The State of the AL West
Spring Training looms, and with the exception of the somewhat odd Khris Davis deal, the AL West clubs seem to have settled in on the rosters they’ll take to Arizona. We’ve also got some early projections of each team, and now that the hot stove is down to a very low simmer, we can start to make more educated guesses about playing time. Put the two together, and we’ve got some early looks at the state of the division. If your initial reaction to this is indifference or a rueful scoff, I get it; the projections were pretty enamored with the M’s last year, and that didn’t work out so well. I think there’s always some value in seeing how the teams stack up on paper, knowing that chance and luck will make most of these look silly by August, if not sooner.
Going a step further, I think it’s interesting to compare the projections and see where they agree and disagree about the M’s and their competitors. As we’ve seen in recent years, the parity of the AL as a whole means the relative position of each team within the division means a little bit less – there are no great teams and no truly awful ones. EACH of the projection systems we’re looking at today sees the best and worst teams separated by 10 games or less. You can legitimately make a case for any of these teams, though obviously the case is a bit easier for the Astros and M’s than it is for Oakland.
This compression/parity/mediocrity/whatever you want to call it has a few ramifications. First, if true talent isn’t going to decide the division, then chance and how teams respond to it means more. That is, teams will need to decide very quickly what counts as a slow start, and what is an unacceptable level of performance in a tight race. The M’s have been remarkably poor at this in recent years, but given the samples, it’s almost impossible to separate noise from signal. The Rangers’ faith in Shin-Soo Choo and Elvis Andrus was rewarded last year, while the Athletics’ freakishly unlucky bullpen* never saw their luck change as their innings piled up. Second, differences in individual forecasts are similarly more important. Fangraphs’ projection for Nelson Cruz is pretty terrible, while CAIRO’s is more bullish. For Kyle Seager, they’re reversed. A lot of this has to do with how they apply aging curves (Nelson Cruz’s WAR by age is almost unprecedented outside of the steroid age), but regardless of why, each individual projection (and thus each individual season or performance) has a larger impact on how the teams rank within the division. Finally, with the division so bunched up, a big part of a team’s playoff odds come down to the strength of the division as a whole vis a vis the others divisions.
Each system sees the teams as tightly bunched up, but there’s some disagreement about *where* they’re bunched up. In CAIRO, the teams are centered on around 79 wins. The Astros 83 projected wins actually paces the division. This contrasts sharply with Clay Davenport’s projections, which see the AL West as perhaps the best division in baseball: the Blue Jays lead the league with 90 projected wins, but then spots 2-4 are all AL West clubs. To CAIRO, the teams are sharing 393 wins pretty evenly, but the Davenport projections show them fighting over 420 wins. Fangraphs comes in right at the midpoint. In a system with this much parity, small differences can get magnified. This isn’t just a mathematical oddity; how the clubs in the West and Central stack up impacts the M’s playoff odds. By CAIRO, if you don’t win the West, you’re pretty much out of luck. The Davenport win totals show three AL West teams with very good odds to make the postseason.
One final comment before we get to some tables: the other thing that parity does is put more emphasis on playing time estimates. Every team projection has an objective part and a subjective one. The projections endeavor to give us an objective look at true talent, but a human’s got to input some playing time estimates, and that’s both tough to do before spring training and hugely important to the overall winning percentage the model spits out. Fangraphs has their own estimates at their “depth chart” page, as does Clay Davenport on his projection page. CAIRO, the projection system from SG at the Replacement Level Yankee Weblog, doesn’t yet have these estimates for every team. He’s done one for the Yankees, which is fair enough, but not for the AL West teams. Using his awesome CAIRO spreadsheet, I’ve taken a stab at it for each club in the West. Fair warning: I’m sure fans of the other clubs might quibble with my guesses.** You tell ME how the A’s are going to use Khris Davis and Mark Canha. It’s just a guess, and it’s still only mid-February, but it’s important to acknowledge that small changes to the line-up – even bench players – move the needle by a win or two.
So, let’s take a look at how Fangraphs, Clay Davenport and CAIRO see the AL West as of Presidents Day 2016:
|Team||Wins -FG||Wins – Dav||Wins -Cair|
The Astros are generally, though not uniformly, the best team, with the M’s a close second. The other three teams are a few wins back, though it’s worth pointing out that the systems disagree on the Angels more than the other clubs.
Looking at run differential helps highlight how each system sees the AL West vis a vis the other divisions:
|Team||RunDif FG||RunDif Dav||RunDif Cair|
The M’s (and Astros’) run differentials stand out. By Davenport, the M’s are projected to score the 2nd most runs in the AL, behind only Toronto. By CAIRO, the M’s improvement in runs per game is much more modest – about 0.05 runs per game as opposed to 0.5. Both systems have identical forecasts for runs allowed, though. Fangraphs and Davenport are very close to each other regarding each team’s runs scored, while CAIRO’s much more bearish. On runs allowed, it’s Davenport and CAIRO that are nearly indistinguishable, while Fangraphs sees a lot more runs charged to AL West hurlers.
One final area of agreement: Fangraphs and CAIRO both see the M’s defense as a net positive, with the bulk of the contribution coming from the OF. Putting Nelson Cruz at DH and Leonys Martin in CF is a 1-2 win swing on its own, and the M’s D is forecasted to save around 5 runs by Fangraphs or 12+ by CAIRO. Martin’s defense is going to be important with so many fly-balling relievers, so it’s critical that Martin hit enough to stay in the line-up. I know he was hurt last year, but the projection systems all see him as a Brendan Ryan-type hitter/player. If he’s significantly better, the M’s are in much, much better shape. If not, they’ve swapped out a black hole at C and 1B for one at CF.
(FG = Fangraphs, Dav= Clay Davenport, Cair = CAIRO. I had nothing to do with creating these systems, I’m just using them. Huge thanks to SG for the editable playing time tab in his CAIRO spreadsheet.)
* I’ll come back to this in a future post, but one of the most significant sources of variance this season – and every season – is relief pitching. Not just “Does Steve Cishek bounce back, or nah?” but in a larger sense: how important to team success is a bullpen? The A’s “unlucky” bullpen absolutely sunk a team that was putting together a decent season by run differential (and it made Evan Scribner available to Seattle), while the Pirates and Yankees’ bullpens propelled them to contention. Depending on your point of view, this is either a huge part of “chance” – random good seasons from volatile set-up men can get you Baltimore’s 2012 or Oakland’s 2015 – or an important way to actually beat projections.
** Some obvious issues with my estimates here: Alex “Chi Chi” Gonzalez wasn’t forecasted, but his not-so-hot forecast shouldn’t do too much to the Rangers’ overall win projection. I have no idea how the A’s OF rotation works now, but gave Davis and Canha plenty of PAs. Wouldn’t be surprised to see the latter moved before opening day. The back of each bullpen is pretty speculative, but with so few innings at stake, it shouldn’t matter too much.