According to B-ref, these are the AL leaders in pitcher WAR:

1. Hernandez (SEA) 5.1
2. Wilson (TEX) 4.8
3. Buchholz (BOS) 4.7
Weaver (LAA) 4.7
5. Liriano (MIN) 4.5
6. Price (TBR) 4.4
7. Pavano (MIN) 4.1
8. Lester (BOS) 4.0
9. Sabathia (NYY) 3.9
10. Danks (CHW) 3.8

According to Fangraphs, these are the WAR leaders among pitchers:

1 Cliff Lee 6.0
1 Francisco Liriano 6.0
3 Felix Hernandez 5.6
4 Jon Lester 4.7
5 Jered Weaver 4.5
6 Zack Greinke 4.4
7 Justin Verlander 4.0
8 Gavin Floyd 3.9
9 Colby Lewis 3.7
9 CC Sabathia 3.7
9 C.J. Wilson 3.7
9 John Danks 3.7

Maybe someone who studies these things could explain why Cliff Lee is not on the B-ref list? B-ref has Lee's WAR at 3.4, which seems odd given his season and given what Fangraphs says about him. Colby Wilson and Clay Buchholz also have wide differences in WAR according to the two sites.

EDIT: Maybe Fangraphs is not updated. Lee's ERA+ with Texas is 92. Maybe his WAR has faded with less than stellar pitching of late?

Halladay is on pace to become just the third pitcher in the live-ball era to lead his league in most strikeouts and fewest walks per nine innings in the same year (joining Urban Shocker in 1922 and Robin Roberts in 1954).

Whoah, that's pretty cool. Also it was nice of the Phillies to trade Cliff Lee so Halladay didn't have to compete with him for the BB/9 crown.

Subjectively, I was beginning to like B-ref WAR (esp 2-7 here) over Fangraphs but I agree what is up with Lee?

After today's dominant 8 innings / 1 hit start, Sabathia is probably up to 7th on Rich's list. And does anyone really think that the CYA is likely to be determined by a 1 game WAR differential, if another pitcher wins 21-23 games and does it in a tight division race?

Not that Hernandez wouldn't be a good choice, since AFAIC he's the best pitcher in the league. But then that may have something to do with those 91, 70 and 82 Game Scores against the best offense in the game, in addition to all those other numbers.

"After today's dominant 8 innings / 1 hit start ... against the sorry A's offense ..."

Maybe Fangraphs discounts the accomplishment when CC dominates the Oakland AAs.

Maybe someone who studies these things could explain why Cliff Lee is not on the B-ref list?

FIP vs. ERA, I believe. I'm not 100% sure though.

How is B-Ref's WAR calculated? I know Fangraphs' uses FIP, which leads to really stupid results, like Jason Hammel being more valuable than Tim Hudson. I suspect that B-Ref's WAR is a bit more reality-based.

After today's dominant 8 innings / 1 hit start, Sabathia is probably up to 7th on Rich's list. And does anyone really think that the CYA is likely to be determined by a 1 game WAR differential, if another pitcher wins 21-23 games and does it in a tight division race?

I think the quality of the Yankees actually works against Sabathia here. It's hard to rationalize that a guy is integral to a team's success, even in a close pennant race, when worst-comes-to-worse the team loses the division but still wins the wild-card by 4 games. I do think that Felix has to at least break .500 in order to win the CYA, though. No matter how enlightened they are, I can't see the voters ever picking a guy who goes 12-12.

Is it possible that B-ref is not using Lee's full-season stats, and that there is a quirk in the software that is splitting Lee's stats between Seattle and Texas?

But remember, this is a performance award, period.

I always get a kick out of the idea that the CYA is supposed to be purely about performance, as opposed to the MVP (which is, of course, about "value"). It says "Most Valuable Pitcher" right there on the freaking plaque.

Maybe Fangraphs discounts the accomplishment when CC dominates the Oakland AAs.

But not when Felix does?

But are voters really ready to ignore that win column completely? We’ll find out. Won’t we?

yep, we'll find out that after today's 1 hitter, CC will win about 80% of the first place votes and win the Cy with plenty of room to spare. I don't think it's right, but that's the way it is.

After today's dominant 8 innings / 1 hit start, Sabathia is probably up to 7th on Rich's list. And does anyone really think that the CYA is likely to be determined by a 1 game WAR differential, if another pitcher wins 21-23 games and does it in a tight division race?

I think the quality of the Yankees actually works against Sabathia here. It's hard to rationalize that a guy is integral to a team's success, even in a close pennant race, when worst-comes-to-worse the team loses the division but still wins the wild-card by 4 games.

Somehow I doubt if too many voters are going to follow that sort of reasoning. If they give it to King Felix, I think it'll be because they recognize how great he's been, not how fortunate Sabathia was that the Red Sox have been crippled with injuries. The truth is that this is one of those years where when you consider all the factors, there isn't any one clearcut CYA winner---and besides, there are still over four weeks to go in the season.

Beyond that, I seriously question that any replacement pitcher for Sabathia would have meant only four fewer wins for the Yanks, not when you factor in all the extra rest he's given to the bullpen. That won't show up in his own personal stats, but it could very well have meant an extra few wins in games that followed one of his starts.

To truly appreciate that last point, and to understand just how valuable Sabathia has been to the Yankees this year, try this:

Number of starts of 7+ innings / Number of starts of fewer than 6 innings

Sabathia: 20 / 3

Hughes: 8 / 11

Burnett: 9 / 10

Vazquez: 8 / 15

Pettitte: 9 / 4

And imagine the relative stress that those five starters have put on the New York bullpen.

BTW King Felix is even more impressive by that metric (21 / 2), but then he's had no pressure in terms of the division race since the middle of May.

I think a guy could win the Cy Young award with a 13-8 record or something like that. But it will be tough for him to win it if he has a .500 record like Felix does now.

In other words, the number of raw wins is important to the voters, but they will also look at winning percentage.

RIFKIN, Not Ripken: "B-ref has Lee's WAR at 3.4"

ANGEL, Not Ahn-Hell: "Is it possible that B-ref is not using Lee's full-season stats, and that there is a quirk in the software that is splitting Lee's stats between Seattle and Texas?"

I think the B-ref stats are updated daily.

FWIW, B-ref shows Lee -- Cliff, that is, not the guy who hated the Discovery Channel -- having 2.9 WAR for Seattle and and 0.5 WAR for Texas. For simplicity -- not Felicity -- I added them together so Cliff could be compared to those who made the list.

Lee's BB Ref WAR looks fine to me. The difference between BB Ref and Fangraphs is basically the question of whether you think that the actual runs a pitcher allows tell you more about how valuable he's been than a theoretical value assuming a league average BABIP tells you.

He had a 2.69 RA in 103.7 innings with the Mariners. League RA is 4.46, adjusted for Safeco it's 4.39. So he was something like 20 runs above average. If you set replacement level for a starting pitcher at 25% worse than league average that's around 32 runs above replacement level.

With Texas he's got a 5.13 RA in 80.7 innings, compared to a 4.55 league average RA (adjusted for TBiA). So he's been five runs worse than average, and about four runs better than replacement level.

Felix Hernandez probably deserves the Cy Young right now, but he's probably going to have to pick a up a win in all of his remaining starts to have a shot at it.

There's no software quirks going on. B-R uses run prevention, adjusted for team defense. Fangraphs WAR uses only FIP. Cliff Lee has still put up ridiculous peripheral stats in Texas, and thus, still gets rated highly by Fangraphs WAR which totally ignores hits allowed, runs allowed, etc.

I know Fangraphs' uses FIP, which leads to really stupid results, like Jason Hammel being more valuable than Tim Hudson.

I think a lot of that has to do with a park effect at Coors Field. In addition to its other peculiarities, Coors has the largest outfield area in the majors, which probably plays a bit of havoc with some of the defensive metrics. Whatever the reason, Fangraphs hates all the Rockies' outfielders, which leads them to think the Rockies have a bad defense, which leads them to overrate the Rockies' pitching.

I think a lot of that has to do with a park effect at Coors Field. In addition to its other peculiarities, Coors has the largest outfield area in the majors, which probably plays a bit of havoc with some of the defensive metrics. Whatever the reason, Fangraphs hates all the Rockies' outfielders, which leads them to think the Rockies have a bad defense, which leads them to overrate the Rockies' pitching.

Huh? Fangraphs pitching WAR does not have any input from defensive stats. It's purely FIP based.

And does anyone really think that the CYA is likely to be determined by a 1 game WAR differential, if another pitcher wins 21-23 games and does it in a tight division race?

Sure, if pitchers actually won games. since CC doesn't even hit, he hasn't won a game this year.

NL MVP:
Our ballot right now
1. Joey Votto
2. Albert Pujols
3. Carlos Gonzalez
4. Adrian Gonzalez
5. Roy Halladay

You don't have to burn out many brain cells to make a case for Joey Votto. He rolled into September in position to become just the ninth winner in the past half-century of the modern Triple Crown (batting average, on-base percentage, slugging percentage). But he also leads the league in RBIs. He's tied for third in homers. ... But unfortunately for him ... Albert Pujols and Carlos Gonzalez also sit right there in the top five in all three traditional Triple Crown Categories.

Triple Crown stats aside, B-ref's version of WAR for the NL does not believe that Votto, Pujols or anyone named Gonzalez currently deserves to be the NL MVP:

1. Halladay (PHI) 6.3
2. Hudson (ATL) 6.2
3. Wainwright (STL) 6.0
4. Gonzalez (SDP) 5.9
5. Johnson (FLA) 5.8
6. Pujols (STL) 5.4
7. Jimenez (COL) 5.3
8. Votto (CIN) 5.1
8. Zimmerman (WSN) 5.1
10. Huff (SFG) 5.0

A lot can change from now to the end of the season, but it would be a miracle if a pitcher wins the NL MVP, despite having 4 very good candidates. ... In fact, I would bet Albert Pujols's house that Adam Wainright does not win the St. Louis MVP. Halladay, on the other hand, might finish in the top 5. But part of that is because A) he is Roy Freaking Halladay and B) everyone loves to say, "What's up, Doc?"

This, FWIW, is the Fangraphs top 10 in WAR:

1 Roy Halladay Phillies 6.5
2 Joey Votto Reds 6.3
3 Ryan Zimmerman Nationals 6.2
4 Josh Johnson Marlins 5.9
5 Albert Pujols Cardinals 5.8
6 Andres Torres Giants 5.7
7 Ubaldo Jimenez Rockies 5.3
8 Adam Wainwright Cardinals 5.2
8 Matt Holliday Cardinals 5.2
10 Brian McCann Braves 4.9

Notice that Tim Hudson is No. 2 on b-ref and nowhere to be found on Fangraphs, and these lists are supposedly measuring the same thing--Wins Above Replacement. Josh Johnson and Halladay, on the other hand, score about the same on both WAR metrics.

You know what? Thinking about it, I'm shocked that nobody won a Triple Crown during Coors Field's old launching pad days. I guess it was the titanic home run totals of McGwire and Sosa that really prevented it, but you would think that Walker and Helton both would have been solid candidates to do it.

Notice that Tim Hudson is No. 2 on b-ref and nowhere to be found on Fangraphs, and these lists are supposedly measuring the same thing--Wins Above Replacement. Josh Johnson and Halladay, on the other hand, score about the same on both WAR metrics.

It's all due to the difference between ERA and FIP:
Tim Hudson: 2.24 ERA, 3.74 FIP
Josh Johnson: 2.28 ERA, 2.50 FIP
Roy Halladay: 2.27 ERA, 2.80 FIP

If WAR is calculated from FIP (FanGraphs), then Hudson ranks low. If WAR is calculated from RA (B-R), then Hudson ranks very high. For pitchers like Johnson and Halladay, who rank near the top in both FIP and RA, either WAR rates them very highly.

Thinking about it, I'm shocked that nobody won a Triple Crown during Coors Field's old launching pad days. I guess it was the titanic home run totals of McGwire and Sosa that really prevented it, but you would think that Walker and Helton both would have been solid candidates to do it.

Todd Helton led the 2000 NL in average and RBI; he was tied for seventh in HR with 42 (the league leader had 50).

Larry Walker's best shot was 1997 when he was 10 RBI behind Galarraga, .006 behind Gwynn, and won the HR title. The three years Walker won batting titles, he never got any nearer than 28 HR and 32 RBI away.

Galarraga led in HR/RBI in 1996, but hit .304,. He led the NL in average in 1993, but with 22 HR and 98 RBI.

Dante Bichette had the HR and RBI in 1995, and was third in batting average (.028 behind Gwynn).

Ellis Burks was 7 HR, 22 RBI and .009 behind in 1996.

In 1998, Vinny Castilla was 10th in average, 4th in HR and 3rd in RBI.

Notice that Tim Hudson is No. 2 on b-ref and nowhere to be found on Fangraphs, and these lists are supposedly measuring the same thing--Wins Above Replacement.

They are trying to measure the same thing, but they make very different assumptions about how games are won. As others have pointed out, Fangraphs' version measures a pitcher by his fielding-independent events, while BR's credits him by actual runs allowed.

And does anyone really think that the CYA is likely to be determined by a 1 game WAR differential, if another pitcher wins 21-23 games and does it in a tight division race?

Sure, if pitchers actually won games. since CC doesn't even hit, he hasn't won a game this year.

Primey for the most original snark of the year. You could win a fortune in bar bets with that one.

You know what? Thinking about it, I'm shocked that nobody won a Triple Crown during Coors Field's old launching pad days. I guess it was the titanic home run totals of McGwire and Sosa that really prevented it, but you would think that Walker and Helton both would have been solid candidates to do it.

Wasn't there a hypothesis at one point that Rockies' hitters, while benefitting greatly while hitting in Coors, were dragged down somewhat below their expected true ability on the road? Possibly because every time they went on the road they had to adjust to pitches that broke more in sea-level air?

ISTR someone (maybe at BP?) doing a study on that, and maybe concluding that it didn't hold water, but I don't know how rigorous or accurate the methodology was.

Whenever the question of B-R vs. FanGraphs WAR comes up, you should just point people to this link.

http://www.baseball-reference.com/blog/archives/7447

It links to two articles by Patrick Flood that clearly lays out the differences. One for hitters and one for pitchers.

The FIP vs. runs allowed calculation for pitchers will result in some extreme differences in WAR. Last year it was Ricky Nolasco. Looking at his runs allowed (also adjusting for defensive support, ballpark, opponents faced) he's an average pitcher. Looking only at FIP he's a great pitcher, one of the best in the league.

Roy Halladay, with his 190/26 K-W and 2.27 ERA, is going to rank high on either system. But for Cliff Lee and Tim Hudson, the system you choose will make them either a Cy Young candidate or merely a good pitcher. Lee was up there for both metrics, but has a 4.69 ERA since he's joined the Rangers, especially his last 5 starts: 29 IP, 44 H, 30 runs, but still a 35-3 K/W ratio, so on the Fangraphs calculation he's still adding WAR while getting the crap beat out of him.

Pit fWAR rWAR
Tim 3.2 5.8
Lee 6.0 3.4

and these lists are supposedly measuring the same thing--Wins Above Replacement.

Take some time to actually read people's explanations, and you will see that despite the same name, they most definitely are not measuring the same thing. I would appreciate it if you stop acting like something must be wrong if they don't come out to the same results.

Or this link.

Here's what I said yesterday on which WAR is better, and whether someone should pick one site or mix/match and use fWAR for hitters and rWAR for pitchers:

That seems like a matter of personal choice. WAR is a framework. It’s very easy for anyone to substitute one component for another. Since both Fangraphs and BB-ref have John Dewan’s fielding numbers, it would be very easy to use WAR and replace the fielding part with Dewan’s if you wanted to. Or use Chris Dial’s, although not quite as convenient.

I wouldn’t have an issue with using rWar for pitchers and fWar for batters, or something like that. The only consistency you need to follow is be consistent with yourself.

For example, if you choose rWar in general as your metric for batters, but happen to be a Ranger fan. So all of a sudden you decide you like how fWar rates Josh Hamilton, and switch to make an MVP argument. That would be cherry picking. Try to avoid that.

I'll give Buster Olney some credit here -- the guys on Sportscenter last night were saying "WHO WILL WIN THE CY YOUNG? SABATHIA OR PRICE?" and went to him, and he spent a couple minutes explaining that Felix Hernandez is the guy to beat from his point of view, though Sabathia could still win it.

Take some time to actually read people's explanations, and you will see that despite the same name, they most definitely are not measuring the same thing. I would appreciate it if you stop acting like something must be wrong if they don't come out to the same results.

Yes. For pitching, at least, you're attempting to measure what actually happens, and Fangraphs is off in la la land.

Good work.

The no love for Bautista in the AL MVP race seems a little unjustified- the Jays have played decently this year, aren't that far behind Texas in W-L, and you'd think the HR/RBI gap between them (43-103 to 31-97) would help to neutralize the huge gap in BA/OB%. Add to it that Hamilton will likely lose more ground in the counting stats as he sits in September, plus his defense between 3B/RF, and it seems like Bautista should at least get into the top 3.

If Bautista should be getting more MVP love, then how about Beltre and Konerko? They play for teams with better W-L than Toronto and one of them is actually in a pennant race.

I guess the WAR thing for a pitcher depends on how much credit you give for the defense behind him-the FIP way apparently is willing to cede a large amount of value to the fielders (if I'm reading it right), while the runs allowed method put all the onus on run prevention on the pitcher (for good or bad).

I would appreciate it if you stop acting like something must be wrong if they don't come out to the same results.

Yes, Miss Offmore.

I thought her name was Miss Othmar?

How is B-Ref's WAR calculated? I know Fangraphs' uses FIP, which leads to really stupid results, like Jason Hammel being more valuable than Tim Hudson. I suspect that B-Ref's WAR is a bit more reality-based.

Reality based is a nice way of putting it. We can debate FIP all we want in terms of how it evaluates true talent. However it is completely inappropriate in this discussion. There is no room for debate on this. Fangraphs WAR for pitchers substitutes a theoretical construct of what happened for what actually happened. It is a pure construction that is only loosely connected with happened on the field. It's as if you played a diamondmind season and when choosing your Cy Young awardd instead of referring to the actual results of the season you tried to look at coding of the player in order to say who had the best season.

Anyone who suggest using Fangraphs WAR for pitchers for an award doesn't understand what they are talking about.

Whereas Baseball-Reference's WAR is a construction only loosely connected with the pitcher's contribution.

@35 - It's runs allowed adjusted for defense at BBREF so it isn't entirely ceded to the pitcher.

I'm with Gaelan on this one. Tim Hudson has had a great year so far because he hasn't given up very many hits. It isn't because the Braves have played otherworldly defense behind him (although they've done okay) - rather, balls just aren't getting hit hard when he's on the mound. Sure, he probably can't keep it up next year, but that has no bearing on his success this season.

Reality based is a nice way of putting it. We can debate FIP all we want in terms of how it evaluates true talent. However it is completely inappropriate in this discussion. There is no room for debate on this. Fangraphs WAR for pitchers substitutes a theoretical construct of what happened for what actually happened. It is a pure construction that is only loosely connected with happened on the field. It's as if you played a diamondmind season and when choosing your Cy Young awardd instead of referring to the actual results of the season you tried to look at coding of the player in order to say who had the best season.

Anyone who suggest using Fangraphs WAR for pitchers for an award doesn't understand what they are talking about.

Totally agree. Cliff Lee's recent stretch is a forceful reminder that DIPS and FIPS doesn't always tell the whole story - you can post a 35-3 strikeout-to-walk ratio and still get your head beat in on the mound.

Whereas Baseball-Reference's WAR is a construction only loosely connected with the pitcher's contribution.

Meh, like Tango says, you can probably just split the difference. I could see using BABIP/FIP as a tiebreaker when two guys are pretty close in value. Like Wainwright and Halladay in the NL - both are neck and neck in runs allowed, but Halladay's BABIP is much closer to his career value, whereas Wainwright's looks pretty fluky this year. So I'd lean toward Halladay for the NL Cy Young Award (well, that and I'm a total Roy fanboy, but I can sort of justify my logic this way).

"Whereas Baseball-Reference's WAR is a construction only loosely connected with the pitcher's contribution."

What model would do you think best measures the pitcher's contribution?

Not sure what you meant by that. If you're a FIP fan then Fangraphs has your WAR. If you posted that for other reasons, I'd be interested in hearing what else should be done to measure pitchers.

"Reality" is likely some factor between weighting ERA and FIP for war. If pitchers get no credit for "easy" BIP than it's 100% FIP 0 ERA. If they get all the credit (except "unearned" runs) then it's 100% ERA 0% FIP.

I think Chris Dial had a post on this earlier in the year. Maybe establishing a baseline for replacement fielding? Maybe 70/30 FIP/ERA?

To embrace FIP and ignore ERA would be inaccurate even if balls in play were 100% the responsibilty of the fielders.

You also have:
1. Does the pitcher pitch better with runners on base?
2. Can the pitcher field his position? Does he allow runners to run wild or can he pick off a decent number of them?

"Whereas Baseball-Reference's WAR is a construction only loosely connected with the pitcher's contribution."

What model would do you think best measures the pitcher's contribution?

Not sure what you meant by that. If you're a FIP fan then Fangraphs has your WAR. If you posted that for other reasons, I'd be interested in hearing what else should be done to measure pitchers.

In #38, Gaelan was (perhaps hyperbolically?) completely denying that Fangraphs's WAR has any validity in a Cy Young discussion, because "It is a pure construction that is only loosely connected with happened on the field."
I was merely pointing out that B-Ref's WAR is also flawed, but in a different way.

I don't really know what the best model is. I'm just a dilettante with these advanced statistics anyway, but it seems dismissive to ignore a statistic that credits a pitcher just on the things under his control. In a very real sense, what "actually happened" (Gaelan's words) insofar as the pitcher is concerned is that he allowed a ball in play. Whether the ball's path happens to coincide with a fielder has little if anything to do with that pitcher.

@46 AROM - #1 is a very real phenomenon - How does the pitcher perform in the stretch v the windup? #2 is why Mark Buehrle always outperforms his Fangraphs xFIP prediction. There is a lot more to run prevention than throwing strikeouts and not allowing home runs or walks.

(I thought her name was Miss Othmar?

This is correct.)

@48, that's very fair, and "full" WAR would no doubt take that into account (to whatever extent it's defensive component is accurate, anyway). But that's not pitching, and to my mind, therefore not relevant to the Cy Young discussion.

I thought her name was Miss Othmar?

Bingo. Somebody give that man a box of broken eggshells!