---

Monday, September 29, 2008

Why Did the 2008 Yankees Disappoint?(Offense Edition)

Coming into this season, the Yankees looked to be a contender for best team in baseball. Although they were taking a calculated risk in relying on some very young pitchers, they were returning the bulk of a team that has scored 968 runs last year.  Even with expected declines in their older players, they still projected to have the best offense in baseball.

When I ran the Diamond Mind Projection Blowout, they were the best team in the AL in most of the projection systems I used.  When I looked at the position players and the pitching staff I arrived at the same basic conclusion, the the Yankees should have won somewhere between 93-97 games.

As we all know, things didn’t work out that way.  The question I want to look at is why?  Were the projections wrong?  Was it injuries?  Was it a lack of testicular fortitude?  Let’s see if we can figure it out.

The projections I’ll be referencing here are the composite projections of several different systems, as detailed in this February blog entry.  I’m going to go through position by position looking at the projected starters compared to what the actual primary starters ended up doing, then looking at the bench as a unit. 

One other thing, while offense in the AL was down for most of this year compared to last year, the difference has narrowed signficantly of late to the point that it’s probably no longer statistically significant. The league average line of .268/.336/.420 equates to a BR/650 of 80 this season compared last year’s line of .271/.338/.423 which equates to a BR/650 of 81.  The BR(batting runs) I’ll be using here are context-neutral raw batting runs using linear weights, not position-adjusted or adjusted for base-out states. 

Catcher
Coming off his career year in 2007, Jorge Posada was projected to give back a fair amount of value in 2008.  Posada was projected to hit .286/.380/.469, a performance that would have been worth 76 batting runs according to linear weights using my assumed 500 plate appearances. 

Posada went down to a shoulder injury after just 195 PA, which made Jose Molina (Molino for the regulars) the primary starter.  While Molino had a strong defensive season, he hit like crap.  Molina hit .216/.263/.313 over 297 PA, which was worth 21 batting runs.  The difference between Posada’s projection and Molina’s actual performance as the primary starter was 55 runs, which was a 5.5 win underperformance.  BTW, disparities in playing time at all the positions will be accounted for when we get to the bench.

First Base
Another risk the Yankees took entering 2008 was that Jason Giambi would be able to
-Out-produce a disastrous 2007
-Play passable defense at first base
-Stay healthy all season

If you had asked me the likelihood of all three of those things happening, I’d have put it around the same likelihood as Mike Mussina winning 20 games. 

Giambi managed to achieve all three things for the most part.  His glove was bad, but not horrifically bad, and he was productive and healthy for most of the year, although 45% of his production came in the 45 games from April 22 - Jun 17 where he hit .319/.441/.694 (40 BR).  Over the other 100 games he played, he hit .213/.342/.414 (48 BR).  I don’t know if this type of split is particularly meaningful, so consider it just a random factoid.

The 2008 projections for Giambi weren’t bad on a rate basis (.245/.387/.474) but his playing time projections were pretty pessimistic (300 PA).  He projected to be worth about 46 batting runs because of that.  Instead, Giambi hit for a little more power but a little less OBP (.247/.373/.502), but thanks to exceeding his playing time projections he was far more valuable than he projected to be, exceeding his projected batting runs by 42 runs, a 4.2 win offensive upgrade.  The 2008 Yankees’ underperformance from their pre-season projections can’t be laid at Giambi’s feet.

Second Base
Pass.

Well, I wish I could pass anyway.  Robinson Cano seemed to be on the cusp of becoming the best 2B in the AL.  Cano projected to hit .308/.348/.482 and play plus defense and was at an age where he could still realistically be expected to improve.  Cano’s projected offense over 585 plate appearances would have made him worth 84 batting runs. Instead, Cano digressed in just about every facet of his game, hitting .271/.305/.410 over 634 PA, which was only worth 68 BR.  This was a downgrade of 1.6 wins.

Third Base
Boo.

Alex Rodriguez was justifiably the AL’s MVP in 2007.  He was rightfully expected to regress somewhat in 2008, projected to hit .300/.406/.574 over 650 PA, which would have been worth 123 batting runs.  On a rate basis, he basically hit his projection, but thanks to a leg injury that cost him 56 PA compared to his projection, he was 14 runs less valuable than projected, a downgrade of 1.4 wins.  It’s also a fact that Rodriguez’s performance in more crucial plate appearances were less productive than his context-neutral numbers show. This was not a problem exclusive to Rodriguez, so I’ll devote a section to that as well.

Shortstop
Derek Jeter came into 2008 projected to hit .307/.379/.438 over 600 PA, which would have been worth 86 BR.  A late hot streak pushed him closer to his projections as he ended the season at .300/.363/.408 over 668 PA, a total of 82 BR.  He was less valuable than projected on a rate basis, but by exceeding his projected playing time his overall value was only about four runs less than projected.

Left Field
Johnny Damon was shifted to LF due to his defensive decline, the seeming emergence of Melky Cabrera and Hideki Matsui’s bad knees.  This looked like a disaster in the making coming off a poor offensive 2007, although it did also seem like a defensive upgrade.  Damon projected to hit .280/.353/.423 over 585 PA, a line that would have been worth 78 batting runs.  He blew that away by hitting .303/.375/.461 over 623 PA.  So like first base, left field ended up as a net gain on the pre-season projections, to the tune of 15 runs, or 1.5 wins.

Center Field
After a promising rookie season, Melky Cabrera declined in 2007.  The projections expected him to bounce back, thanks in part to his youth.  Melky was projected to hit .282/.344/.406 over 550 PA for a total of 68 BR.  Instead, he got even worse, hitting .249/.301/.341 over 453 PA before mercifully being demoted to Scranton.  Cabrera’s line was a downgrade of 27 BR, or 2.7 wins.  And that’s all I have to say about that.

Right Field
Bobby Abreu came into 2008 projected to hit .277/.383/.439 over 600 PA, which would have been worth 89 BR.  He exceeded that by hitting .296/.371/.471 over 684 PA, worth a total of 101 BR.  Luckily for Abreu, we’re ignoring defense for now, and this was an offensive upgrade of 12 runs or 1.2 wins.

Designated Hitter
Hideki Matsui was penciled in as the primary DH coming into 2008 and his projection was pretty good, .287/.367/.477 over 500 PA, equivalent to around 74 BR.  Unfortunately, after a hot start to the season, Matsui’s knees gave out on him and he missed a significant part of the season and when he managed to come back was very unproductive, ending 2008 with a line of .294/.370/.424 over 378 PA for a total of 50 BR.  That’s a 24 run/2.4 win falloff.

So just looking at the starting nine, we see that the Yankees projected to hit .289/.372/.465 over 4870 PA, which would have been worth 724 BR.  Instead, the starting nine hit .280/.352/.443 over 4896 PA, for a total of 653 BR.  That’s a 71 run drop off or 7.1 win drop off.

The Bench
Projecting the bench is usually tricky, but at least on paper the Yankee bench looked to be serviceable this year, with Wilson Betemit, Morgan Ensberg, Jose Molina, Shelley Duncan, Brett Gardner, Alberto Gonzalez, Nick Green, Jason Lane and Chris Woodward the likely candidates to see playing time.  I had the bench projected to hit .247/.319/.404 over 1811 PA, which would have been worth 209 BR.  Instead, the collection of Betemit, Ensberg, Posada,  Duncan, Gardner, Gonzalez, Justin Christian, Francisco Cervelli, Juan Miranda, Chad Moeller, Xavier Nady, Ivan Rodriguez, Richie Sexson, Chris Stewart and Cody Ransom hit over .244/.306/.380 over 1339 PA for a total of 140 BR. That’s a 69 run underperformance.

Clutchness
Regular watchers of the 2008 Yankees as well as regular readers of this blog also know from a variety of different methods that the Yankees underperformed with runners on base.  Using the method I like best (Fangraph’s batting runs above average based on run expectancy), here’s a look at how the Yankees’ players context neutral batting runs compared to their contextualized batting runs in 2008.

Overall, the difference was small, but I think we can all pick out several games where the Yankees failure to come up with hits in big spots cost them wins.

cnBRAA: context-neutral batting runs above average.
cxtBRAA: contextualized batting runs above average.
Diff: cxtBRAA - cnBRAA.

I don’t know how good or bad the team baserunning has been aside from stolen bases, which are already included in the batting runs above, so we’ll assume they were average.

Let’s roll all that up.

So we had a team that was supposed to score 933 runs or so.  Their starting nine created 71 fewer runs than expected, and their bench created 69 fewer than expected, for a total shortfall of around 141 runs.  Subract another three runs for unclutchness.

933 run projection - 141 actual underperformance - 3 clutchness = 789.

How many runs did the 2008 Yankees score?  789. That’s creepy, huh?

If you were to rate the biggest reasons for the Yankees’ offensive underperformance, it’d be:
1) Losing Jorge Posada (55 runs)
2) Melky Cabrera sucking ass (27 runs)
3) Hideki Matsui’s knee causing him to both miss time and to underperform when he returned (24 runs)
4) Robinson Cano becoming Wayne Tolleson (16 runs)