July 7, 2014

Pebble Hunting

Why Do Teams Pitch to Trout?

by Sam Miller

Here’s a story about Mike Trout and intentional walks: During his junior year, in the state playoffs, the opposing Cherry Hill East manager got so spooked by Trout’s batting practice display that he intentionally walked him to lead off the game; then again in his second at-bat, with a runner on first; and once more in his third at-bat, with the bases loaded.

There are very, very few stats in baseball that aren’t part accidental. A player hits 40 home runs—that’s telling, but that number is an accident; he’d much rather have hit 600 home runs, and the pitchers would have much rather he hit zero, and nobody was trying to hit exactly 40. Sacrifice bunts show intent, but plenty of sacrifices are called for but turn into pitches taken, pitches fouled, pitches whiffed, bunts popped out, etc. Stolen base attempts are close to a perfect expression of intent, though plenty of stolen base attempts don’t get recorded—when the pitch was fouled off or put in play.

Intentional walks, though: Every one of those was intentional. There are no accidental intentional walks, and but for the rare exception every five years or so, there are no thwarted intentional walks. Intentional walks tell us, with almost 100 percent accuracy, that an opposing team would rather face the man on deck (with an extra runner on base) than the man at bat. The Cherry Hill East manager had an opinion; he expressed his opinion; he acted on his opinion; we know his opinion, with certainty.

So what does this say: Despite being the best hitter in baseball over the past three years, Trout hardly ever gets intentionally walked. He leads the American League in OPS+ this year and has been intentionally walked just once. Howie Kendrick has been intentionally walked twice.

There are probably several factors:

• He’s right-handed, and right-handers get intentionally walked something like two-thirds as often as left-handers.
• The guy behind him has generally been right-handed, so there’s no platoon-advantage-based strategy to justify it. Since the start of 2013, there have been nearly 1,200 intentional walks, ignoring those that precede a pitcher batting. About 60 percent of those switched the platoon advantage from the batting side to the pitching.
• He’s super fast, which makes it slightly more costly to put him on base.
• He bats second or third, which means it isn’t just the hitter after him who tends to be qualified to provide protection; it’s the series of hitters after him, all of them better than the bottom-of-the-order types a pitcher might salivate over.
• And intentional walks in general keep going down, to the lowest per-game average in IBB-recorded history. That’s probably in part because of the league’s general appreciation for the fact that intentional walks aren’t often justified, and in part because of the lower run scoring environment of the past few years. As The Book noted, the intentional walk is favored slightly more in high-offense environments (though it doesn’t generally change the calculus by much at all). Trout doesn’t live in an IBB era.

But here’s another hypothesis: Trout is still underrated. And the guy behind him, his lineup protection, the guy who has been intentionally walked seven times this year, is still overrated.

In Albert Pujols’ best years—2003 through 2010—he had an OPS+ of 177. Trout’s OPS+ since he was promoted to the big leagues in April 2012 is 175. Unless you’ve got a beef with the park and league adjustments that go into OPS+ (and True Average, which assesses the players similarly), there’s very little difference there. And yet if Trout were intentionally walked in each base/out situation at the same rate that Pujols has been in his career—note: in his career, not just in his best eight years, when he was comparable to Trout—then Trout would have eight intentional walks this year, not just the one:

Split	Pujols IBB Rate	Trout Expected IBB	Trout Actual IBB
0 out, ---	0.0%	0.00	0
0 out, 1--	0.0%	0.00	0
0 out, -2-	2.2%	0.30	0
0 out, --3	8.7%	0.09	0
0 out, 12-	0.0%	0.00	0
0 out, 1-3	2.1%	0.08	0
0 out, -23	22.2%	0.00	0
0 out, 123	0.0%	0.00	0
1 out, ---	0.1%	0.07	0
1 out, 1--	0.0%	0.00	0
1 out, -2-	14.1%	1.69	0
1 out, --3	18.2%	0.91	0
1 out, 12-	0.0%	0.00	0
1 out, 1-3	2.9%	0.12	0
1 out, -23	44.4%	0.89	1
1 out, 123	0.0%	0.00	0
2 out, ---	0.2%	0.10	0
2 out, 1--	0.0%	0.00	0
2 out, -2-	28.6%	2.57	0
2 out, --3	17.6%	0.53	0
2 out, 12-	0.5%	0.05	0
2 out, 1-3	1.7%	0.15	0
2 out, -23	38.9%	0.39	0
2 out, 123	0.0%	0.00	0
Total	3.2%	7.93	1

To reiterate: that’s based on Pujols’ career IBB rate. In 2010, at the end of his most dominant eight-year stretch, when he had the exact same True Average that Trout has this year, he was passed intentionally 38 times, about 70 percent more frequently than his career rates.

We had plenty of factors for Trout’s low intentional walk rate. How many of those apply to Pujols in 2010?

• Right-handed? Check.
• Guy behind him also right-handed? In the Cardinals’ lineup that year, yes. (And of Pujols’ 43 intentional walks since the start of 2011, just six created a new platoon advantage for the defense.)
• Super fast? No, though Pujols stole 14 bases and had a reputation for good baserunning. He produced 3.5 baserunning runs above average that year; Trout has produced just .5 runs on the bases this year.
• Bats second or third? Third, which isn’t quite the same as second (the different words for each spot are a dead giveaway) but is similar (different words notwithstanding).
• Low-IBB era? Not as low—an intentional walk every four games, rather than the current five—though on the downward slope. The run-scoring environment in the National League in 2010 was very similar (4.33 runs per game) to the American League in 2014 (4.27 runs per game).

So if Trout and Peak Pujols are comparable hitters, and they shared key characteristics and circumstances, then the difference must be the guys batting behind them, right? Trout has Pujols, a pretty good hitter with substantial power, batting behind him. Pujols, by contrast, had Matt Holliday, a very good hitter with substantial power, batting behind him. Indeed, the difference between Trout and Pujols now is larger than the difference between Pujols and Holliday was then:

Pujols, 2008-2010: .367 True Average
Holliday, 2008-2010: .324 True Average
Difference: .043 (12 percent)

Trout, 2012-2014: .359 True Average
Pujols, 2012-2014: .293 True Average
Difference: .066 (18 percent)

Note: I didn’t weight these to put more emphasis on recent performances, as a more rigorous process might have. If I had, then all four numbers would have changed to benefit Trout in this comparison—his TAv would go up, both of Pujols’ would go down, and Holliday’s would go up. I didn’t do that, because the gap here is already substantial enough to make the point without adding an extra layer of making stuff up as I go along.

So that’s pretty compelling: Opposing managers were more afraid of Pujols then than they are of Trout now, even though Trout is roughly as good; and they’re more afraid of Pujols now than they were of Holliday then, even though Holliday was substantially better.

(Aside: Google “ever pitch to Pujols,” as in “why would you ever...,” and there are 19 non-duplicative results.

• “Ever Pitch To Ortiz”: 16

• “Ever Pitch To Stanton”: 9

• “Ever Pitch To Cabrera”: 8

“Ever Pitch to Trout”: 1)

None of this is to suggest that managers should be walking Trout more. The Book prescribes intentional walks to a hitter of Trout's ability in front of a hitter of Pujols' ability in very, very few cases, such as runners on second and third with one out in the ninth inning with certain leads. What it suggests is that we should expect managers to be walking Trout more. Recall: We have about a 66-point True Average gap between Trout and Pujols, if we take their past three seasons (unweighted) as each player’s true talent. Managers routinely issue intentional walks to gain an advantage that is much smaller than that.

We looked at all the nearly 1,200 intentional walks issued in the past two seasons, excluding those followed by a pitcher batting. We used PECOTA, and each player’s multi-year splits, to derive a true talent for each potential matchup—the batter’s true talent against the handedness of the pitcher walking him, and the following batter’s true talent against the handedness of the pitcher who ended up facing him (either the same pitcher, or one brought in from the bullpen to face the man on deck). Example:

David Aardsma intentionally walks Eric Chavez to face Cody Ross. Chavez’s .253 overall True Average is adjusted to .272 against right-handed pitchers; Ross’ .266 True Average drops to .248; so Terry Collins’ decision gains him a 24-point edge (at, obviously, a cost of a baserunner).

That example is just a little bit less advantage than the typical intentional walk gains: About 29 points, since the start of 2013. In fewer than 15 percent of walks during that time has the gap between the batters been as large or larger than the 66-point gap we’re crediting to Trout over Pujols.

So what the heck? Logically, one of two things: Either Trout isn’t actually 66 points of True Average better than Albert Pujols, or Trout isn’t perceived to be 66 points of True Average better than Albert Pujols. PECOTA, still skeptical that anybody can be as good as Trout has been, supports the former; it considers him a .330 True Average hitter, with a minimal platoon. When he’s facing a right-hander, our method assigns him a .328 true talent True Average. Pujols, meanwhile, still projects to a .314 True Average, with a small reverse split that makes him a .316 hitter against right-handers. In this way of looking at the players, there’s almost no difference between them; that, given this information, almost no manager would walk Trout to face Pujols. (About 27 percent of intentional walks produce a gain of 12 points or smaller, though many of those are heavily influenced by situation—late in games, where the cost of allowing one run is disproportionately larger than the cost of a big inning.)

This goes to the fundamental question of projection systems, and of PECOTA’s long memory: Does it save us from our tendency to focus too much on the present and forget too quickly? (Yes!) Or does it sometimes change its mind more slowly than we are capable of doing based on new information? (Yes!)

In this case, a smart projection system regresses Trout a bit, because a player who produces consistent .360 True Averages at ages 21 and 22 is an every half-century type of player, the sort of player whom the smart odds will never favor existing. And yet, everything we’ve seen in the past 27 months suggests an every half-century type of player; regressing him toward a league-typical standard implies that he belongs in the same league as the Jason Kubels and the Michael Cuddyers and the Matt Kemps of the world. He doesn’t. He’s playing here only because a better game hasn’t started that can challenge him, just like when he was facing Cherry Hill East in the state playoffs. Is there anybody willing to bet the under on a .330 True Average from Mike Trout?

And, in this case, PECOTA credits Albert Pujols with sharing the same tissue as the 2009, 2010, 2011 Albert Pujols. That .314 True Average—well, he hasn’t topped that (or come all that close) since 2011, three sub-.314 seasons ago. There is no part of him—approach, health, athleticism, performance--that hasn’t sloped steadily downward since then, and the league-wide trend toward infield shifts against the most extreme-pull right-handers seems to have made him particularly defendable. Is there anybody willing to bet the over on a .314 True Average from Albert Pujols?

The Indians weren’t. They intentionally walked Trout in front of Pujols this year, in late April. It wasn’t a case where the situation demanded it. It was the fourth inning, Corey Kluber was pitching, there were a couple guys in scoring position and one out. Terry Francona didn't even wait for Kluber to fall behind in the count; he immediately passed on Trout to face Pujols. (Eric Wedge and Bo Porter made similar decisions last June, when Pujols was more visibly hindered by leg injuries.) As the cameras cut to Pujols in the on-deck circle, the commentary from the Indians' broadcast crew was what you'd expect:

Matt Underwood: How many times has a batter ever been intentionally walked to face Albert Pujols?
Rick Manning: How many times do you walk a guy to get to a guy who has 500 homers?

Probably not very often, and that’s the point. Most guys who have 500 homers have one thing in common: They old. They’re not nearly as good as they were when they had 100 homers, 200 homers, 300 homers. Most of them have ceased to be the guy you walk, and should be the guy you consider walking to get to. I believe in the Pujols we see. I believe in the Trout we see. I believe the gap between the two is pretty big; given that gap, I believe that if Trout weren’t 22 and if Pujols didn’t have three nicknames on his Baseball-Reference page he’d be walked intentionally a lot more. Again: Managers probably shouldn’t walk Trout any more than they currently do. But by the standards managers go by, we'd expect them to. Given those standards, why would you ever pitch to Trout?

Huge thanks to Rob McQuown for research assistance/heavy lifting.

Sam Miller is an author of Baseball Prospectus. Follow @SamMillerBB
Click here to see Sam's other articles. You can contact Sam by clicking here