'This item is sold out or not available for shipping.' Indeed.

Probably my favorite regular feature at John Sickels' excellent Minor League Ball blog is  "Prospect Retrospective," in which he takes a current or former major league player and goes back to see how he did (or would have) graded the player given what evidence was available at the time. Here is a recent example.

As a sideways sort-of tribute, I might start doing a short occasional "Contract Retrospective," which is pretty much exactly what it sounds like -- looking at a contract signing from the past while looking at the how the team might have valued (or misvalued) the player at the time. It's easy enough to criticize or praise a contract a few years into it, but how did they do at the time?

Given that it seems to be "Cleveland Week" (or was that last week?) here at Driveline Mechanics, I thought I would kick off this National Sensation by going over a contract signing that many Rust Belters probably see as a bad one now, particularly with the latest DL stint, but someone who was (and still could be) one of my favorite non-Royals: Travis 'Pronk' Hafner's four-year, $57 Million extension signed in July 2007.

Keep in mind that this isn't meant to be an exhaustive historical or statistical analysis by any means, just a look back at the past. And again, while this deal might look bad now, we're trying to get an idea of how good or bad an idea it might have been at the time it was signed.

In July 2007, the Indians signed Travis Hafner to a four year, $57 million dollar extension. It replaced the last year of the three-year deal he signed in April, 2005. To be honest, Cot's (where I'm getting the details I'm basing this on) doesn't make it clear to me (and I'm not smart enough to reconstruct his service time status) if any of the years of the 2007 extension would originally have been arbitration years, so I'm going toassume the 2007 extension just deals with years that would have been free agent years.

Given that the deal included part of 2007, it's difficult to reconstruct it precisely on the scale. Keep that in mind as well. For simplicity's sake, I'll use Tom Tango's handy free agency salary chart for the 2007-2008 offseason. We see that four years, $57M means that Cleveland was paying for a 3.5 WAR (Wins Above Replacement) player, assuming a 10% salary inflation a year, and a 0.5 WAR decline a year. Once the $400,000/year "replacement salary," is substracted, the total is pretty close to the $54.9 listed on the chart.

Was 3.5 WAR a reasonable expectation for Hafner? Let's do some very basic "retro-jection." He's was already a full-time DH, so worrying about his defensive numbers isn't a concern. He's not a a great baserunner, but he wasn't much worse than -1 a season at that point, according to Rally's numbers, so the main thing to worry about is what a DH gets paid for: hitting. As I mentioned in my post on DH salary last week, the a DH's value above replacement level is simply his offensive runs/wins above average (note that that post looked at yearly salaries, here, we're looking at the value of a contract as a whole).

So let's engage in a bit of fun "retro-jection" for Hafner using his numbers at the time. wOBA gives us a good offensive "total value"  stat that is easy to convert to linear weights runs above average. I'll use the version implemented at Stat Corner  on Hafner's player page since they give a park-adjusted version of wOBA (wOBA*) as well as the league average right there.

We can't just take the numbers from each season, of course. Following The Book (as usual), we need to regress each seasons' wOBA* against 220 plate appearances of league average. So I'll so that, then take a 5-4-3-2 weighted average of Hafner's performance 2004-2007. I'll also use some linear weights aging curves suggested by Tango, et. al. elsewhere, although that's a whole different discussion.

After implementing aging adjustments and a weighted average, we get a .390 projected wOBA. Given a similarly derived league average, we get a .344 wOBA, meaning that over a "full season" of 700 PAs, we can see Hafner as a about 28 runs above average. Assuming a (slightly generous) 10:1 runs:wins conversion, that gives us 2.8 wins over 700 PAs (which Hafner has never had in a season). If that's how we evaluate his deal straight up, it seems pretty far off, especially after adjusting for playing time, he seems like a less than 2.5 WAR player. But we can be a bit more subtle than that. Hafner had shown that he was "used" to DHing, so I think we can go ahead and assume that he wouldn't suffer from the "hitting off the bench" penalty, or if he did, his projection already assumes it. So instead of the -22.5 DH positional adjustment, it can be -17.5. Moreover, he's always been an AL player, so replacement level should be 2.5 wins/25 runs per 700 games rather than the standard 2.25/22.5 (or for the NL, 2/20). We get:

+27.8 offense - 17.5 positional adjustment + 25 AL replacement level = ~ 35.3 x 85% playing time = ~30 runs above replacement, or about 3 WAR.

That's better, but suggests something like 4/$45M rather than 4/$57M. That's still a lot of money, and even that might be generous given that 2007 was the first time Hafner had logged 85% playing time (700 PAs being 100% for our purposes) in his major league career and was already 30.

How else we might think about this? One obvious issue is that the contract was signed halfway through the 2007 season, and my "projection" used the stats from after the end of the season. Even with Pronk not lighting it up at that point in 2007, that's pretty early to be using the current season's stats on an equal footing with previous performance. So I re-did the "retro-jection" using 2003-2006 numbers and came up with:

After doing the aging adjustments and converting to runs above average, we get:

+31.1 offense - 17.5 positional adjustment + 25 AL replacement level = ~ 36..6 x 85% playing time = ~33 runs above replacement, or about 3.3 WAR.

That's pretty close.Still, starting with 85% playing time (about 600 PAs) seems optimistic, given that pre-2007, Hafner had never had 600 PAs in a season. Moreover, 3.3 WAR is probably a bit high, given that at the time, the AL runs to wins conversion was closer to 10.5 to 1 than 10 to 1.

There are a lot of other angles here -- some will want to say that Hafner had "old player skills." This is a good point that deserves better treatment than I can give in a post that is already about 3 times as long as I'd planned. Without getting into every little thing, I'd simply say that a) the assumed 0.5 WAR decline curve is fairly steep to begin with; and b) that Hafner had too high of a batting average (over .300 2004-2006) to fit in with the Dunn/Burrell group, and his strike zone judgment was too good for him to belong in the the Richie Sexson category.

Mark Shapiro, et. al., were and are a smart group. I still think they are one of the better front offices in baseball. This doesn't mean they are immune to mistakes, of course. Maybe sales of Pronk Bars were expected to recoup the shortfall, I don't know. It's easy to criticize the deal now, of course with Hafner on the DL again (although he was hitting well) and having missed most of last season. But that's not the point -- was the deal a good idea then? Tenatively, I'd say this: the deal wasn't good (from the team's perspective), but it wasn't as bad as you might think.

[Update: I'm a moron. Okay, that's not news. But I completely missed Jay from Let's Go Tribe's excellent interview with Chris Antonetti that is quite germane to the subject at hand.]