Throwing the Book at the M’s Lineup
The M’s 2010 offense was a spectacle none of us will forget, however hard we might want to. If you’ve been reading for any length of time, you know that scoring runs is just one part of the WAR equation, and that run suppression is just as important. But clearly, the M’s gave their pitchers/defenders a hopeless job last year, and while regression to the mean will help the M’s score more runs, the gap between where they were and plain old ‘bad’ is still huge. This team’s upgraded the offense, particularly at DH, but while regression and luck may giveth, aging taketh away (Jack Wilson, Milton Bradley, Chone Figgins and Jack Cust are all 30+, and while Ichiro doesn’t fit traditional aging curves, the man IS 37). The M’s offense has a lot riding on a good year from its prospects, the health/improvement of Jack Wilson and a bounce-back year from Figgins or Bradley. All of this means deploying these meager resources wisely is more important to the M’s this year than it is for other teams.
Using a neutral projection system (I’ve chosen CAIRO, as it breaks down projections by batter handedness), I’ve taken a stab at an optimal line-up for this decidedly sub-optimal collection of hitters. Steel yourself, then take a look.
To construct this line-up, I’m following sabermetrician (and M’s consultant!) Tom Tango’s data-driven template. This post by Sky Kalkman is a great overview, and Tango/Lichtman/Dolphin’s book, er, “The Book“, devotes a chapter to batting order. Some of their research flies in the face of received baseball wisdom, particularly regarding the relative importance of the #3, #4, and #5 hitters. That said, it’s important to keep in mind that batting order exerts a minor influence over expected runs scored. This thread at Tango’s blog calculates one reductio ad absurdum: batting the pitcher in the clean-up spot.
So…I can’t delay this any more. Here’s the M’s line-up as of the evening of 3/1/11 by The Book. Versus left-handed pitching, the line-up would go like this (the number is the CAIRO projection versus left-handed pitching):
1: Ichiro .331
2: Gutierrez .322
3: Olivo .320
4: Bradley .360
5: Cust .329
6: Smoak .314
7: Figgins .307
8: Jack Wilson .305
9: Ryan .295
In this line-up, Jack Wilson plays SS while Brendan Ryan plays 2B. This gives the M’s a formidable middle-IF defense, which is a better thing to talk about than their batting projections. Milton Bradley hits in the crucial clean-up spot against lefties (assuming he makes the team) while Cust moves to the #5 spot in his (much) weaker platoon split. You could make a case for Cust in the #2 spot, as he’s got on-base skills but not much power versus lefties, and Gutierrez projection seems oddly low. To be clear, I’m making an editorial decision here, and this probably wouldn’t be the way you’d do it if you relied PURELY on these batting projections.
If you’re wondering where Michael Saunders would hit, well, good look finding a projection that has him near league average as a hitter; for a corner OF, that’s a bad sign. Even with an overall league-average projection, he’d be below average against lefties, so Cust stays in as the least worst option.
Against righties, things look a bit different:
1: Ichiro .338
2: Bradley .339
3: Smoak .319
4: Cust .368
5: Figgins .325
6: Langerhans .314
7: Kennedy .309
8: Olivo .290
9: Ryan .279
This construction takes advantage of Langerhans’ and Kennedy’s platoon splits, and pushes Brendan Ryan to SS, Adam Kennedy to 2B and Jack Wilson to the bench. Let me be clear: this is largely based on my opinion of Ryan’s defense vis a vis Wilson, and is NOT based solely on their hitting projections. The hitting projections are so close, defense can break the tie. Based purely on batting projections, you’d probably scrap them both and put Luis Rodriguez in the line-up; he projects to be about 5 runs better than Ryan over 300 PAs. Thus, I’m tacitly (or at least it WAS tacit) arguing that Ryan would add 5 runs with the glove over Rodriguez in half a season.
The gap in hitting between Langerhans and Gutierrez is about 0.20 points of wOBA, or around 5-8 runs over half a season. Again, if you want Gutierrez in the line-up against lefties and righties, you’re saying that his defense will more than make up for the platoon advantage Langerhans enjoys. That’s certainly a defensible position, but it would probably entail rejiggering the line-up to push Guti down to the 7/8 hole and pushing Kennedy up. It would also give the line-up THREE hitters with split wOBA projections under .300. Is Nick Franklin ready yet?
Anyway: what do you think? I’ve taken some liberties with regard to defensive projections – what would you change? Jack Wilson’s batting projections are superior to Brendan Ryan’s, but is this because Wilson’s aren’t weighting his recent seasons heavily enough? That is, should we push Wilson’s projections down because a standard aging curve doesn’t fully account for how bad he’s deteriorating? Should we give Michael Saunders a break and assume he’ll be closer to league average, and thus in the running as an option against RHP? How about Smoak? At what level of production would you remove Cust from the line-up against lefties? I’m sure even Cust thought he was going to face mostly righties, and many observers noted his poor performance against lefties over the past few seasons, but if there’s one thing the 2010 M’s hitters were adept at, it’s putting phrases like “poor performance” in context.
Some of the big take-aways for me:
* This line-up is still bad at hitting. CAIRO projects the M’s to score 606 runs, by far the worst in the AL and only 2 more than the hapless Astros.
* Projections for Saunders/Smoak/Moore/Ackley are pretty ugly. If you think CAIRO’s too pessimistic, check ZiPS. On the one hand, not getting a lot of production from the prospects means the team has to rely on Milton Bradley’s mental state and Jack Cust remembering how to hit lefties. On the other hand, projections of prospects have lower reliability than projections for veterans (see here), so these guys have a better chance of blowing their projections out of the water than the vets do.
* Milton Bradley’s been talked about as a potential DFA candidate, but he stands out as one of the few MLB-caliber hitters by CAIRO (and ZiPS).
* CAIRO/ZIPS/PECOTA all have the M’s right around 70 wins. Matthew Carruth’s projections at LL imply a team that’s a bit better than that. I tend to agree with Matthew, but again, it’s fun to pinpoint where the disagreements are. Justin Smoak’s a big one, as is Franklin Gutierrez – M’s fans tend to rate these guys as better hitters than the cold, unfeeling CAIRO/ZiPS. You see this reflected in the fan projections at Fangraphs. Gutierrez’ fan projected overall wOBA is about .020 higher than CAIRO’s projection, and Smoak’s is a massive .035 higher.
* The projections against RHP are so bad that it’s tempting to throw caution to the wind and start Ackley in Seattle. You could calculate the gap between his WAR projections and Brendan Ryan’s or Adam Kennedy’s, but what you really need to do is weight the expected WAR improvement against an extra year of club control. That’s a valuable thing, but so are runs on the scoreboard. Fans give these vastly different weights.
Comments
6 Responses to “Throwing the Book at the M’s Lineup”
Leave a Reply
You must be logged in to post a comment.
“Regression to the mean”? Don’t you mean progression?
Whether the results in game terms are positive or negative, motion towards the mean is still termed “regression” by convention.
Smoak’s projections are low from most system, in my opinion. He played injured the second half of 2009 and didn’t hit for any power at all. It’s almost a quarter of the data that systems are working with, except Oliver, which has MLEs for his college days. (Oliver only projects him at .322, so it’s not like he’s a stud over there.) I’m also mistrustful of the way projection systems do park adjustments over small samples, and Smoak is penalized for hitting in Texas over a 275 PA stint. By the way, ZiPS lists him as a lefty, so I’m not sure what to make of that projection. Even so, I think he’s probably only about a .340 wOBA this year. We’ll see what happens with him, but I think he will legitimately be the best hitter on the bad team this year.
606 runs, ~70 wins, depending on Milton Bradley to be a core hitter on this team, and Saunders, Gutierrez and Smoak doing a collective bellyflop is a prescription for Jack Zdurencik getting fired at the end of the year- since it basically means complete incompetence at assembling a credible offense (606 runs would be the M’s third worst full 162 game season ever, only behind 2010 and 1983).
If you’re three years into being a GM and you can’t assemble a better offense than when the Mariners were an expansion franchise (and not really a better team than then, either- the 1979 M’s were 67-95, with a Pythag of 70-92), you’ve failed at your job. And you can’t exactly blame Bill Bavasi for trading for Justin Smoak- and if that assessment is correct, that Franklin Gutierrez really isn’t a better hitter against RHP than a guy who’s a fringey MLB 4th OFer, Death To Flying Things is really a 4th outfielder like Endy Chavez.
marc,
Out of curiousity what was the Cairo projection for Saunders against righties? Was he that far off from Langerhans? Langerhans & Gross are not that appealing as back-ups for a team that has two starting corner outfielders that swing lefty sticks.
BTW, I’m not being lazy. I tried to look it up on my own first without much luck.
Lailoken – Saunders projection v. righties was .298, or essentially the same as Gutierrez’s.
eponymous – well, if they actually score 606, then I think you might be right. Again, we know where the FO would probably differ from these projections: Smoak and Gutierrez, maybe Saunders. If Smoak develops into some semblance of a middle of the order hitter and if Saunders is average-y, this team looks a bit different. Then there’s Ackley. So this team COULD beat the RS projection handily, but that’s true for any team.
philosofool – a .340 wOBA from Smoak is attainable and would legitimately change the line-up. I was pretty surprised by his projections as well, but he really hasn’t put up a lot of power for a 1B to date. Good point about park effects messing with the projections, but I think MLEs/chaining is just as big of a problem – remember the issues PECOTA had a year ago.