Tim Lincecum and the Slow Death of ERA

Last night, Jeff sent me a text that said simply “Lincecum re-signed, 2/35”. My immediate reaction was that this was a hilarious overpay. I had just published a piece earlier in the day explaining why I didn’t see why the crowd thought Lincecum would get 3/40 when Dan Haren was projected for 2/19 and wouldn’t come with the qualifying offer tag. 3/40 for Lincecum, coming off two mediocre years, just seemed like an overpay. He was the kind of guy you should buy low on, and that’s not buying low.

And 2/35, with a full no-trade clause, is even more player friendly than 3/40 would have been. And this is what the Giants paid to keep him from even testing the free agent market; the presumption is that they think his price would have been even higher had they let other teams start bidding. When you factor in the value of the draft pick that would have been tied to signing Lincecum, and the value of the no-trade clause, this contract essentially bets that Lincecum’s market value is somewhere around $20 million per year.

That seems crazy. Last year, he had an ERA- of 124, ranking 74th out of the 81 pitchers who qualified for the ERA title. He ranks right between Jerome Williams and Kyle Kendrick on that leaderboard. And that was an improvement over his 2012 season, in which he ran an ERA- of 139, the very worst mark put up of the 88 pitchers who qualified that year. Over the last two seasons, the only qualified pitcher with a worse ERA- than Lincecum is Edinson Volquez, who the Padres released during the season. By runs allowed, Tim Lincecum has been basically replacement level for the last 400 innings. And he just got valued at around $20 million per year. Crazy, right? Well, maybe not.

We have been writing here, for years, about the flaws of evaluating a pitcher by ERA. There are a ton of variables that are large factors in ERA which are influenced by things other than the pitcher, and just crediting or blaming the pitcher for everything that happens when he’s on the mound can lead to some mistaken conclusions. Run prevention is not synonymous with pitching; run prevention is pitching and defense, along with the sequencing of when the various events occur. The pitcher is a major factor in run prevention, but we should be more interested in isolating his role in the result than in holding the entire result against him.

And that’s just looking backwards, trying to assign credit and blame for things that happened in the past. When we get into trying to project the future, single season ERA becomes even less useful. As Bill Petti noted back in April, the year to year correlation of ERA from 2002 to 2012 was lower than almost every other metric for starting pitchers. It’s not that ERA contains no useful information, but that a lot of the things that drive ERA in one season simply don’t carry over to the next season.

Knowing this, we should not decide that a significant contract for a pitcher with a high ERA is demonstrably nuts. In fact, the crazy thing would be continuing to cling to ERA as a barometer for what a pitcher should be paid for his performance going forward when we know it’s not very good at predicting a pitcher’s future performance. It isn’t clearly crazy to pay a pitcher $35 million when he’s posted an ERA of 4.52 over the last two years; it is clearly crazy to look no further than ERA when deciding what kind of contract a pitcher should get.

And Major League teams have been making that adjustment for a while. This actually looks like part of a trend that began last winter. For instance, here are a few of the more notable free agent starters (without health concerns, which changes all the calculations) who signed last off-season, and their respective performances over the prior two seasons:

Hiroki Kuroda, 1/$15M, 82 ERA-/90 ERA-
Kyle Lohse, 3/$33M, 83 ERA-/103 xFIP-
Zack Greinke, 7/$147M, 93 ERA-/74 xFIP-
Anibal Sanchez, 5/$80M, 95 ERA-/87 xFIP-
Edwin Jackson, 4/$52M, 98 ERA-/96 xFIP-
Ryan Dempster, 2/$27M, 102 ERA-/95 xFIP-

Kyle Lohse was, by ERA, maybe the best starter on the market last winter, once you factor in Kuroda’s age and geographic preferences limiting where he’d sign. Lohse didn’t sign until almost Opening Day. We all chalked it up to the power of the qualifying offer, but his ERA was significantly lower than his peripherals would suggest, and he spent the entire off-season without a serious suitor. Meanwhile, the Dodgers, Tigers, Cubs, and Red Sox spent lavishly on pitchers who had peripherals that were better than their ERAs. Last winter, ERA did little to help one understand how free agent pitchers got paid.

The first deal of the 2013 off-season continues that trend. By ERA, this contract for Tim Lincecum looks crazy. But, more and more, it looks like we are entering an age where teams will bid significant dollars for pitchers who are expected to be better than their recent ERAs suggest. And, of course, Lincecum fits perfectly into that group.

You don’t need a further rehash of Lincecum’s last few years, as everyone by now knows his xFIPs have been pretty good and his ERAs pretty awful. The numbers that are better at projecting future performance think Tim Lincecum is still pretty good. Over the last two years, Lincecum’s 96 xFIP- ties him with Mat Latos and Derek Holland. It puts him almost dead even with Jon Lester, and ahead of Jake Peavy. 2/35 for any of those four pitchers would be hailed as a huge bargain, because their ERAs match up with their xFIPs, and so the public perception is that they are good pitchers. The public is still evaluating pitchers almost entirely by ERA; Major League teams, increasingly, are not.

We have become used to identifying pitchers with good peripherals and bad ERAs as nifty buy low candidates. This, to me, looks like part of a market correction, and the price for pitchers with bad ERAs but good peripherals seems to be going up pretty fast. 2/35 for Lincecum might seem like a dramatic overpay on the surface, but in retrospect, should we really have expected him to sign for much less when we saw what the market said about Greinke, Sanchez, and Jackson last year?

I think the shocking part of this deal is that this is the first big contract for a guy whose results were truly awful. With Greinke, Sanchez, and Jackson, you could make a case that ERA undervalued each of them, but ERA still said they were decent. Teams showed that they were willing to pay high prices for okay ERA/good xFIP guys, but this is the first time a team has shown that they were willing to pay a premium price for an awful ERA/good xFIP guy. This is maybe the strongest rejection of ERA that we’ve seen in free agency.

But this is also something of a logical conclusion to believing in peripheral data more than ERA. If the Dodgers were willing to bet on Greinke’s BB/K/GB rates to to the tune of $150 million, is a $35 million bet on those same numbers really that absurd? It’s the same concept, just on a smaller scale. The Jackson deal is perhaps a better comparison, because Lincecum’s peripherals are closer to the slightly above average numbers that he put up rather than Greinke’s dominating ace type numbers, but again, Jackson was seen as a wildly inconsistent pitcher whose results never matched up to his stuff, and he got $13 million a year for four years. Relative to that deal, the Giants premium in AAV can simply be seen as a trade-off for limiting their risk to only two years.

This deal seems shocking, because we haven’t seen others exactly like it before, but we should have seen the seeds of this kind of deal being planted a year ago. Last winter, the market for free agent pitchers started leaning away from ERA, and the Giants may very well have seen the writing on the wall, knowing that the days of a pitcher with Lincecum’s pedigree and projections being a bargain were over.

Now, that doesn’t mean that the Giants should have signed this deal, or that there weren’t better ways to spend $35 million this winter than on hoping for a Tim Lincecum rebound. I think, realistically, $17.5 million a year, and a no-trade clause, and the lost value from getting a draft pick had Lincecum signed elsewhere, probably does make this an overpay. The Giants may very well have been better off letting Lincecum leave and targeting a pitcher like Dan Haren to replace him, saving money and landing the draft pick in the process.

But I think we now have to revise our expected price for Haren, and pitchers like him, up by a decent amount. The days of simply sorting a leaderboard by the difference in xFIP and ERA and finding bargains might be over. The entrenched hold that ERA has had on pitcher valuations appears to be dwindling. It’s time we stop expecting pitchers like this to sign for peanuts simply because of their ERA. That’s not how major league teams are evaluating pitching anymore.

We hoped you liked reading Tim Lincecum and the Slow Death of ERA by Dave Cameron!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Dave is the Managing Editor of FanGraphs.

newest oldest most voted
Dave
Guest
Dave

Lincecum’s average fastball lost over 2mph over the past few seasons. Are you honestly telling me that GM’s don’t use that info when signing pitchers?

Justin
Guest
Justin

doesnt seem the giants were too concerned, right or wrong.

AK7007
Member
AK7007

The giants saw a no-hitter and said “hold on to that guy.” Jonathan Sanchez was kept way too long based on K rates and a glimmer of what could be. Bet the same things happen with Petit. Point being, I think the Giants really look at a single super awesome performance, and try to extrapolate that into glory, as insane as it is.

Mr. Jones
Member

There is absolutely no way that this could possibly be the case.

Pirates Hurdles
Guest
Pirates Hurdles

LOL, the Asst Gm actually mentioned the no-hitter in an interview with Joel Sherman yesterday.

Brett
Guest
Brett

“There is absolutely no way that this could possibly be the case.”

Mr. Jones, clearly you are not a Giants fan who has nearly torn his hair out over the multitude of ridiculous contracts Sabean has handed out to veterans.

(Sabean deserves a lot more credit than he gets, but this is his most glaring weakness.)

Randyhalo
Guest
Randyhalo

Your gonna compare a guy with 2 Cy Youngs to Sanchez?

cs3
Member
cs3

Well Id say they *were* concerned considering the last time they offered him a contract it was for 5 years and 100 mil.
If they werent concerned then why is this deal 3 years shorter and $65 million poorer?

Sure, Timmy is 2 years older, but c’mon.

dainla
Guest
dainla

There are sixteen starting pitchers with at least 380 innings pitched over the last two years who also struck out more than 380 batters. Tim Lincecum is one of those pitchers.

Of those, 10 are under 30. One is Lincecum.

There are eight qualified starting pitchers who make batters swing and miss on at least 11% of all their pitches. Tim Lincecum is one of those pitchers.

Giants relievers allowed over 60% of the runners Lincecum left on base to score. The league average is 28%.

Bip
Member
Member
Bip

That last stat is misleading. It’s possible that in Lincecum’s “pull” innings, he typically left runners farther along on the bases and managed fewer outs than the average starter. If a starter leaves in the 7th with a guy on first and two outs, then he is leaving his reliever in a better position than a guy who leaves with the bases loaded an none out.

What we would have to do is look at all of Lincecum’s bequeathed runners and add up the probability of each scoring, look at the number who actually scored, and compare that ratio to the league average ratio. One can also add the run expectancies of each base-out state that Lincecum left and subtract the RE of none-on, and same number of outs to remove any possible runs that Lincecum would not be responsible for that may score in that inning.

david k
Guest
david k

I was thinking the same thing, and also, it depends on how deep the starter can get into a game. If he makes it to the 8th, then he will likely be replaced by the “set-up guy”, probably he 2nd best reliever on the team. If he only makes it to the 6th, he’ll probably be relived by the 4th best reliever on the team, or worse.

I don’t have the time to look at Lincecum’s starts right now to research this myself, so it’s all conjecture on my part.

Bip
Member
Member
Bip

Smart managers often have a kind of “fireman” reliever who probably isn’t as good as the closer but may be the second or third best reliever. This guy will not be dedicated to the 8th, but will instead come into a jam, when the opposing team’s chance of scoring is highest. If not that, then managers will instead play to matchups to get out of the inning, bringing in their lefty killer if a LHB is coming up, for example.

I don’t know if the Giant’s did any of this or if they stuck to a rigid “4th best guy for the 6th, 3rd best for the 7th, etc.” pattern. I hear that Bochy is a decent tactician, so I presume the bullpen was handled better than that though.

Subfield
Guest
Subfield

There are four starting pitchers with at least 350 innings pitched over the last two years who also walked 10% or more of the batters they faced. Tim Lincecum is one of those pitchers.

There are three qualified starting pitchers who threw first-pitch strikes less than 56% of the time. Tim Lincecum is one of those pitchers.

There is one qualified starting pitcher who had the highest line drive percentage. Tim Lincecum is that pitcher.

TKDC
Guest
TKDC

“There is one qualified starting pitcher who had the highest line drive percentage. Tim Lincecum is that pitcher.”

There is one qualified starting pitcher who had the third highest line drive percentage. Adam Wainwright is that pitcher. Is this really that useful a stat?

Bip
Member
Member
Bip

People have often quoted line drive percentage as if it is supposed to be some stable determiner of BABIP. It seems that the quality of contact matters more than the trajectory in actual fact.

Wainwright does have a worse-than-average BABIP though, and his ERA is almost half a run worse than his FIP. It only looks like a bad comparison because Wainwright’s peripherals are much better than Lincecum’s.