Marcus Semien’s Defense Is a Team Effort

Let’s start with the obvious: Defense hasn’t been the big surprise of Marcus Semien’s 2019. That story was written a few years ago. The big surprise of Semien’s 2019 has been that in consequence of a .373 wOBA and 7.5 WAR, he has precisely doubled the value he produced last year in the second-best season of his seven-year career. I wrote about Semien’s offensive breakout in May, noting his much-improved plate discipline, and suggesting that its probable cause — a change in approach at the top of the strike zone — augured continued success. Dan Szymborski picked up the thread in July, finding much the same, and also discussing Semien’s defense at some length. Dan concluded:

[D]efensive numbers are volatile, so having a second year of improved defensive numbers significantly betters the chances that Semien’s reinvention with the glove is for real. That the improvement is largely driven by error rate is an even more promising development because though errors themselves aren’t a great measure of defense, error runs tend to be more predictive on a year-to-year basis than range runs. This isn’t surprising given that range numbers necessarily have to jump into evaluating theoretical plays that never happened. In error runs, Semien’s +4.6 ties with Paul DeJong for the best in baseball at any position in 2019. If he continues on this pace, he will have added roughly 20 runs compared to the 2015 season, simply from avoiding errors.

Going into the final day of the season, Semien was +6.9 ErrR, which put him 19.5 runs ahead of 2015’s abysmal mark, and contributed no small amount to his overall WAR figure (which, incidentally, Dan optimistically projected for six runs in July, and which Semien blew past in August). In this piece, I want to pick up on something a commenter on Dan’s piece pointed out: that the major improvement in Semien’s defensive numbers came not right after his much-ballyhooed 2015 heart-to-heart sessions with Ron Washington, but over the course of 2017 — right when Matt Chapman became the A’s everyday third baseman.

Marcus in the Field
Year ErrR DRS UZR/150
2015 -12.6 5 -12.2
2016 -3.5 -6 -5.5
2017 0.5 -9 -6.2
2018 0.0 9 6.4
2019 6.9 3 5.0

Chapman took over for good halfway through the 2017 season, and the numbers diverge some overall, so I hesitate to put too much heft behind Chapman’s position at the hot corner as the sole explanation for Semien’s improvements. Semien himself clearly valued Washington’s instruction, and it’s certainly possible that it took a solid year or so to put the lessons fully into practice. And whatever benefit playing next to Chapman might have given Semien (we’ll get into that shortly), it was clearly easier to take advantage of because Semien was a better defensive player in 2017 than in 2015. This is a “both and” situation. Still, the dynamic interested me enough that I asked Semien and Matt Williams, the A’s third base coach, about it last week. They both saw something there.

“I think the deeper he [Chapman] can play, the closer to straight up I can play,” said Semien. “And that really helps me going right to left. I think he has the ability to play so deep because he reads guys who are bunting and their mannerisms before they do it. He has a good sense of that. He’s really good at reading hitters and reading what’s different about this, so even when playing way back he’s already on the bunt. It’s been a good combination these last two years.”

Semien told me what should be obvious to anyone who’s watched him play defense: That he’s more comfortable going to his backhand (towards third base) than his forehand (towards second). Chapman’s range and overall skill at the position to his right — his 8.3 ErrR this year is first to Semien’s third — mean Semien has been able to play far closer to the second base bag than he had earlier in his career. That makes plays to his left — up the middle — much more within range, and allows him to stretch into the no-man’s land between him and Chapman on his strong hand.

“Chappy helps him because he covers great ground,” said Williams, who joined the A’s at the beginning of the 2018 season. “At the same time, Marcus has become an elite shortstop because he’s paid attention to the details. Not just catching the groundballs. He’s past that. He’s past having any confidence issues with regard to throwing or catching it or anything like that. Now he can just simply play, and he’s taken it to the next level where he’s actually anticipating. Over the last couple years, he’s really taken to that. Understanding hitters, understanding our pitching staff, anticipating a pitch that’s thrown, and being in the right spot.”

Inside Edge suggests that in 2019, Semien has achieved career-high success rates on Even (40-60%), Likely (60-90%), and Routine (90%+) plays. Meanwhile, at least in part due to his positioning closer to the middle of the field, more of the groundballs he’s faced have fallen into one of those three categories than ever before: 92.2%. Some of that is just selection — balls in the hole that previously might have fallen into the Impossible or Remote categories aren’t even within the realm (or reach) of possibility any more, and fall within Chapman’s remit. It doesn’t seem to have hurt the A’s team defense overall; they’re still, by any reasonable metric, among the best defensive units in the game.

“I think that since Chappy has been here, both of their confidence levels have risen,” said Williams. “That means that they’re certainly comfortable playing next to each other, they communicate well. They pick each other up with regard to a grounder that is hit to that side. They’re both very studious in the scouting aspect of it. They understand the hitter, the pitcher, and what they need to do defensively. But most of all, they take great pride in it and understand that the reason that we are where we’re at is, granted they’ve both got 30+ homers, is because they’ve played great defense. The little things, they’re very attentive to. It comes from a lot of work. Lots of work. And understanding what he can accomplish. There’s a reason he’s turned himself into an elite shortstop and it is that he doesn’t have to worry about the minutiae of that.”

Semien, for his part, is pleased but not satisfied with his success in the field and at the plate. “I’d like to keep working on throwing, finishing my throws, no matter what grip I end up in my hand, knowing that you need to finish your throws so that the ball stays through. Catching the ball I’ve felt pretty good, but throwing is what I’ve been up to this year.” Anyone who’s watched Matt Olson stretch to rescue a Semien throw will know that. But then, that’s the same story again: These A’s can catch it and they can throw it, and they do it as a team. They’ve won 97 games two straight years, and later this week they have a chance to finally break through the one-game playoff and play their first full postseason series since 2013. Team defense will be a big part of the reason why.

We hoped you liked reading Marcus Semien’s Defense Is a Team Effort by Rian Watt!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Rian Watt is a contributor to FanGraphs based in Seattle. His work has appeared at Vice, Baseball Prospectus, The Athletic, FiveThirtyEight, and some other places too. By day, he works with communities around the world to end homelessness.

newest oldest most voted
London Yank
Member
London Yank

I’m rather dubious of infield defense numbers across the board, because a huge part of infield defense is positioning, and this is now largely set by the teams before each pitch. The numbers may reflect something real about defense and run prevention at the team level, but assigning all of the variance to individual players is probably not right.

Towel
Member
Towel

As stated here in the article:

“Meanwhile, at least in part due to his positioning closer to the middle of the field, more of the groundballs he’s faced have fallen into one of those three categories than ever before: 92.2%.”

…positioning is clearly baked into the pie.

London Yank
Member
London Yank

My point is that players should not gain or lose credit for positioning that is set by their coaches.

Towel
Member
Towel

Being in the right spot has always been a part of defense. I see no reason why Matt Chapman shouldn’t get credit for his ability to play deeper than anyone else. And I see no reason that Devers should lose credit for trying to do the same and failing.

London Yank
Member
London Yank

Positioning used to be the prerogative of the players themselves. Now it almost entirely set by the teams. If Chapman was traded across the bay to the Giants and they forced him to play shallower do you think that is Chapman’s fault?

Towel
Member
Towel

I think it should be reflected in his numbers yeah. That’s what they are there for. If another team had Matt Chapman bunting the runner over instead of trying to hit the ball it’s reflected in his wrc+. Why should defense be any different?

London Yank
Member
London Yank

I agree that manager’s decisions are reflected in batting numbers like you correctly point out. However, the difference is the magnitude. 100% of defensive plays are influenced by coaching decisions these days. Only a small fraction of manager decisions affect wRC+. These numbers are meant to help us distinguish between player ability, but in the case of the defensive numbers they actually help us distinguish between coaching staff abilities.

London Yank
Member
London Yank

I’ll add that incorporating coaching decisions into WAR is inconsistent with the philosophy of this website at least with respect to pitchers. The elements of FIP were chosen because they are thought to be things that pitchers alone have control of.

Towel
Member
Towel

Pirates had Cole throwing his worst pitches in the wrong part of the strike zone. That was reflected in his FIP. Cole gets traded to a better org who changes how and what he pitches. Also reflected in his better FIP.

Defense is no different.

You find a way to take FO/ coaching decisions out of WAR and someone will probably pay you a lot of money. Good luck!

London Yank
Member
London Yank

It is easy to take coaching decisions out of WAR for defense. Just consider the difficulty of each play with respect to where the player starts. I’m sure most teams already do this internally and that is why their defensive valuations are often quite different than the published valuations.

Towel
Member
Towel

Thats what they do now

London Yank
Member
London Yank

It is not what they do now: https://blogs.fangraphs.com/the-fangraphs-uzr-primer/

For UZR the difficulty of a play being made is determined by a database of balls in play and whether they were hits or outs. UZR will consider some plays to be almost impossible to make even if they are hit right at the fielder, and it will consider some plays routine even if the fielder is starts off miles away from ball.

Towel
Member
Towel

UZR uses BIS data and BIS takes positioning into account.

London Yank
Member
London Yank

Are you sure? I can find no evidence that UZR takes the starting position of the fielder into consideration. Here is the description of all the things that are taken into account, none of which are the starting position of the fielder:

“How does UZR determine how much credit, positive or negative, to award a fielder on each batted ball? First it goes through 6 years of batted ball data and determines how often each type and location of batted ball is fielded by each defensive position, making adjustments for the speed of the ball, and the handedness, speed, and power of the batter. Later on, further adjustments are made, such as the outs and base runners, and various park adjustments, like the size and configuration of the OF, the speed of the infield, and the speed of batted balls in general, as influenced by temperature, altitude, and the ground ball percentage of the pitcher (e.g. ground ball pitchers allow easier to field ground balls and harder to field air balls). For example, UZR might find that from 2004-2009, of all hard-hit line drives hit by a LH batter with above-average power to a certain location in an average OF, 15% are fielded by the CF’er, 10% by the LF, and 75% fall for a hit. Remember, those would be average numbers across all MLB parks.”

London Yank
Member
London Yank

I was doing a bit of googling around this issue and I found the following article here on FanGraphs from 2015 by Neil Weinberg that makes the same point that I am making here. Weinberg’s conclusion in 2015 was that he hoped statcast data would eventually be incorporated so that player starting position could be factored into the measurements.

https://library.fangraphs.com/the-difference-between-range-and-positioning/

nevinbrown
Member
nevinbrown

Reading this thread lets me know that you did not play baseball at any real level. The fact that you think a LARGE portion of infield defense is positioning is laughable. I don’t want to sound like that guy but you’re putting an absurd amount of emphasis on the statistical element of baseball and completely diminishing the athletic components of what goes on. It is honestly disrespectful to the professional athletes that play the game. Just standing in the right position doesn’t mean you’ll make the play. It takes thousands of reps to be able to do that. The positioning doesn’t tell you the spin on the ball. How to play the bounce. How to make the throw off balance. You’re basically saying the defensive components of WAR are irrelevant and shouldn’t be attributed to the players but rather some dudes that aren’t even on the field.

Cool Lester Smooth
Member
Cool Lester Smooth

Both UZR and DRS ignore positioning, meaning that positioning is absolutely not “baked into the pie” when it comes to WAR numbers.

London Yank
Member
London Yank

I find it remarkable how unclear people are on how these numbers are actually calculated. This is a shortcoming of the information providers rather than the community trying to learn from these numbers.