Why We Overvalue Relievers

Earlier this offseason, Joe DeLessio of New York Magazine did a countdown of the most important Yankee players. His number 1: Mariano Rivera.

Those of you who are sabermatrically-inclined probably just responded with a collective eye-roll. Wins Above Replacement ranked Rivera as only the 5th most valuable pitcher on the Yankees last season, behind CC Sabathia, A.J. Burnett, Andy Pettitte and Phil Hughes. When you include position players, Rivera drops to 15th, just behind Brett Gardner.

How is that possible? Dave Cameron explains:

While the quality of [relievers] work is very high, the quantity is low, which limits their total value. It’s nearly impossible to rack up huge win values while facing less than 300 batters per season. Yes, each of those batters faced are more critical to a win than a regular batter faced, but this is accounted for in WAR.

It's not to say that WAR is a perfect measure - I don't think anyone believes that Brett Gardner is more irreplaceable/valuable than Rivera - but if the numbers are even close, then it's clear that people (media, fans, etc.) tend to overvalue relievers. When you look at these numbers, it becomes clear that some front offices share this skewed view and are willing to overpay them as well.

Why is that? Perhaps a series like the one the Yanks just wrapped up can shed some light.

None of the six pitchers Joe Girardi and Terry Francona called upon to start the last three at Fenway games factored into a decision. Only Sabathia and Lackey were particularly close - both of them watching their lead evaporate under the watch of the man who relieved them. Therefore, each contest was decided by pitchers who entered the game via the bullpen.

Chan Ho Park was on both sides of that equation, taking the loss on Sunday night after allowing a two run homer to Dustin Pedroia and getting the win last night after throwing three scoreless innings while the game was knotted at 1. Jonathan Papelbon had a similar experience, converting a save on Friday and blowing one last night. Each played the role of goat and hero just a few nights apart.

While WAR can objectively weight the contributions of pitchers by leverage, we as fans can't hope to be nearly as unbiased. As a close game progresses, stress and anxiety in the attentive viewer build. Our joy and frustration are multiplied by those factors and relief pitchers are the one major variable in the equation. The lineups are essentially the same but as the stakes within the game increase, the faces on the mound change.

And that's why someone would try to make the case that Mariano Rivera is the most important player on the Yankees. He might not be the most important from a zero sum sabermetric perspective, but he is on a purely observational standpoint, if you have a rooting interest in the team, Mo is the man. CC Sabathia throws far more innings, but they don't seem to have as much on the line. Mark Teixiera plays in almost every game, but most of his contributions occur under ordinary circumstances.

When Rivera enters the game, as a fan, you can exhale. We've seen him do it so many times before, it's hard not to be confident. You trust that he's going to get the job does until he doesn't - and then you assume that he'll do it next time. Conversely, the three innings that Chan Ho park pitched felt significantly more tense and uncertain. The difference between them is in that respect more than commensurate with their respective abilities.

Of course, Mo did what he usually does during the past two nights. He gave up just one baserunner and the go-ahead run never came to the plate. A couple of late nights at the office and two saves in the book.

It might not show up in advanced stats or translate to as many wins above replacement as we would assume, but Rivera and other trustworthy relievers contribute greatly to the enjoyment of rooting for the team. If you spend enough time reading about and understanding the principles of sabermertics, you should realize that his importance is magnified in your mind. But when a save situation rolls around, he really does seems like the most important guy on the team.

Gaudin Waived

According to Ed Price of Fanhouse, Chad Gaudin was placed on waivers last night. Price further tweets that the waivers are outright waivers, making them irrevocable and meaning the Yankees cannot pull Gaudin back if another team claims him.

This leaves Gaudin with a few potential fates: he gets claimed by another team who assumes the entirety of his $2.95M salary; he clears and the Yankees outright him to the minors, remaining responsible for his contract; if Gaudin, has been outrighted before, he has the option to refuse the assignment and forfeit all but 25% of his contract; or he clears and the Yankees release him outright, owing him 25% of his contract.

Gaudin is a capable pitcher, so I'm assuming someone will claim him. FanGraphs' fan projections for 2010 have Gaudin at 0.8 WAR, which would be worth $3.6M on the open market. He's averaged 1.5 WAR over the last three seasons, good for approximately $6.75M on the open market. By those standards, he's a worthwhile pitcher at $2.95M.

Much like yesterday's return of Jamie Hoffmann, I'm a little bit surprised and somewhat disappointed by this move. Gaudin was essentially competing with Sergio Mitre for the longman/sixth or seventh starter position on the roster. While Mitre has been the better pitcher this spring, Gaudin is younger, has the better track record, has been effective in the AL in the past, and has shown himself to be serviceable both out of the pen and at the back of the rotation. Every available projection system predicts Gaudin as the better pitcher this year. Perhaps their salaries figured into this decision; or perhaps this move is a precursor to a rumored trade to secure Hoffmann's rights.

The recent transactions with Hoffmann and Gaudin leave the Yankees with two open spots on their 40 man roster.

What About Paginski?

Here's an interesting thought experiment from the Orioles Blog Camden Crazies (via BBTF):

This idea came up when discussing the possibility of the Yankees signing Matt Holliday on Twitter:

Doesn’t make sense for NY to spend that money on him now. The could put me in LF and still maybe make the playoffs.

Which begs the question, just how many wins could I cost a team if I played a full season? Any guesses, before I try to figure it out?

Seems easy enough to figure out. Same procedure would be used as when evaluating any other player. Take a guess, and then let’s work it out.

First off, the stipulation was that I would get 600 plate appearances in left-field, but I think it’s the case that if I had to play in the field all the time I wouldn’t actually make it to 600 PA. So instead, I’ll go easy on myself and just DH for the season.

The author, Daniel Moroz, is very conservative in projecting his offensive contributions, giving himself a .050/.100/.050 line for the season. Much of that comes from the assumption that if you just kept your bat on your shoulder, you'd work the occasional walk.

Click through to find out just how many runs and wins below replacement a .150 OPS is and to see whether or not the Yankees could hypothetically make the playoffs with him at DH.

It also reminds me of the following classic New York State Lottery commercial:



"Paginski" in the commercial was only batting .027 and he was also costing him team lots of runs with his awful defense at third base, meaning the Designated (Anti-)Hitter Moroz probably isn't the worst player in imaginary baseball history.

The moral of the story is that even the worst major leaguers are several orders of magnitude better than the average person watching from the stands or their couch at home. Although Joe Schmo would probably walk more than Bengie Molina, Yuniesky Bentancourt or Jeff Francoeur.

'Tis The Season For Analysis And Reason

Good morning Fackers. One of the positive things about the slow news cycle that we are going through in the baseball world at the moment is that it gives some very smart writers the chance to do long-form pieces that don't get done when the offseason news is reaching its peak. While some may yearn for the heat of the hot stove, a whole lot of writing and thought gets expended on the the same signings and trade with a relatively small amount getting added to the discussion. Without many rumors flying, some have been using their time well and posting tremendously interesting analytical pieces that have in some cases taken on a life of their own.

Joe Posnanski recently cranked out 5,000 words comparing the Baseball Hall of Fame with the Baseball Think Factory Hall of Merit. Larry from Wezen-ball explored the illogical nature of the voting process for Cooperstown. There was an interesting discussion yesterday about putting a third team in New York which began with a piece on SI.com by Tim Marchman, and was picked up by Craig Calcaterra of Circling the Bases and Kevin Kaduk of Big League Stew (among others) before Marchman responded to the responses. Dave Cameron of FanGraphs discussed the diminishing value of a marginal win for the Yankees which was further explored by River Ave. Blues and Rob Neyer. Cameron's colleague R.J. Anderson used a quote from Peter Gammons to explain the Time Value of Money as it relates to Jason Bay's contract.

Perhaps the best example of this analytical collaboration are the Mike Silva Chronicles that Tom Tango is currently posting over at The Book Blog. Silva is notoriously suspicious of advanced statistics while Tango is responsible for creating a great many of them and inspiring still more. Recently, Silva compared the necessity for the newly-created stat wRC+ (just like OPS+ but based on wOBA) from FanGraphs to "Cool Ranch Doritos, New Coke, or a colorful cover for the iPod". Tango replied in the comment section, "As I keep saying on my blog time and time again, summary opinion without evidence is the very definition of bullsh!t. And Mike’s statement here is pure bullsh!t."

What might have spiraled into a war of words has instead turned into something productive. Tango agreed to answer ten of Silva's questions about advanced metrics, ranging from the accuracy and utility of UZR to the concepts of replacement level and win shares, pulling back the curtain a bit in attempt assuage some of the doubts that many have about newer stats.

Although these two guys are coming from ends of the statistical spectrum, they are meeting at an important common ground. One of the reasons old school baseball thinkers like Silva distrust advanced metrics is because they doubt their accuracy. The people who understand these measures the best understand their limits because they created them. It could be argued that the people who use UZR and WAR without understanding their blindspots are just as bad as those who refuse to acknowledge their utility at all.

Even if you aren't a stathead, I'd recommend stopping by The Book Blog and taking a look at Silva's questions and Tango's responses. They did five questions yesterday and I believe the next 5 should be coming soon. Although the names of some of the newer stats make them seem vague and complex, I think you'll find that hearing someone explain what they are and why they were created while simultaneously acknowledging their limitations to be quite interesting.

Counterpoint: Who Cares?

In the six months I've been writing here, rarely have Jay and I disagreed on any major points. We disagreed about a hypothetical recall of Austin Jackson in the wake of a potential Melky Cabrera injury, and we differ as to how the Yankees should approach Hideki Matsui this off-season. But in instances like these, I think we both can understand, if not agree with, the other viewpoint.

Today though, I think we may have an instance of us being distinctly on opposite sides of the coin. This morning, Jay parsed the AL MVP results and had some valid criticisms of the order of finish as well as some of the more obscure individual votes. Moshe Mandel at The Yankee Universe had something similar yesterday, and I've seen traces of the same sentiment elsewhere in the blogosphere. To which I say: who cares?

Perhaps this is hypocritical of me. Just last week I put up a post here essentially criticizing the "stats based community" for not calling out Tyler Kepner, Zack Greinke, and Brian Bannister for their misuse and misunderstanding of advanced metrics in the wake of Greinke's Cy Young Award victory. Though perhaps I didn't make the point as clearly as I wanted, my issue wasn't so much with the misuse and misunderstanding as it was that everyone was so happy that Greinke both won and acknowledged FIP in the process, that they withheld the usual cantankerousness and I'm-smarter-than-you responses that normally follow such a slip up. Now, when the reaction is a bit truer to character, once again it's me who's complaining about the complaining, just as last week I complained about the lack of complaining.

That said, I still think this is really, really unnecessary. Yes, by any worthy metric Derek Jeter was more valuable than Mark Teixeira. Yes, Ben Zobrist probably finished much lower than he deserved. Yes, there were several players who received individual votes that were higher than they deserved or not deserved at all. Yes, I'm surprised/disappointed that Jason Bartlett didn't receive a single vote. Yes, a first place vote for Miguel Cabrera is so patently stupid that it's probably a waste of energy to even explain why.

But still, do we really need to break down all the grave injustices that happened behind Joe Mauer's cavernous margin of victory? Should we even care what happened beyond Mauer taking the hardware? Isn't the whole point to get the winner right, not whoever comes after him? Don't we often preach about sample size, and isn't the point of casting 28 ballots with 10 slots each for a pool of over 400 players that the "most valuable" player will rise to the top? Sure there will be some oddball votes in there, but the right guy won, and he came within one vote and five points of doing so unanimously. Does any of the rest matter?

Please don't take for me anti-statistic. If you read here with any degree of regularity during the season you know that I often sigthed OPS+, wOBA, UZR, WAR, FIP, etc. That isn't the point. The point is how much is enough? We've seen the deservedly-maligned BBWAA get all four major awards "right", and three of the votes weren't particularly close. In doing so they've eschewed traditional biases that would have favored less deserving candidates in years past. Shouldn't this be enough to keep us content for now? Remember, this round of off-season awards represents the Battle of Saratoga, not the Treaty of Paris.

At their core, the MVP, the Cy Young Award, and even the Hall of Fame for that matter are subjective, loosely-defined awards that have little value beyond whatever worth we assign to them. I have a hard time understanding the utter outrage year after year as the awards season comes and goes. While we'd all like to believe that "objective journalists" are the stewards of these institutions, the fact of the matter is we're not talking about Edward R. Murrow or Walter Kronkite here. These are sportswriters. And while many may still exhibit signs of homerism or may be hopelessly clinging to archaic and ineffective means of measuring performance, I still think they're in a much better state today than they were in the days of Grantland Rice, or Ford C. Frick, or Jimmy Cannon. We're not seeing even the likes of Barry Bonds getting jobbed out of a deserved MVP because he's an asshole to the writers. Ask Ted Williams' frozen detached skull what he thinks about that.

So, with a plethora of better methods to assess value at our disposal, why do we even care who wins these things anymore, let alone who finishes second through tenth? If we want to know who is truly valuable, then why not just pull up the WAR leaderboard, or whichever future metric proves to be the most accurate means of assessing performance? Why do we care which pitcher is given an award named after the all-time wins leader, when we know wins are a misleading indicator of true performance, and we know that Cy Young was an inferior pitcher when compared to contemporaries like Christy Mathewson and Walter Johnson? Do we even need the BBWAA to give the awards to the "right" players to validate what the metrics have already told us?

As Joe at RAB pointed out in the wake of the Cy Young voting, this isn't a culture war anymore. This is the true state of baseball these days, and it's an amalgamation of what were once two distinct schools of thought. As much as some seem to define themselves by it, this is no longer us vs. them, Moneyball vs. old-school, stats vs. scouts, RBI vs. WAR. The game - or more specifically how we look at the game - has changed, is changing, will continue to change, and will do so across the board. It's no longer just Bill James, a few forward thinking executives, a handful of enlightened websites. It's widespread; it's prevalent.

Few if any of GMs are traditional "baseball men"; nearly all have a business background in addition to their baseball experience. Every front office has some sort of statistical analysis taking place. Bill James, Voros McCracken, Tom Tango and others were or are employed by, or are consultants to, Major League clubs. David Cone's routinely referencing Fangraphs on Yankee telecasts. Keith Law, Will Carroll, and other Baseball Prospectus folks were given votes in the post-season awards process. High-profile national sportswriters like Rob Neyer and Joe Posnanski are at the forefront of the "statistical revolution", and if their work wasn't enough to force their colleagues to learn about advanced metrics it was at least enough to create a palpable buzz about the truly worthy candidates.

The times they are a changing folks. And while we might not yet be living in a sabermetric utopia, we've seen great strides made this past week. As we approach Thanksgiving and presumably stop to reflect upon what we are thankful for, shouldn't we at least be satisfied, if not grateful, that the most deserving candidates won both MVPs and both Cy Young Awards rather than griping about the idiot who voted Jason Kubel seventh? I think we should; what about you?

2008 vs. 2009: It's a WAR

Last night, the always brilliant Beyond the Boxscore, in light of the announcement of the Comeback Player of the Year award, took a look at who (among the top 10 vote-getters) had the biggest swings in Wins Above Replacement. I'll refrain from saying "who deserved the award" because I think Comeback Player of the Year is one that should be somewhat subjective and context-heavy because players can "comeback" from a variety of things and that should probably be taken into consideration. That and no one really cares that much about who gets the award anyway.

Chris Carpenter won in the National League and looking at WAR backed that up along with the top four finishes in the voting.

However, the numbers did not match the voting in the American League. Aaron Hill was given the award, but he was actually had the 4th biggest comeback in WAR behind Juan Rivera, Victor Martinez and our own Jorge Posada who had the largest. However, Jorge finished 4th in the voting behind Martinez and Michael Cuddyer probably because he missed time to injury and was relatively productive in his time with the club.

It seems like this year, there have been a ton of Yankees who have improved on last year's campaign, not to mention the three marquee free agents and Nick Swisher who weren't on the team last year. Below, you can find two charts displaying the differences in WAR from 2008 to 2009. BtB measures the dip between the last good year they had and 2009 but I'm only looking at the last two years. (All data via FanGraphs)

First, the position players:

Some interesting stuff here...

• Pretty cosmic shift in run production there, 15.6 wins. The Yanks won 89 games last year and 103 this year. You do the math.
  
  
• I didn't use the more complicated method that BtB did, but those measurements show that Posada wasn't even the most improved player on the Yankees in relation to last year.
  
  
• Jeter made some major strides defensively but his OPS also ended up exactly 100 points higher than in 2008.
  
  
• Obviously, Teixeira represents a vast improvement over the 'stache, taking absolutely no one by surprise.
  
  
• Even with the great year Bobby Abreu is having out in L.A., he's only been worth 2.7 WAR. Still a great pick-up by the Angels for the money ($5M) but the Yanks have been better off with Swish, even though his defense isn't exactly sparkling either.
  
  
• Who would have guessed that Johnny Damon was more valuable in 2008? Hands, please. He's been better with the bat but his fielding has declined by almost 10 runs.
  
  
• If you extrapolate A-Rod's production to a full season he comes awfully close to replicating his 2008 value.

And the pitchers:

• How about the fact that this slice of the 2008 staff was actually worth more in 2008 than in 2009? Not a complete measurement but it's obvious which side of the ball had the most impact in taking the team from third in the AL East to first in all of baseball.
  
  
• I chose Sabathia to replace Mussina since they were the best pitcher in each year. I paired Burnett with Darrel Rasner, Sidney Ponson and Carl Pavano since those combined had about the same number of innings and they would have needed to find similar scrubs to fill in had Burnett not been acquired.
  
  
• Phil Hughes proves that point that if you can be a great reliever, you're more valuable than a bad starter, but I think we already knew that.
  
  
• Before looking, I thought Andy Pettitte would have been more valuable this year than last, but apparently not. I'm honestly not sure what's causing the variance there.
  
  
• Mo saved 39 games last year and blew only 1, while this year he converted 44 and blew 2. His W/L record last year was 6-5 and this year it was 3-3. Wins and saves are bad stats but we are talking about Wins Above Replacement here.
  
  
• Joba derived a good amount of value pitching out of the bullpen last year, which accounts for a lot of the difference.
  
• Our poor Wang got beaten pretty severely this year. He alone makes up for the difference between '08 and '09.

Thanks, But That Would Be A Complete Waste Of A Roster Spot...

Good morning, Fackers. Last night, a suddenly contrite and shockingly self-actualized Joba Chamberlain told reporters that he would "fold towels" if that's what it took to win with the Yankees during the postseason.

Aside from the towel quote that Sam Borden transcribed in his post, here is another exchange that caught my ear from the audio he provided.

Joba: You know, I didn't have fastball command at all. It could have been a lot worse than it was but it was just lack of fastball command from the get-go... Mechanics-wise, I wasn't great. My delivery wasn't great and the delivery affects your command in every aspect.

Kim Jones: You were so good your last time out. How do you explain the inconsistency?

Joba: If I had an explanation, I could probably fix, you know, quicker than things go. That's part of this game and that's why people go in and do bullpen work and look at watch video and figure out what you can do better.

So he did have a bad outing because the players on the other team were "great hitters" or "a couple pitches" that went wrong as he claimed in the past.

Perhaps this is only because the Royals are astoundingly bad at hitting, but more likely it's that someone with the Yankees organization finally explained to Mr. Chamberlain how the media works. They aren't your parents trying to get you to admit your shortcomings as a way to make you take responsibility for them. Reporters need to pair your performance with a reaction so they can write a story about it that conveys what happened to the fans and get the fuck out of there. That process doesn't work when you give up 7 runs and say everything is fine and dandy.

It turns out that Joba wasn't just a delusional headcase, unable to acknowledge his mistakes. He was a kid from Nebraska being forced to come to terms with his failures in front of cameras and reporters with microphones and was getting a little defensive.

Joba mostly dodged questions about about his place on the postseason roster but made clear with his offer to do the work of a clubhouse attendant that he's not going to pout about being used out of the bullpen. I personally think that he is still enough of an asset to warrant occupying a spot on the postseason roster. I'd prefer a hypothetical ALCS Game 4 start to be made by Chad Gaudin and backed by Alfredo Aceves but Joba is still a better option to pitch an inning or two out of the bullpen than Brian Bruney. Maybe this is wishful thinking, but Joba seems to find his stroke when the pressure builds; Bruney just walks a lot of guys.

Since today is an off day, it's conceivable that Girardi uses Chamberlain sometime during the last two games of the Tampa Bay series to get him an inning or two out of the 'pen if indeed they plan on bringing him along for the ALDS. It would give him 3 or 4 days of rest which should be enough to free him up for a limited number of pitches.

I'm a lot less worried about Chamberlain's role in the postseason than I was a week or two ago, though. He's going to be one the the least important players not named Eric Hinske on the postseason roster if he is even ultimately included on it. The Yanks have won 102 games this year in spite of Joba, not because of him.

Delicious

Because everyone loves pie, here is a graphical representation of the Yankees Wins Above Replacement via UmpBump. Dig in!

Quick thoughts:

• Interesting that Phil Hughes checks in slightly higher than Joba Chamberlain (and Mariano Rivera), isn't it?
  
  
• I didn't think CC would be more valuable than Teixeira, did you?
  
  
• Even if you adjust for the time that A-Rod missed to begin the season, he still wouldn't be as productive as Jeter.
  
  
• If you project How is Chien Ming Wang OVER replacement?

More Worthless: Wang Or Papi?

The parallels between these two didn't become apparent to me until I read this post on Mass Hysteria, which seemingly took just as much pleasure in Chien Ming Wang's demise as we have in Papi's. They have more in common than just their names being euphemisms for one's penis.

In the first inning of last night's game, we saw a clash of fallen heroes: Papi, the populist champion of Red Sox Nation and Wang, the most revered athlete in the island nation of Taiwan. The Dominican native has long dominated the battle between the two. In 44 plate appearances against Wang, Ortiz was hitting .432/.523/.703 with 7 walks. He worked another base on balls and won the symbolic battle of two stars wrapped in momentarily coinciding downward spirals.

Both have been a terrible detriment to their teams this year, but in different ways. If these two were actual methods of torture, Wang would be more like waterboarding (brutal for short stretches) whereas Ortiz is the equivalent of Chinese water torture (a slow drip that eventually drives you insane). So which is worse?

According to FanGraphs, Papi is far less valuable than Wang. He clocks in with a value of -$4.7M and a WAR (Wins Above Replacement) of -1.0. Surprisingly, Wang is "only" costing the Yankees $600,000 and one tenth of a win. It would seem that the formula FG is using is letting Wang off the hook since he only pitched 21 1/3 innings so far this year. This should instead count against him, since his lack of length taxes the Yankees bullpen and has stranded them in every one of his starts. You can do a lot of damage in 21 1/3 innings. And Wang certainly has.

A starting pitcher has the ability to sabotage a game in a way no other player does. A starter giving up 6 runs in two innings is less valuable than anything a position player could do. Yes, even going 0-7 with 3Ks and leaving a small army on base (while DH'ing).

Here are Wang's 5 starts this year:

There's no way those only cost the Yankees 1/10th of a win.

In order for the Yankees to win any of those games, they would have to score more than the number of runs he allowed, obviously. Based on the Yankees offense this year, here are the probabilities that they score X or more runs in any given game.

Of course, this assumes that the bullpen is going to be perfect throughout the rest of the game, which is extremely unlikely when Wang averages only 2 2/3 innings per start and the Yankees bullpen has an ERA of 4.91. That number might even be a little favorable considering when you need to get 6 1/3 innings from the pen, you are using mop-up men or overtaxing your competent relievers.

So if you project how many runs the bullpen is likely to allow on top of Wang's performances and line those up with how probable it was for the Yankees to score more runs than that, you'd come away with a much better estimate of just how bad Wang has screwed the Yanks in his starts so far this year.

SPW% is the likelihood that the offense would score more runs than the pitching staff allowed in any given game. The most runs the Yanks have scored this year is 12 and that was only once, hence the 0% chance of winning the game against the Indians (which of course they lost 22-4) and the 1% chance against the Rays (2% chance they score 12 runs and 50% they win in extra innings).

Add those percentages up and you get a 54% win probability out of a total 500%. That's not costing the Yankees 1/10th of a win, friends, it looks a lot more like negative 4.5.

This obviously isn't overly scientific or definitive, so what do you think? Who would fetch the ham sandwich on the trade market and who would come with chips? Which one of these guys would you rather not have?