
GO!
Look, I'm not that great at math. I struggled to a a B- in my college calc 1 class, and that was with a lot of extra help from the TAs. I don't have a mind to see a bunch of symbolic language and process it quickly and easily.
That said, stats like DER, WAR, OPS etc aren't that complicated to understand. It takes about 5 minutes for a non-math type to wrap their head around linear regression, which most advanced baseball stats that are now mainstream are based on. UZR sounds complicated but then you read what it's comprised of and it makes perfect sense.
I played baseball in college, DIII level granted, from 2001 to 2004, it takes nothing away from my enjoyment of sabermetrics.
What if I want to know how many times the batter deserved to arrive at fist base safely? Why wouldn't I leave out ROE then? And maybe IBB too? And what about reaching on a fielder's choice? Batter arrived at first safely, but he also made an out.
Then I think you're moving beyond what OBP is. On-base percentage is a descriptive statistic: what percentage of the time did the batter reach base without making an out? And as a descriptive statistic, excluding ROE is a mistake (in my opinion). If you want to get into "deserves" then you would also have to look at how many times he hit a popup that fell between two fielders and was ruled a hit, and the ground balls "past a diving Jeter" that would have been a routine 6-3 groundout against two-thirds of all MLB shortstops, and at those balls that would have been a hit against most fielders, but the batter had the misfortune of having his great blast run down in the left-center field gap by Brett Gardner.
- The ability to avoid strikeouts,
- The ability to hit homeruns, and
- The ability to turn batted balls into hits, which in turn can be a function of several things - not making a lot of partial contact that leads to pop-ups or easy grounders, for instance, or being able to place-hit well a la Ichiro or Rod Carew
Avoiding strikeouts is *definitely* in that equation. In terms of runs expected, sure, they aren't significantly worse than any other out (I think back in the 80s BJ or maybe Palmer/Thorn said you'd lose around a run per 100 Ks - the loss of baserunner advancement just barely outweighs the fact that double plays are rare), but baseball players aren't collections of probabilities. They are people with differing attributes and skillsets. The inability to avoid being struck out has torpedoed a *lot* of careers - just on the Mariners it and it alone turns Trayvon Robinson into a AAA hitter, makes any chance that Carlos Peguero will be a major league hitter very small, and transforms Michael Saunders from a potential All-Star to a guy who might be average overall if everything lines up right.
In the specific case of Jax, the dingers outweigh the strikeouts, and on top of that during his prime he hit a lot of line drives (his BABIP was pretty consistently in the .300-.320 range throughout the 70s). Once he stopped being able to control his K rate (he moved from 19-22% up to 1981 to 25% and up from '82 to the end of his career) and he either lost a step or started hitting a lot more pop-ups (his BABIP dipped a lot at about the same time), he stopped being a good hitter. I just don't see how Reggie makes us have to re-evaluate the way modern sabremetrics grade players.
I can't even call this a straw man argument against an outdated form of sabremetrics, because as noted even this idea that Ks matter is a relatively recent introduction to the field (and mainly useful in player evaluation as opposed to base-out evaluation). Back in the 80s guys like Bill James were lionizing high-K guys who were denounced by traditionalists. James even created a fantasy team called the Ken Phelps All-Stars in part because of this sense. I remember getting into arguments with traditionalists about the relative value of Rob Deer.
I suspect that's the reason, but the numbers don't support the idea that ROE are randomly distributed. Some guys reach on errors a lot, others not so often. ROE should absolutely be a part of OBP, and it's a major flaw in OBP that they aren't.
IBB too.
Fielder's choice with an out? No, they shouldn't be included. I don't see any justification for that. But a FC without an out? I can see that. As Kiko said, the stat records what happened, not what should have.
Jobs. Nobody likes when the whippersnappers take over because all the existing old guys get pushed out. What these articles are really saying is "I'm still an expert and deserve to continue being treated as such." Which in many cases is true - but everyone gets their back up when jobs and money are at stake.
The idea that players are often only 1 or 2 win upgrades is intuitively clear once someone points out that the spread between the best and worst teams in the game is usually 40 wins or less -- there's no room in there for any but the most occasional superstar being worth 10 extra wins.
This is a good way of describing it. Another fallacy I believe is when "they" think of taking for example Albert Pujols away from a team, they think just about that loss as if no one would take his place and that everything he produced would be gone with no replacement. If the Cardinals hit 200 HRs last year now they will hit 165. In truth, someone would take Albert's place - someone is taking his place - that someone just wouldn't be as good. Instead of hitting 35 homers in 650 PAs that person might hit 20. It's not as good as 35 ... but 20 is something and would help win a few games. Just not as many as Albert's 35. The player may drive in 70 runs instead of 100. When you get it down to more manageable numbers, 15 homeruns or 30 RBIs (and then you start to consider how many of those HRs or RBIs are game changers), then it starts make more sense how a superstar could be worth "only" 6-8 more wins than their replacement.
Or, to put a broader point on it, the Cardinals won 90 games last year. They then lost Albert Pujols in the offseason. A baseball fan as a result doesn't expect them to win only 65 games this year. We expect them to win 85 (I have no idea actually what the Cards were expected to win this year - just an example). People know this stuff, they just don't connect it that way.
Then you're just assuming that the errors were somehow forced by something the batter did. It's still a descriptive stat whether the ROE are in or out; you can just as easily argue that counting them the same as hits or walks is a mistake since AFAIK most fielding errors are not, in fact, forced.
I am aware of studies that show the existence of both high ROE and low ROE players, but I was not aware of anything that purports to show that the majority of players aren't pretty similar in ROE. And I would think that it would have to be true that a large majority of players have a repeatable skill level at ROE for it to be a "major" flaw in OBP.
For a fan who knows nothing other than the basic, "traditional" stuff about baseball statistics, think about what it takes to hop on that train:
1) Hear things that often sound like esoteric gibberish to you, and care enough to try to learn and understand what it means.
2) Locate a good explanation/explainer, and actually learn/understand.
3) Admit to yourself that there are some important things that you either didn't know or that were just wrong--about a seemingly fairly simple game you may have followed for decades and thought yourself pretty knowledgeable about.
3) a) Also overcome the fact that there are probably quite a few people (players, sportswriters, other fans) whose opinions you put some stock in who call all these newfangled stats a bunch of malarkey.
It's daunting I'd say if you don't happen to have the right mental approach that helps you be open to it. For a longtime fan, I'd imagine it's like a milder version of questioning your religion.
I don't think anyone is saying it's a "major" flaw, just a flaw.
I actually love it when old guys get pushed out of jobs by better young whippersnappers. It means the world is progressing. I really dislike old people who refuse to adopt better ways of doing things. I respect guys like Joe Maddon who is old, but still thinks he has something to learn....much more than the guy who wrote this article.
Not at all. A descriptive stat makes no assumption whatsoever about whether the batter was the "cause" of something or not. It merely describes what happened. And on an ROE, what happened was that a plate appearance ended with no outs being recorded and the batter being safely on base.
I've got to say that a *lot* of the work done over the past 10-15 years has had the effect of putting statistical analysis much in line with conventional wisdom, particularly old-timey, pre-statty CW. That whole "clutch" thing that so-called anti-statheads trot out for instance is more of an outgrowth IMO of the over-investment of the importance of a particular statistic (RBIs) and the cognitive dissonance that comes about when a guy with otherwise blah numbers (hi Joe Carter!) gets a lot of them because he hits behind players who get on base a lot. Or it's about guys who happened to have performed well during a short period of time while on the national stage. Modern sabremetricians understand small sample size; what these "anti-stat" guys are doing in this case is *not* discounting statistics which ought to be looked at with a jaundiced eye.
Basically, because it wipes out playing time. A RP who pitches 40 innings at a slightly above average rate, is not more valuable than a SP who pitches 220 at slightly below average rate. But if you used average, that's the result you would get.
except SoSH:
...
58 is old? Damn!
I think that the word "major" here is an over-reach, but I do think that hit/error distinctions are generally not terribly helpful. As a descriptor of what happened, if the batter reaches first base safely, I see no purpose in distinguishing how he got there. And as a measure of what a batter deserves "credit" for, the hit/error distinction is merely one of many such judgments that one might want to make - poorly fielded balls that aren't ruled errors (popups that fall between fielders, "past a diving Jeter"), great fielding plays that rob batters of hits that they otherwise "deserved", even long fly balls that are/are not home runs depending on the particular dimensions of the field on which they were hit. Depending on what I'm doing, I would think that I should either care about all of these things or none of these things; limiting my "caring" to only hits v. errors seems wrong to me in both cases.
Damn right. You have players such as Derek Jeter and Ichiro who reach on errors quite often. Even if it's not always a direct skill, it is undeniably an indirect skill* and a year-in, year-out part of their game. And a stat that's sole function is to record how frequently a ballplayer reaches base that excludes, for no consistent reason**, one way players reach base is flawed in a major way. It's a bigger flaw than BA not counting sac flies as at bats, and that one pisses me off too.
It would be a minor flaw even if there were no evidence that some players reached on errors at greater rates than other players, because it would still be ignoring what it purports to measure. That it fails to measure something where differences in rates are obvious and consistent elevates it into major flaw status.
* By indirect skill, I mean that the way they hit/play (not a lot of Ks, hit it on the ground a lot, good speed and a willingness to use it) is why they get a lot of ROE.
** If you're going to argue that ROE shouldn't be a part of OBP, then the same should be true of Hit Batsmen. In both cases, the ultimate fault rests with the player not in the batter's box. And in both cases, the player inside the batter's box actually plays a large role in the frequency of the stat's occurrence.
Are you arguing to get rid of the distinction altogether, or just for the purposes of OBP? If it's the former, that's one of the few areas where you are miles apart.
I was just responding to Ivan's missing the fact that I was directly quoting that word, but I don't want to make a "major" deal out of it.
I guess I come down on the side of caring about all of them, while acknowledging that this certainly would complicate things and take OBP out of the realm of being descriptive. I just think that errors are rare enough that we don't need to get all that worked up about them. MLB fielding %age was .983 in 2011. ROE is less than 1% of PA.
Well, I suppose that's true. There were only 1274 sac flies in MLB last year, while there were 1816 ROEs. I guess I will now concede that you probably didn't mean "major" in a quantitative sense.
The league leaders in ROE were all over the map. Yeah, you had guys like Andrus and Ichiro but guys like Michael Cuddyer, Carlos Lee and Aubrey Huff were there too. I'm pretty sure James studied this in some detail and found that there was not a strong relationship between speed and ROE but I won't swear to it.
I don't think it's that tough to identify a great many players for whom the hit by pitch IS a "skill." There have been more than a few players who regularly got hit. Some (Youkilis for example) seem to just be bad at getting out of the way while others (Shoppach for example) seem all too willing to just let the ball plunk off of them.
I can see some relevance to distinguishing between a player who has no range but is extremely sure-handed with what he gets to (e.g., Derek Jeter) vs. a player with exceptional range but who is prone to botching routine plays (e.g., Starlin Castro) as a scouting/descriptive tool to describe players.
But in terms of evaluating the relative fielding values of Jeter v. Castro, ultimately, it seems to me that in both cases, it still boils down to how many outs did they make versus how many non-outs did they allow that they shouldn't have - either because Castro threw the ball away or because Jeter couldn't move two steps to his right to get to the ball.
For pitchers, I'd also rather start with total runs allowed (RA) and when it comes time to adjust for defense, again, the issue is how many extra base runners did the fielders behind the pitcher prevent/allow compared to average - but, again, allowing extra runners because of fielding errors or because of lousy range have the same effect.
EDIT: I'll add that I appreciate baseball history enough that I wouldn't want to necessarily lose the concept of "errors", and I can appreciate a "no-hitter" that included one or more fielding errors as still being worth celebrating as a no-hitter. And if "range" vs. "error-proneness" project differently in terms of projecting players' future fielding, that's worth taking it account. But beyond that, I just don't see a lot of practical places where I'm inclined to distinguish between errors and hits.
That's just one potential factor. The point is, ROE numbers aren't completely random, with one batter equally likely to reach base via error as the next guy. It is dependent on a number of factors (handedness, ballpark, GB/FB tendency, speed, hustle). In that way, it's similar to how HBP is not random. They shouldn't be treated differently.
No, it has nothing to do with the frequency of occasion, but more my desire to see these stats measure what they're designed to measure.
OBP's flaw is more meaningful for a few reasons. I think there's less randomness in ROE than in sac flies (just a guess), and you could make an argument that guys really are trying to hit fly balls to the outfield with a runner on third (I don't buy it, but I can't disprove it either).
Also, because the first involves the numerator, and the latter affects the denominator. So ignoring ROE as reaching base has a larger result on the stat itself than pretending sac flies should not be considered at bats.
That's why I asked if you want to get rid of it altogether. If all you're using the stat for is to asses value, that's one thing. But stats are more than just tools to measure value. They tell us what happened. Baseball is a lesser place if Billy Buckner's boot or Luis Castillo's drop is simply scored as Ball In Play, Runner Reached Safely.
I can agree with that (as I hope my edit made clear).
Sorry. I hadn't seen the edit.
If nothing else, the reading of the linescore at the end of each half inning would only be 2/3 as good if we got rid of errors.
3/4 as good
"no runs, one hit, no errors, one man left"
Wrong thread...human error is so repulsive.
I agree - I'm just saying that's the mindset. I find it a sign of weak character when people continue to cling to things that aren't true, that deep inside they probably know aren't true, just because it's easier that way or on the flipside, hard to admit you might have been wrong. When those people then flock to anti-intellectualism as many in the anti-stats crowd do, it becomes disgusting. And dangerous, beyond the scope of baseball.
Of course, we'll see what we all think when we're old guys and the new research shows that throwing strikes is actually bad.
in the crumbled financial institutions of this land
(there will be snacks)
There are at least a couple of reasons why using zero for the average player doesn't work as well as using a "replacement player" (I used quotes because there is no real consensus as to what the replacement rate actually is, except that it's well below average).
1. The negative numbers cause more problems when you look at a career than they do when you're only looking at one year. If you're just comparing guys within one year, it doesn't matter much where you put the zero point. But when you're looking at careers, and you use zero to represent the average, you're going to find many, many players whose career values are LESS than their values were five years before they retired, which makes no sense. That's because they played those last five years as below average. Replacement rate gets that zero point down to where very very few careers end up with a run of negative numbers because very very few players can get away with five years playing below replacement rate. This is one of the problems that plague Pete Palmer's career numbers in Total Baseball. Check it out. You'll see a lot of players whose career values maxed out five years or so before they retired. That's the main reason that too many guys with negative numbers is bad, if you're an analyst. You get the career values wrong.
2. The whole concept of replacement player is designed to identify who can actually play at the major league level and who can't. That's why there are many different replacement rates. Analysts differ as to where that in/out line is. However, the idea is that, if a player has a positive number, then he's a MLB player. Maybe not much of one, but a MLB player. If the number is negative, then he's not really a MLB player and, which is the most important thing, YOU CAN FIND A BETTER PLAYER WITHOUT SERIOUS SACRIFICE. That's why the term is "replacement." The idea is that, if you've got a guy below that level, you can bring up someone from your AAA team, or make a very cheap trade, and get someone who is better. You can replace your below-zero guy, and you can replace him cheap. This, if you're running a ballclub, is obviously a very useful thing to know. Replacement rate analysis helps you know that.
I don't know how much math you have in your background, but here's the point stated in more mathey terms: The distribution of major league players is NOT the normal bell curve, where there's a big hump in the middle and the same number of people above that hump as below it. In fact, most major league players are below average. The best players are further above average than the weakest ones are below it, but there are more weak players than strong ones. If you try to use the average as your zero point, you will find out that there are no teams who have all-positive-number players. Even the best teams have SOME guys who are below average. The replacement rate is supposed to be the point at which you CAN fill out a team with positive number players. If you visualize math curves well, you can think of the distribution of MLB players as looking like a Hershey's kiss candy, with a big bulge at the bottom and a quick narrowing rise to a point. If you draw a line halfway down that distribution, you will end up with a lot more candy (players) beneath it than above it. Replacement rate is designed to be right underneath that big bulge in the kiss candy.
I hope that helps, - Brock
The irony of course is that, IMO, James is one of the most interesting "human" writers out there yet everyone thinks he's some nerdy guy who just punches a calculator while watching games.
I read a bunch of literature on handicapping and the more I read, the more confused I became. Then I got married and just wasn't into horse racing as much. I never got to the point of reading trainer stats, but I think they may hold a key to the game.
It is true that everybody's got a pet theory. In a way, handicapping races is analogous to defensive evaluation in baseball. Overall, they point in a consistent direction, but once in a while there's a severe discrepancy in the way two analysts will evaluate a horse race or a shortstop.
guys like Joe Maddon who is old
Maddon is five years older than I am. Great, now I know how close I am to geezerhood :)
James didn't have much data to work with when he first looked at the issue, but even so he noted that it was right-handed line drive hitters who seemed to do best. And came up with a reasonable theory as to why. Most errors come on hard hit balls. And on a hard hit ball to first, the first-baseman may still be able to make the play after bobbling the ball. Much less likely for a third-baseman.
Ah but my respect for you is now immense
Well, if you took the theoretical average player that player wouldn't exist either, as there is no player who is exactly average. At some point, no matter what level of player you use, you're going to have to create an averaged out barrier where above is good and below is bad, and that barrier isn't going to represent a single particular player because no player is going to have those exact stats, or if he does not for long.
The best way (I think) to think about it is to compare it the number zero. Replacement is simply the zero point. Above zero equals worthy of playing in MLB, i.e. one of the top 750 or whatever players in the world, even if not well. Below the barrier equals not worthy, and can go as low as you want. The barrier in our case is slanted - there are far, far, far more baseball players in the world who are not capable of playing in MLB than that are, so it makes sense logically for that to be the zero point - those above contribute to winning games in MLB, no matter how small that contribution, those below don't.
FWIW the number zero also doesn't exist and represents nothing, and we don't have any trouble comparing other numbers to it. We have lots of numbers that actually represent something - why not compare against one of those? That's more of a semantic point though.
What are you, a nihilist or something?
Dang, that was David Murphy's only claim to fame.
OPS+=100, WAR fielding runs=0, minimum 2000 PA
Player WAR/pos OPS+ Rfield PA From To BA OBP SLGTom Brown 13.3 100 0 8206 1882 1898 .265 .336 .361
FPH says: "Headshot."
You must be Registered and Logged In to post comments.
<< Back to main