The WAR in the (2) is simply a combination of the details stated prior to the summary.

WAR is nothing other than the articulation of a set of value stats which have been expressed via a single unit of measure, to enable their combination. If you don't object to that, you don't object to WAR.

Tom - Can I ask for your response to #91?

How do I choose who to sign without evaluating their contributions in some common unit of measure?

What does your team look like? Who is your CF? Does he cover a lot of ground defensively to make up for Sluggy's sloth? What kind of pitching staff do you have? Are they primarily ground ballers or do they rely on their OF's to chase down balls in the gaps? What does your farm system look like? Is there a reasonable defensive replacement OF in AAA, not much of a prospect but good enough to fill in for Sluggy when you have a late lead, or is your best young player a hammer of a 1B who could PH late and hit a few homers if you need them?

None of those bits of knowledge have anything to do with WAR.

WAR is nothing other than the articulation of a set of value stats which have been expressed via a single unit of measure, to enable their combination. If you don't object to that, you don't object to WAR.

Then let me be clear; I don't object to (2) because you've provided the more useful, worthwhile data, so the oversimplification of the combined display can be properly ignored.

Those are all useful bits of knowledge.

However, if you can't compare Sluggy and Speedy, you don't know how wide the difference between them is, and how much consideration you need to give to team construction as compared to player quality. Getting the right player is always a matter of balancing the need to get the best player with the need to get the player who fits the team the best. You are right that we need to consider the latter problem, but you can't ignore the former problem.

It's neither the first nor the last time a guy with less stat-nerd chachet loses and MVP race.

Trout didn't just have more stat-nerd cachet. He had a better year.

Then let me be clear; I don't object to (2) because you've provided the more useful, worthwhile data, so the oversimplification of the combined display can be properly ignored.

You aren't actually ignoring it. If you look at 10 + 2 and don't see "12" as well as 10 + 2, you are probably illiterate. It is important, in comparing a 10 + 2 player to a 7 + 3 player, to compare 10 and 7, to compare 2 and 3, and to compare 12 and 10.

You are right that we need to consider the latter problem, but you can't ignore the former problem.

I don't suggest that we should ignore the former problem. I simply think that trying to shoehorn the former problem into a single value number like WAR so oversimplifies the math that it becomes meaningless. You have to value and evaluate the player. You don't need to cram it into a single number; ever. And cramming it into a single number creates false confidence and false precision, such that eventually you have folks arguing that a 1 WAR difference between disparate players on drastically differing rosters is the difference between MVP yeses or nos.

I don't suggest that we should ignore the former problem. I simply think that trying to shoehorn the former problem into a single value number like WAR so oversimplifies the math that it becomes meaningless.

You need the methodology that can produce the "single value number" in order to do the comparison.

A methodology that can't produce a single value number also can't do the comparison.

Just to be clear. Would you object to one, two, none, or all of these sentences?

1) I project Nick Swisher to +27 RAR
2) I project Nick Swisher to +15 Bat, -2 Run, +21 Rep, -8 Pos, +1 Def, +27 RAR
3) I project Nick Swisher to +15 Bat, -2 Run, +21 Rep, -8 Pos, +1 Def

I don't object to any of it; you may project Nick Swisher to do anything you like. Just don't pretend it's anything other than estimates built on estimates. As to preferring one over the other, as Sam said, the granularity is far more preferable than the blunt tool. (I've also noticed that you've moved away from using WAR at all, which is a good thing; the fact that this thing is sold as telling us how many wins each player is responsible for is a complete oversell.)

I still don't see what it has to do with front office decisions. The choice for the Indians is not to sign Nick Swisher at +27 RAR for one year as opposed to Outfielder X at +21 RAR for one year. The choice is to sign Nick Swisher for four years, or trade a couple of AA ball pitchers for Denard Span on a short-term contract, or plug in some kid from AAA, or any number of other things. Coming up with a very rough estimate of Nick Swisher's value for 2013 is only a small part of that equation.

That's my basic objection to the whole concept: Baseball is really complicated. It's too complicated to reduce each player to a single number and think that answers any important questions.

Trout didn't just have more stat-nerd cachet. He had a better year.

Sure. And Miggy won the triple crown. Post-season awards aren't always about statistical performance.

You need the methodology that can produce the "single value number" in order to do the comparison.

We're talking past one another at this point. I believe I've made my position pretty clear.

Slash stats allow me to assess how a player I don't see play very often contributes to his team (or fails to)

Quick, which player was the better offensive player?

a) 326/399/564/83 RBIs

Or

b) 333/393/606/139 RBIs?

Wait, player A hit leadoff a lot while stealing 49/54 bases and grounding into only 7 double plays, while player B hit cleanup and was 4/5 on steals. while grounding into 28 double plays. Player A played in a pitchers park in a tougher division while player played in a hitters park in a weaker division.

Still maintain that slash lines are all the meaningful data you need to evaluate players you haven't seen?

but we don't need them if we see the players. If you watched Mike Trout play baseball on a daily basis, you wouldn't need his slash line or his zone ratings to ascertain his value. Those stats are most useful for general managers looking to evaluate potential roster additions more than anything else.

Who watches any player on a daily basis? Can you really watch all of Mike Trouts games and Miguel Cabreras games and know which one is the more valuable offensive player? You couldn't tell us who had the higher OBP without recording data, and likely couldn't come close to guessing ther batting or slugging averages.

Through out this thread you continuously use components of WAR as measures, you clearly believe in measuring offensive and defensive contributions and estimating heir impact in wins for their team. You just don't like summing those contributions in a singe number because it hides data. So what? So do BA, RBIs, pitchers Wins, ERA, etc, every single one is an imperfect summary of a player skill. The difference is that WAR hides less data and has far fewer imperfections than all of the standard stats.

I mean you and the author don't even believe that the vast majority of pitchers have little control over their BABIP, a concept that is about as solidly proven as gravity and evolution by this point (and easily disprovable if it wasn't so true). You and the author should create your own baseball Truther society to deny all sabermetric principles and studies, when you have free time away from attempts to prove man walked with dinosaurs and NASA faked the moon landings.

I don't object to any of it; you may project Nick Swisher to do anything you like. Just don't pretend it's anything other than estimates built on estimates.

This is unresponsive, and your anti-intellectual tactic of repeating basic facts about statistical methods with a sneer as if that is somehow an argument is seriously bothersome. Anyone with any understanding of statistics knows when they are working with estimates. If you want to object to modern social science in its entirety because it uses "estimates", I guess I can't stop you, but I wish you'd stop arguing like you're in front of an audience of idiots who don't understand how statistics work.

The choice is to sign Nick Swisher for four years, or trade a couple of AA ball pitchers for Denard Span on a short-term contract, or plug in some kid from AAA, or any number of other things. Coming up with a very rough estimate of Nick Swisher's value for 2013 is only a small part of that equation.

It is indeed. But it is also a necessary part of that equation, and previously you were saying it was useless. I agree with the quoted text in its entirety, and you have completely shifted your position from unreasonable to reasonable.

Here's a thought: we don't actually need slash stats to know Cabrera could rake or that Trout was an outstanding five tool player. Slash stats allow me to assess how a player I don't see play very often contributes to his team (or fails to) but we don't need them if we see the players.

Yeah, this is nutty. We can often get value from subjective observations, but the idea that we can consistently and confidently evaluate player value without numbers is a load of crap.

At the end of the day, I think this is what drives most hardcore WAR believers these days; the terrible, horrendous, crime against humanity that happened when the voters awarded Miguel Cabrera the 2012 MVP because he was a triple crown winner, over the beloved youngster, Mike Trout. Good god, son. Get over it. It's neither the first nor the last time a guy with less stat-nerd chachet loses and MVP race.

Why? It so wonderfully captures the detachment of sportswriters and people like you from reality. Especially the phrase "hard core believers". LOL. If Trout had lost to Cano, it wouldn't have been so beautiful. I mean we all knw how imprecise WAR is, so you can make the case that Trout might not have been that good defensively, etc. and Cano had certain intangibles that could narrow a 20% gap.

But Cabrera trailing by so much? LOL, its should be the Most Narrative Award (and still a travesty).

WAR exposed Miggy for being the out maker, rally killer, gloved butcher he was. He produced the highest most shiny stats, but gave back a bunch of runs in much less obvious ways,, but WAR caught him.

Trout also had shiny stats, but in the tougher to see and appreciate skills he created many more runs than Miggy, by doing little things that WAR measured and valued.

You and the dumb sportswriters just got up to get another beer everything Miggy grounded into two outs.

Based on my observations of a dozen or so 2004 Red Barons games, Jim Rushford was a five-tool player who would have been another Frank Thomas if given half a chance.

Here's a thought: we don't actually need slash stats to know Cabrera could rake or that Trout was an outstanding five tool player. Slash stats allow me to assess how a player I don't see play very often contributes to his team (or fails to) but we don't need them if we see the players.

Yeah, this is nutty. We can often get value from subjective observations, but the idea that we can consistently and confidently evaluate player value without numbers is a load of crap.

Lets say you watched every Tigers game. Without tracing stats, could you really know that Cabrera was a better hitter than Fielder? The difference between Miggy (.330/44/139) and Prince (.313/30/108) was approx 1 HR every other week*. Without tracking stats, can you really tell the difference?

*Turn 13 Prince outs into HR and his slash line is .336/43/128, the last number assumes 1.5 RBI per extra HR.

I don't suggest that we should ignore the former problem. I simply think that trying to shoehorn the former problem into a single value number like WAR so oversimplifies the math that it becomes meaningless. You have to value and evaluate the player. You don't need to cram it into a single number; ever. And cramming it into a single number creates false confidence and false precision, such that eventually you have folks arguing that a 1 WAR difference between disparate players on drastically differing rosters is the difference between MVP yeses or nos.

If I change the numbers and make Sluggy a 150 OPS+ hitter, he becomes so radically better than Speedy that all you need is the single number.

WAR can't tell you if a 4 WAR player is better than a 3.5 WAR players, but it sure as hell can tell you that a 4 WAR player is better than a 1 WAR player.

Having a contextual framework, and a common unit of measurement helps you evaluate all the other factors (when necessary) by suggesting what orders of magnitude you're dealing with.

Has anyone measured whether pro athletes really experience this in pressure situations?

It can be done. Some NFL and soccer teams put heart monitors on their players during practices and games.

Who watches any player on a daily basis?

Baseball beat writers.

This is unresponsive, and your anti-intellectual tactic of repeating basic facts about statistical methods with a sneer as if that is somehow an argument is seriously bothersome. Anyone with any understanding of statistics knows when they are working with estimates. If you want to object to modern social science in its entirety because it uses "estimates", I guess I can't stop you, but I wish you'd stop arguing like you're in front of an audience of idiots who don't understand how statistics work.

I am genuinely sorry you feel that way. I certainly didn't intend my answer to be insulting.

But the fact that we are talking about estimates seems to be the whole crux of the matter, isn't it? You think you can take estimates of the values of Nick Swisher's past seasons with regard to hitting, fielding, baserunning, positional value and whatnot, add them together, apply some forward thinking, and come up with a reasonable estimate of his value for 2013. You ask me if I object to various ways of presenting it, and my answer is: No, I have no objection at all, because I don't think it has a whole lot to do with reality. It's all guesswork.

Look, Bbref thinks Swisher was five runs better than average on defense in 2012, and six runs below average in 2011. Fangraphs think he was 3.7 runs above average in 2012 and 6.9 runs above average in 2011. And you think we can come up with a reasonable estimate for the value of his defense in 2013? Those guys know way more about baseball than I do, and they can't figure out if Swisher is getting worse or getting better. So why should I have an opinion about the best way to present that information?

I don't think it's anti-intellectual to remind people that we really don't know all that much about the accuracy of these numbers. I don't think it's anti-intellectual to acknowledge the limits of our intellect.

You and the dumb sportswriters just got up to get another beer everything Miggy grounded into two outs.

You are an arrogant son of a #####, huh? The least you could do is perhaps read for comprehension before spewing out at the mouth like this.

I have stated already that Trout was the better player. Dumbass.

I don't think it's anti-intellectual to remind people that we really don't know all that much about the accuracy of these numbers. I don't think it's anti-intellectual to acknowledge the limits of our intellect.

Absolutely I agree.

You ask me if I object to various ways of presenting it, and my answer is: No, I have no objection at all, because I don't think it has a whole lot to do with reality. It's all guesswork.

This is the thing I was objecting to. "Estimate" != "guesswork".

Estimates are logically derived, rigorously argued attempts to model reality. Reducing them to guesswork is both wrong as a matter of description and disrespectful to the people who've put in the work to create these estimates.

Now, estimates can be very, very wrong. I don't mean all estimates are good estimates. If you have a specific argument regarding a specific estimate or a specific method of estimating, that would be a useful contribution. Objecting to the entire process of estimating, that's not.

Look, Bbref thinks Swisher was five runs better than average on defense in 2012, and six runs below average in 2011. Fangraphs think he was 3.7 runs above average in 2012 and 6.9 runs above average in 2011. And you think we can come up with a reasonable estimate for the value of his defense in 2013?

Yes! Those numbers are very, very close to each other. We're looking at a range of about 10 runs. He's probably within 5 runs of average. That's super useful for estimating his defensive value.

Yes! Those numbers are very, very close to each other. We're looking at a range of about 10 runs. He's probably within 5 runs of average. That's super useful for estimating his defensive value.

Exactly. Those numbers tell you he's probably averagish. Most years we'll expect him to be +/- 5.

Just b/c estimates have an error range, doesn't make them guesses. Our offensive projections have error ranges too.

All projections of value should incorporate error ranges. Many of them do. The CAIRO projections explicitly state various percentiles of performance.

For me the useful thing about this article was to discover that WAR believes that Rick Reuschel > Bob Feller. Nor sure what to make of that. (Yes, I know that WAR credits no war credits, but still.)

Yes! Those numbers are very, very close to each other. We're looking at a range of about 10 runs. He's probably within 5 runs of average.

Break out the champagne folks, we have a possible value with of +/- 5 "runs." How could we have ever evaluated Nick Swisher's defense otherwise?

Whats the standard error in runs on offense? Pitching? Its not that far from 5.
The purpose of war is to put offense, defense (including positional adjustment), and baserunning on the same scale.

94:

The idea of converting value to a single unit of measurement and comparing player value between different categories based on these units is not something that was active in old-time baseball. It's reasonably new.

I disagree. People have been doing this since the first argument about which of two players was better. It was just done in a more haphazard way.

I disagree. People have been doing this since the first argument about which of two players was better. It was just done in a more haphazard way.

Not really, no. It's still just as haphazard as before. But now we want to pretend it's not. That's the whole damned problem.

Heart rates: They used to do this on some poker broadcasts. Not athletes obviously but high pressure situations. Not that I watched a lot (much less have the data) but it seemed the top guy's heart rates still went up an appreciable amount.

I'm sure cycling teams, etc. do this but they don't really have many high pressure situations -- they have work at insane levels for bursts of time that of course send heart rates through the roof (e.g. sprints). Those moments coincide with high pressure situations but they're not being asked to come through in the clutch exactly they're being asked to work harder than the other sprinters.

I disagree. People have been doing this since the first argument about which of two players was better. It was just done in a more haphazard way.

It was mainly done in a useless way.

"Player A's defense and baserunning make up for the hitting gap with B."

"There's no way his defense could be good enough to make up a gap of (20 points of OPS+, 60 points of OPS, 30 points of BA, 12 HR, 39 RBI or whatever you feel like plugging in)."

"Can too."

"Can not."

"Idiot."

"Moron."

At least now we can get to the point of

"See, it's perfectly reasonable to think A's defense and baserunning make up the gap, WAR says so."

"That just proves that WAR's defense and baserunning numbers are wrong."

"Does not."

"Does too."

"Idiot."

"Moron."

It's called the advance of science.

RBI needs a rate component to be useful, but everyone who has put it into a rate component form, has screwed it up by being too technical about it and making it a non-popular or incomprehensible number

CFB, you wrote this on page 1. BP's OBI% is simpy the percentage of all runners on base batted in.

Whats so incomprehensible about that ?

Here is the top 20, Min 300 PA's (258 PLayers in all) LINK


Hamilton 22.2%
Cabrera 21.4
Gonzalez 21.1 (Boston PA's only)
Hunter 20.8
Lee 20.5
Nelson 20.3
Colvin 20.2
Ruiz 20
Headley 19.9
Craig 19.8
Montero 19.7
Lucroy 19.4
Soriano 19.4
Longoria 19.4
Arencibia 19.4
Inge 19.2
Butler 18.9
C. Gonzalez 18.7
A Ramiriez 18.7
Posey 18.4

There are quite a number of guys on this top 20 list that I would not put anywhere near the top 20 most valuable hitters, top 20 "best" hitters, or even top 20 "best RBI men"....yet here it is, in a very simple, comprehensible stat.


Other players of interest

42nd Trout 17.3


The Yankees were especially bad as a team

The BEST Yankee, Swisher, clocked in at 16.5 %, 64th overall

Tex 15.5% 95th
Granderson 14.9% 118th
Cano 12.8% 184th
Jeter 12.6% 188th
A Rod 11.0 230th
Martin 10.6% 236th

Thats stunning really.....if just one or two off those guys cracked the top 50 their offense would have looked a bit different perhaps.

And of course...Jeff Franceour, 9.6%, 246th out of 258 with 300 PA or more.

CFB, you wrote this on page 1. BP's OBI% is simpy the percentage of all runners on base batted in.

Whats so incomprehensible about that ?

Does it account for players who had an unusual percentage of their runners on base on first base? Or those who saw a lot of runners actually in scoring position? Is it adjusted for ballparks? Does it account for the baserunning ability of the runners on base?

It's not nearly as illustrative of pure "RBI skills", as you pretend it is.

There are quite a number of guys on this top 20 list that I would not put anywhere near the top 20 most valuable hitters, top 20 "best" hitters, or even top 20 "best RBI men"....yet here it is, in a very simple, comprehensible stat.

And those guys weren't on the list last year. And probably won't be on the list next year. OBI% is prone to huge year-to-year variations, with little rhyme or reason. But over time, the best hitters tend to end up at the top. There is really nothing there that you aren't better off figuring out from raw hitting stats.

If you're measuring a player's value for a season, aren't players who hit better when there's the opportunity to drive in somebody providing value beyond their raw hitting stats?

Does it account for players who had an unusual percentage of their runners on base on first base? Or those who saw a lot of runners actually in scoring position? Is it adjusted for ballparks? Does it account for the baserunning ability of the runners on base? 

BP reports the runners on base and scored by base for each batter.

It is not RBI%+. That does not make it useless .

If you're measuring a player's value for a season, aren't players who hit better when there's the opportunity to drive in somebody providing value beyond their raw hitting stats?

If you want to measure situational value, just go with WPA. RBI and OBI% really add nothing of value that isn't already captured there. I mean it's still mostly useless, but at least it's comprehensive.

It is not RBI%+. That does not make it useless .

No, being a glorified random number generator takes care of that.

No, being a glorified random number generator takes care of that.

The sport of baseball is a glorified random number generator.

No, being a glorified random number generator takes care of that.

The sport of baseball is a glorified random number generator.

http://xkcd.com/904/

WAR is just wins.

This is what people mean by arrogance. War is a structure that attempts to qualify the totality of a players performance in a wins like system. To claim it is wins, is ridiculous, the accuracy is not even close to being there, and outside of Win Shares, none of the systems even start with wins as the baseline.

Nerds didn't work for years to custom design value measures in WAR to show Mike Trout in the best possible light and expose all of Miggys flaws. They created the best possible value measures, and those measures said Trout was great.

It also exposed the flaws of War. To think for a second that there is a 4 win difference between those two performances is nuts. Yes Trout was the Most valuable player last year, but it wasn't a 4 win difference.

You can reject defensive measurements that said Trout was otherworldly and Cabrera bad, and Trout still crushes Miggy.

Not really, to get the "crush" you have to accept the positional adjustments applied to the two, the park factors that says that Anaheim was an unwordly great pitchers park, and that Tiger stadium was a hitters park. Etc.... The difference between the two was ultimately in the hard to see areas(Park adjustments, the fact that there is no positive benefits for playing everyday with war if you are average or less defensively, and positional adjustments) As it stands, Miggy put up the better offense over the course of the 162 game season, Trout just gains ground by better defense, park adjustments, positional adjustments and the silliness of a system that thinks there is no benefit to a guy who puts on a uniform 20+ times over the other guy.

Trout should have been the MVP, but war believers that push this 4 win difference is the ones who need to take a step back.

CFB, you wrote this on page 1. BP's OBI% is simpy the percentage of all runners on base batted in.

It doesn't tell the full story, and the masses don't care for it. I would have no problem with it, provided it doesn't try any weird ####, like using the expected run matrix and comparing that player to average and coming up with a decimal number that doesn't really give any 'meaning' to a casual fan. On top of that, for some reason obi ignores homeruns, and I really, really hate any stat that use RBI and ignores homeruns (absolute dumbest stat ever created is the Rbi+Runs-HR stat that some people like to tout out. This makes productive outs to be the greatest stat ever invented in comparison)

Eventually someone is going to come up with the proper rbi stat, which is to assign .25 bases to every remaining base of the runners on, including a full 1 for the batter, and then track batters advanced, batters scored, and total opportunities. Until that day arrives you are going to have the stat nerds doing everything with the expected run scoring matrix, and not one normal person giving a rats ass about it.

Does it account for players who had an unusual percentage of their runners on base on first base? Or those who saw a lot of runners actually in scoring position? Is it adjusted for ballparks? Does it account for the baserunning ability of the runners on base?

You can't expect everything, no stat in existence accounts for baserunning ability of runners on base. In fact no stat accounts for any variable involving the quality of players involved. That is one of the big flaws with WPA and the theoretical break even point of sb/sb% it's based upon assuming everyone else is average.

That link he gave has obi for men on first, second and third along with overall percentage so it has some of the information you are asking for if you look at it close enough.

If you want to measure situational value, just go with WPA.

If you want to act like you have had half your brain removed, use WPA for anything. It's gotta be on the short list of Worst Stat ever created, while grounded in a basic logical framework(it's a toss up between this or productive outs)

the fact that there is no positive benefits for playing everyday with war if you are average or less defensively

This is completely untrue. WAR is a counting stat. As long as the totality of your contributions is above replacement level, playing more games will increase your WAR total. If Cabrera had played 15% less, at the same rate he did, he'd have 15% less WAR.

Maybe you shouldn't criticize a system you clearly don't fully understand.

If you want to act like you have had half your brain removed, use WPA for anything. It's gotta be on the short list of Worst Stat ever created

Oh I agree. And it still runs circles around OBI%, which is really all you need to know about OBI%.

the fact that there is no positive benefits for playing everyday with war if you are average or less defensively

This is the opposite of a fact.

This is completely untrue. WAR is a counting stat. As long as the totality of your contributions is above replacement level, playing more games will increase your WAR total. If Cabrera had played 15% less, at the same rate he did, he'd have 15% less WAR.

Maybe you shouldn't criticize a system you clearly don't fully understand.

War does not do what you think it does. If you understood it, you would realize that a player who plays league average defense for 162 games is considered to be lesser defensive player than a guy who played one game and was slightly above average. Simple as that. I'm talking only about the defensive component, I don't give a rats ass about the offensive component here. War is a totality of several numbers including defense. It doesn't reward playing in the field everyday, it's a clear flaw in the system. Same goes for baserunning.


Considering that everyone on here defending war has stated that war is the summation of running, defense, offense etc, the fact that it clearly fails on defense and running is a problem. I know why they do it that way, but it still doesn't change the fact that it exaggerates the difference between a player like Cabrera vs Trout.

Win Shares is the one system that seems to have the gap correctly 35 for Trout, 30 for Cabrera.

War does not do what you think it does. If you understood it, you would realize that a player who plays league average defense for 162 games is considered to be lesser defensive player than a guy who played one game and was slightly above average. Simple as that. I'm talking only about the defensive component, I don't give a rats ass about the offensive component here. War is a totality of several numbers including defense. It doesn't reward playing in the field everyday, it's a clear flaw in the system. Same goes for baserunning.

That's just completely wrong. The value for playing time is included in the replacement level component. An average player who plays a lot racks up more value than an excellent part-timer.

You saying so doesn't mean it's not there.

It doesn't reward playing in the field everyday, it's a clear flaw in the system. Same goes for baserunning.

One, it's not true - adequacy in defense is already rewarded in positional offensive replacement level. Why would we double-count being average?

Second, replacement level for defense is very high, adequate defensive players are a dime a dozen if you don't care how they hit.

That's just completely wrong. The value for playing time is included in the replacement level component. An average player who plays a lot racks up more value than an excellent part-timer.

Again, in the defense component it does not. A player who is purely average defensively gains absolutely no value whether he plays 1 game or 162 games. His defense component is exactly the same.

A hitter can hit below average and gain value for every game he plays, but only on the offense side of the ball. A player who plays in the field for the team doesn't get that same generosity.

One, it's not true - adequacy in defense is already rewarded in positional offensive replacement level. Why would we double-count being average?

Second, replacement level for defense is very high, adequate defensive players are a dime a dozen if you don't care how they hit.

Again. I know why they do it the way they do it, but it then overstates the value when you have two players, one a below average defender at a non-premium position who played 150+ games versus a guy who was good defensively but only played 110 games.

Second, there is no replacement level for defense, they are using league average for defense and baserunning. They use that standard because the theory is if you hit at replacement level, you have to perform at league average to be worth anything. A replacement level hitter wouldn't get playing time if he was also a replacement (which doesn't exist in war) level defender.

Again, in the defense component it does not. A player who is purely average defensively gains absolutely no value whether he plays 1 game or 162 games. His defense component is exactly the same.

A hitter can hit below average and gain value for every game he plays, but only on the offense side of the ball. A player who plays in the field for the team doesn't get that same generosity.

The replacement component is not part of the offensive component. I think it was when WAR was first introduced, but now it's independent of offense and defense.

The replacement component is not part of the offensive component. I think it was when WAR was first introduced, but now it's independent of offense and defense.

Huh? Replacement is the level that the maker of war determines as the baseline for whether a player is performing above. It's has to part of the component or else there wouldn't be any negative numbers possible. The fact that a player can put up a negative oWar is proof that replacement is part of the offensive component.

If you are talking about rBat that is different, that is relative to average.

But here is what they say about dWar A defensive measure of wins above replacement, but given only the defensive stats of the player and his position adjustment. For this calculation we use a replacement level on defense is the league average.

Huh? Replacement is the level that the maker of war determines as the baseline for whether a player is performing above. It's has to part of the component or else there wouldn't be any negative numbers possible. The fact that a player can put up a negative oWar is proof that replacement is part of the offensive component.

If you are talking about rBat that is different, that is relative to average.

Except oWAR and dWAR aren't quite components of WAR. They're groupings of a few components, and oWAR includes the replacement level one. In WAR (overall), the offense and replacement level are separate.