Does Outfield Alignment Actually Matter?

The idea for this post came about from reports last week that the Yankees were considering flipping Brett Gardner and Curtis Granderson on defense, with Gardner taking over in center and Granderson shifting to left field. Of course, Granderson had his forearm broken by a J.A. Happ fastball yesterday, so now the decision has been made for the Yankees, as Gardner will start the season in center field. However, the Yankees will still have to decide what to do when Granderson returns from the DL in May, and they aren’t the only team looking at pairing a couple of center fielder in their outfield this year.

The Angels are shifting Mike Trout to left field because they’re going to give Peter Bourjos a chance to play regularly. The Indians are moving Michael Brantley to left field because they signed Michael Bourn. The A’s are going to use Chris Young all over the outfield because they have Coco Crisp already. In some cases — Trout and Bourjos, for instance — the defensive excellence of both means that there’s probably no wrong answer, as either could play center field and have it look like the right decision. But, in New York, the question was a little more interesting, as Gardner is generally considered to be a better defensive player, and flipping their positions was a consideration based upon improving the overall defensive quality of the Yankees outfield.

It makes sense, after all, to have your best defenders play where the ball is going to be hit the most often, but how much does it actually matter? What’s the magnitude of the difference between having Gardner in center and Granderson in left versus the alignment the Yankees have run out the last few years?

Let’s start off by saying that this isn’t a question we can answer with exact precision. We can make some informed comments based on the data that we do have, but we all know that defensive data isn’t perfect. But, we do have some information, and we can glean some idea of the scope of the impact with an optimal outfield alignment.

I started off by asking Jeff Zimmerman for some data on balls in play — excluding home runs — distributions for the last three years. Here’s essentially how the total breaks down for 2010 to 2012:

Left: 119,713 – 31%
Center: 147,305 – 39%
Right: 114,032 – 30%

Included in those numbers are ground balls and pop flies caught by an infielder, however, and we don’t really care about those data points when trying to talk about the overall opportunities for an outfielder. So, if we just focus on balls in the air that are fielded by an outfielder, the numbers break down like this:

Left: 53,054 – 29%
Center: 78,460 – 43%
Right: 52,823 – 29%

Given how strongly the pull effect is on ground balls, it shouldn’t be too surprising that balls in the air appear to skew slightly more towards the middle than the entire population of balls in play. There’s likely some bias in these numbers since center fielders tend to take charge on balls that multiple players could get to, and a ball is more likely to be tagged as being fielded in center if the center fielder catches it, but it’s not going to be large enough to erase the conclusion that center fielders do have more balls hit in their direction than corner outfielders. Not exactly a new finding, but it’s at least good to know that the data supports common knowledge here.

So, now that we have something approximating a distribution spread, we can note that this data suggests that the distribution of air balls to the outfield is something like 30/40/30. And, interestingly, this corresponds almost perfectly to the distribution of putouts as well. Last year, Major League center fielders made about 12,000 putouts, while right fielders and left fielders made just over 9,000 each. So, we have both the distribution of chances and the distribution of outfield putouts coming in at something close to 30/40/30.

With that distribution in mind, we can now at the overall number of opportunities and get a sense for how many more plays a CF might be involved in than a corner outfielder over the course of a full season. Last year, 184,179 batters came to the plate. If we remove all the walks, strikeouts, hit batters, and home runs, we’re left with 127,055 balls that were put in play in some fashion, or right around 70% of the total number of plays. Of that subset of plays, 45% were hit on the ground, leaving us with just under 70,000 balls hit in the air — either classified as a line drive or a fly ball — in 2012.

That’s 2,330 balls in the air per team per season, and 10% of those are infield flies, which leaves with almost exactly 2,100 balls hit in the air to the outfield per team per season. Applying the 30/40/30 distribution, that would make the average chances per outfield spot on a team per season look something like this:

Left Field: 630
Center Field: 840
Right Field: 630

200 more plays per season for a center fielder than a corner guy. That’s an average that can fluctuate based on a team’s strikeout and groundball rates, of course, but it’s a nice number to keep in mind in terms of a baseline. Just based on the distribution of batted balls, a center fielder is going to get one more ball hit his way each game than a corner outfielder. Clearly, playing your best defensive outfielder in center is a good idea.

But, at the same time, it also highlights that the magnitude of who plays where isn’t all that enormous. For one, these are season totals at the team level, and very few players play every inning of every season. Last year, the only outfielder to play 1,458 innings — 9 times 162 — was Adam Jones. Only Alex Gordon and Hunter Pence also cleared the 1,400 inning mark. The average innings spent in the OF for the 90 outfielders who played the outfield most often last year was 1,015 innings, or about 70% of the total innings available. Even if we limit it to the top 60 outfielders in terms of innings — assuming every team has two regular OFs who play most of the time — we come out to an average of 1,160 innings, or 80% of the available playing time.

So, unless you happen to have two iron man candidates in the outfield, they’re not going to experience the full effect of those 200 additional chances for the CF. Accounting for a realistic amount of playing time, moving a corner outfielder to CF will result in something like 150 to 175 additional balls in play hit in his general direction. But, of course, not all of those balls will be catchable.

Last year, the overall BABIP on balls in the air for MLB hitters was .336, meaning that 1/3 of those extra balls in play are likely to go for base hits no matter what kind of alignment a team is running out there. So, we’re already chopping 50 to 60 chances off the total opportunity difference, just based on the fact that many of those extra chances are simply fielding a screaming line drive that is going to hit the wall no matter who is chasing it.

Now, we’re left with something like an extra 100 to 115 catchable air balls for a center fielder over the course of a full season. But, just like there are non-catchable balls, that total also includes a large number of plays that any reasonable Major League outfielder will be able to make. The selection process for Major League outfielders weeds out the completely hopeless, leaving us with a spread of talent that isn’t all that large.

Of the 107 Major League outfielders who were given at least 500 innings last year, J.D. Martinez caught the fewer number of balls that BIS graded as being “in his zone”, coming in at 82.4%, and that’s dealing with a small sample of just 108 chances. Of outfielders who have played at least 1,500 innings over the last three season, Logan Morrison grades out with the lowest rate of catching balls in his zone, coming in at 84.7%. Essentially, even the most immobile outfielders get to something like 85% of all the balls that BIS thinks that they should be able to get to.

On the other side of the scale, the top-end OFs are in the 95% range. This isn’t exactly apples to apples, since a large part of being able to be a premium outfielder is to run down balls out of your zone, but it gives us an idea of the spread of talent between even great and terrible outfielders. Even the very worst Major League OF is going to catch a majority of those extra ~100ish balls that are hit in his direction and have some chance of being caught. And, of course, in each case that we talked about, we’re not dealing with the differences between a big range guy and a guy who should be spending a lot of time at DH. The spread in talent between guys like Gardner (caught 93% of balls in zone from 2010 to 2012) and Granderson (91% of same) is even smaller, as we’re dealing with differences between players who are already selected for having been among the better defensive OFs in the game.

What’s the exact number of extra plays that the Yankees might get from switching the two? Or that the Reds might get if they determine that Jay Bruce is rangier than Shin-Soo Choo? That requires more precise tools than we have available, but we can make a pretty good guess that we’re talking about a handful of extra outs per year. Maybe it’s a dozen if there’s a big gap in skills, as there might be with Gardner and Granderson now. But, overall, it’s not that big of a deal. There are a lot of balls hit to the corner outfield spots too, and defensive value isn’t simply constrained to up-the-middle positions. There are more balls hit to center, but once you factor in how many extra plays we’re talking about and the spread in talent between decent Major League outfielders, we’re not talking about a huge number of extra runs.

The main point is to get the right guys on the field; how you align them once they’re starting is more of a minor problem. Don’t worry too much about whether a center fielder is being wasted in a corner spot. There are a lot of chances to go around, and center fielders don’t get so many more meaningful chances that there’s a huge advantage to be gained by shuffling players between OF positions. There’s enough extra opportunities that your best guy should go in center field, but if it isn’t patently obvious who the best guy is, don’t stress about it too much. In the end, just put them side by side in some fashion and the results won’t be drastically different.

Print This Post

Dave is a co-founder of and contributes to the Wall Street Journal.

86 Responses to “Does Outfield Alignment Actually Matter?”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Benzedrine says:

    Interesting that LF and RF end up with virtually the same percentage, always thought the LF would get more chances due to the league having more RHB pulling.

    Vote -1 Vote +1

  2. CircleChange11 says:

    but if it isn’t patently obvious who the best guy is, don’t stress about it too much. In the end, just put them side by side in some fashion and the results won’t be drastically different.

    Sometimes, certain statements just strike me as humorous or entertaining. Like this …

    So, basically, if you have guys that are roughly equal fielders it doesn’t matter which one goes in CF or LF.

    I think that pretty much goes without saying, when the real aspect is finding out which one serves best in CF. There are some differences, obviously, namely there’s more ground to cover and reading batted balls “right at you” is a little different than reading them at an angle or “off to the side” a bit. You can see certain fielders really struggle with batted balls hit right at them and not being able to decide whether to come in, stay put, go back, etc.

    But, if they’re roughly equal fielders, then yeah, it doesn’t matter.

    I give the Angles credit for being willing to start Trout and Borjous … I just wish they ha done that without acquiring Vernon Wells and/or trading Napoli.

    Another aspect to look at or consider is games played at CF vs. LF, is one more taxing than the other? Are there more collisions with the fence in CF or LF? More hammy pulls? Blah Blah?

    Vote -1 Vote +1

    • Krog says:

      Anecdotal comment from a guy who last played baseball in high school: I always struggled reading the ball off the bat while playing centerfield. Batted balls don’t fly straight, they have spin on them which can cause them to break in strange directions. Balls hit to the corners tend to have a similar break on them, but balls hit to center can do all kinds of crazy things.

      Vote -1 Vote +1

      • Jarrod Saltalamacchia says:

        As a former outfielder myself, I’ve always liked centerfield more because the balls stay more true. In right or left, the ball tends to hook differently based on right- or left-handed hitter, but center is mostly true all the time.

        Vote -1 Vote +1

    • Rob says:

      If the players are defensively equal, then they should have left Trout in CF.

      Vote -1 Vote +1

  3. MyrEn says:

    Interesting article concept and great exploration, but the eventual conclusion was a little bit disappointing. Basically says that if two guys have almost the same skills, it doesn’t matter where you put them. Seems obvious when you sum it up in one sentence.

    I’d almost rather see how arm strength in the outfield at each of the positions affects total bases or something like that.

    Vote -1 Vote +1

  4. Baltar says:

    I thoroughly enjoyed your data but don’t agree with your conclusion. The data still support the ideas that the best outfielder should play CF and it does make a difference. As for COF, since the number of balls is the same, it still makes sense to put the guy with the better arm in RF due to the longer throw to 3B, but this difference must be very small.
    In other words, accepted wisdom finally wins one.

    Vote -1 Vote +1

  5. Pete says:

    This article distinguishes between CF and corner outfield regardless of the pitcher over the course of a full season. It seems to me that it’d be important to look at BIP data when a lefty or righty is pitching in a given game. I would think when a lefty is pitching the data might be more 40/40/20 and vice versa for a righty. If that’s in fact true then “hiding” a poor outfielder in either left or right depending on the SP may be well worth it. Would be interested in your feedback/analysis.

    Vote -1 Vote +1

  6. Will says:

    Bryce Harper (17.7 UZR/150 in CF) was also shifted to LF to make room for Denard Span’s career 4.6 UZR/150 in CF. Good to know it won’t make any difference.

    Vote -1 Vote +1

    • TFINY says:

      While your point is valid, you really shouldn’t use a single year sample for UZR. Also something to note, that looking at Span’s UZR in the last three years (when he was not playing in the metrodome) his UZR/150 is substantially better.

      Vote -1 Vote +1

    • adohaj says:

      Good thing Harper will get to all those balls in left center that span wont!

      Vote -1 Vote +1

    • JCA says:

      this was not done as much to make room for a superior fielder in CF as it was done based on Rizzo’s and Johnson’s belief that CF is a more physically demanding position and that Harper, with his size now and in the long run, would wear down in CF with the extra running. Basically, this was to conserve Harper physically. It would not surprise me if this had somethign to do with the thinking on Trout and Bourjos.

      Vote -1 Vote +1

      • Jason B says:

        Especially with Trout coming into camp looking like one of those twins on motorcycles from the 1988 Guinness Book of World Records…

        Vote -1 Vote +1

  7. kevinthecomic says:

    The conclusion here is eerily similar to the conclusion regarding batting order: “The main point is to get the right guys in the lineup; what order you bat them in once they’re starting is more of a minor problem.” I recall reading a study that suggested that you could play an entire season with an inverted batter order (ie. change 1 thru 9 to 9 thru 1) and it would only cost you 1 or 2 wins. Meanwhile, we get managers obsessing over whether to hit a guy 3rd or 5th.

    A great example of not having the right guys in the lineup, as you all know, was the playoffs in 2011. LaRussa (Cardinals) routinely started Nick Punto at 2B instead of Allen Craig due to Punto’s superior defense. Punto was a superior defender but couldn’t hit water if he fell out of a boat.

    The American League World Series representative, the Texas Rangers and their manager Ron Washington, did little better with their decision to keep starting an injured Josh Hamilton and batting him 4th!! Nothing like a 0.214 BA out of the 4-hole to help you lose a World Series. If you had to keep Hamilton on the World Series roster (because you were afraid that he was one step away from being a homicidal maniac), then he should have been a bench player only and maybe been given an opportunity to have a “Kirk Gibson circa 1988″ moment.

    Vote -1 Vote +1

    • Dave Cameron says:

      Right – defensive alignment is basically the fielding version of batting order. That’s a good way to think about it.

      Vote -1 Vote +1

      • Sandy Kazmir says:

        Isn’t one of the points of sabremetric research to find these areas with which to sqeeze 1-2 marginal wins? If you’re in the sweet spot of the win curve and can squeeze a win through batting order optimization and another through outfield alignment then you’re talking about drastically improved playoff probabilities.

        Vote -1 Vote +1

        • kevinthecomic says:

          With respect to batting order you’re probably not improving your playoff chances. Remember, the 1 or 2 win advantage is if you INVERT the batting order for the ENTIRE season. No manager would ever do this, obviously. Managers tend to concern themselves with who bats 3rd or 5th or maybe whether to move a struggling 2-hole hitter down to the 8-hole. These decisions are made (and I am guessing here) maybe 30 – 50 games per year, depending on how ‘active’ the manager is. So, if inverting the lineup for an entire season makes a difference of 1 or 2 games, then screwing around with one or two players 30 – 50 times per year is likely negligible — to the tune of 1 or 2 wins every 10 years.

          Vote -1 Vote +1

        • Sandy Kazmir says:

          No manager will invert their batting order, but some might, say, give Drew Stubbs 393 PA in the 2-hole (one of the two most important spots) then watch him go on to provide a .237/.300/.401 line in front of a very good 3/4 combination. Using the Book linear weights I get a wOBA of .302 in that spot and if league average is .316 then you’re talking about a wRAA of -4.7 just for 393 PA of Drew Stubbs in the 2 hole. That’s a serious problem. That’s almost half a win that could have been provided by merely having an average hitter in that spot. You can poo poo these things all you want, but the fact is there are still plenty of teams that aren’t even close to acting optimally.

          I’d love to see Cistulli or one of the other fine authors here look at this from a broader scope. Hell, I’ll just do it myself.

          Vote -1 Vote +1

        • kevinthecomic says:

          I think you just made my point for me. The meaningful decision is who to put in the line up not where to bat him. Putting Drew Stubbs in the line up at all was the error since those numbers are well below replacement level. Batting him second only made it a little worse.

          Vote -1 Vote +1

        • payroll says:

          There’s also the reasoning that while moving a guy around in the order might net a loss in strategy, it regains that in production. For whatever reason (psychology would be my guess), guys may produce better or worse depending on their spot in the order. Same deal with moving Hanley back to SS. Probably will cost the Dodgers a few runs, but Hanley has, historically, elevated his game when his coach strokes his ego like that. The net result is, probably, zero, but a happier team.

          Vote -1 Vote +1

    • Nathaniel Dawson says:

      That notion you have about batting order is not true at all. There’ve been many people that have looked at simulations and how batting order affects runs, and what comes out of them is that an optimal batting order probably gets a team an extra 5 to 8 wins over a pessimal batting order. Simply inverting an order wouldn’t be as bad as that, because it keeps at least a couple of your best hitters near the middle of the order, but it would likely result in a few losses more for a team over a season.

      Vote -1 Vote +1

      • Neil says:

        The point about batting order, if I’m understanding his point correctly, is still a fine one. Even a “bad” line-up will usually have the best hitters at 3 and 4 – sure, we’d like them at 2 and 4, but that’s close. And the traditionalists still tend to value good players near the top. That’s why you only lose one or two wins.

        Vote -1 Vote +1

      • Tangotiger says:

        The difference between the absolute best and absolute worst batting order is close to five wins. I doubt anyone has run a sim and is showing eight wins.

        In any case no one would ever put the absolute worst order together. So, the difference between absolute best and “normally crummy” is 1 to 2 wins.

        Vote -1 Vote +1

        • Baltar says:

          Thank you for clarifying that, Tango.
          Though too much is made of batting order (mainly because everyone, including me, likes to suggest their own), 1 or 2 wins is worth having–enough to make a difference in not making the playoffs.
          I know the Rays use their own batting order optimization program for each game; I wonder how many other teams do.

          Vote -1 Vote +1

  8. Dan says:

    If you have two centerfielders and you put the CF and LF, then there are going to be plenty of balls in play that either of them could get to.
    How about some outside of the box thinking?
    What id you took your two centerfielders, played the LF and RF and expected each of them to cover additional area in CF. Then put you slower outfielder in CF.
    I’m not saying I think this is a great idea, but something I would be interested in hearing more about.

    Vote -1 Vote +1

    • TKDC says:

      People suggest this sometimes, but I think in reality most balls are either only catch-able by one player or are easily catch-able by 2 players. Very rarely do you see a play where two separate players could possibly make an outstanding catch. Thus, I think it would be a bad idea to put your worst fielder in the position where the most balls go.

      Vote -1 Vote +1

  9. Lothar says:

    On balls that are fielded by an outfielder, the percentages calculated in the article are incorrect. They should have been:

    Left: 53054 – 28%
    Center: 78460 – 41%
    Right: 58823 – 31%

    This does not change the conclusions of the article; it’s just a nit.

    Vote -1 Vote +1

  10. existing says:

    I am curious how ballpark plays into this. I would imagine that each ballpark has to be a huge factor. For example, I was surprised when the Red Sox wanted Crawford to play LF rather than RF because I figured his range would be better used in LF, but I also know that Theo knows a lot more about it than I do. I wonder how, if at all, each specific ballpark plays into this analysis.

    +8 Vote -1 Vote +1

  11. polka bear says:

    Would it hurt the Angels to put Trumbo in CF sandwiched between Trout and Bourjos.

    Vote -1 Vote +1

  12. Kinanik says:

    Is a missed ball the same for a corner outfielder and a center fielder? There were 231/261/435 triples to pull/opposite/center field. Does a missed ball in CF do more damage? Obviously the magnitudes won’t be huge either way, but it might bump up the importance a bit.

    Vote -1 Vote +1

    • Cookierojas16 says:

      Yes, I had the same thought. My gut feeling is that a ball that gets past a CF is much likelier to be a triple than one that gets past a C OF.

      Vote -1 Vote +1

      • stuck in a slump says:

        I wonder if this has to do with CF’s not always having the best arms like a RF, or if it’s completely the fault of the ball being hit to the deepest part of the park.

        Vote -1 Vote +1

  13. Matt says:

    Dave, you did an article recently about how hitters tend to go opposite-heavy on fly balls. So if you have 2 CFers in your outfield, why not swap your corners every at bat depending on the handedness of the batter?

    I’m sure there’s worry about diminishing a player’s effectivness by moving him back and forth like that, but still I’m surprised no team’s tried this yet. Seems like done right you could get a bunch of runs saved out of it.

    Vote -1 Vote +1

    • Matt says:

      In case it wasn’t clear, an example of this would be to have Bourjos always in center, but have Trout swap with the other corner outfielder on an at bat by at bat basis, going to LF for LHBs and RF for RHBs.

      Or to simplify you could just have them take a side depending on who’s due up and stick there for the whole inning.

      Vote -1 Vote +1

      • Ian R. says:

        I’m in favor of this purely because it would make for some hilarious box scores.

        Vote -1 Vote +1

      • Baltar says:

        I have often thought about something like this. I would go so far as to suggest moving OF’s according to the batter’s tendencies, which would often result in flipping.
        The box score, unfortunately, wouldn’t need to show this. You would just have the RF sometimes playing in LF.

        Vote -1 Vote +1

    • Mike Newman says:

      RE: moving OF back and forth

      Over a 162 game season, swapping the LF/RF depending on the situation would result in extra wear and tear from the running involved. The season is a grind. Heck, standing up in the outfield for a number of hours per week isn’t the best for a player’s legs and back. With 30? seconds between batters, the outfielders would have to run a couple of hundred feet to swap spots to not slow the flow of a game.

      Let’s say a lineup was 6 RHH and 3 LHH and 39 plate appearances occurred throughout the course of the game, that’s 13 trips back-and-forth or so. 13 x 200 feet is about half a mile. Half a mile times 150 games is 75 miles of extra running when players are already exhausted by September/October.

      Vote -1 Vote +1

      • Matt says:

        Interesting, and good point. Still think you could squeeze a few runs by switching depending on batters due up in the inning.

        Vote -1 Vote +1

        • Jay29 says:

          You could save the switch only for extreme options, based on spray charts and other data, to where it’ll really pay off. E.g., it’s not worth it to tire out your OFs for a percentage point advantage, but it might be worth for, say, a 3 and 4 hitter with vastly different FB tendencies.

          Vote -1 Vote +1

  14. PackBob says:

    Seems there is one more layer here, one that can’t be predicted. If you take Gardner and Granderson’s 93 and 91%, if those 2 extra flies per 100 that Gardner gets to makes a difference in the game result, say a 2-1 game in ninth, man on second and third, that could be 2 wins, important for a team on the playoff bubble. Then too, the 2 extra flies may be meaningless to a game result. Overall across many seasons there is little difference, but in one season it could make a difference if the game circumstances were meaningful to a win. If a couple wins might be the result of playing the slightly better outfielder in center, even if the odds are low for that happening, I think you do it.

    Vote -1 Vote +1

    • Record Keepers, Highly Leveraged Center Field Drop Division says:

      Pro, Josh Hamilton dropped a ball in a season ending game in Oakland against the Athletics that opened the door to an improbably comeback that saw the A’s take the division on the very last day of the season. Con, Coco Crisp dropped a ball in Game 2 of the ALDS that allowed the two runs to score in a ballgame that ended 5-4 in the 9th on a walkoff fielder’s choice.

      Every time you put an inferior defender in center you risk inferior defense, but, in a highly leveraged situation, apparently, you do when you put the superior defender there, too.

      Vote -1 Vote +1

      • baycommuter says:

        In the same game, Crisp caught a fly ball that Cespedes also wanted, and because Crisp has a noodle arm, a Tiger runner advanced from first to second and came around to score. Not Coco’s finest hour.

        Vote -1 Vote +1

  15. George says:

    Does your conclusion that it it matters very little who plays in what position for equally talented defenders in any way affect your positional adjustments in WAR calculations? It would seem that if there is really not much of a difference, the positional adjustment of +2.5 runs for CF and -7.5 runs for LF and RF should be changed. Or am I thinking about this completely wrong…

    Vote -1 Vote +1

    • Matt says:

      If you put equally talented players in CF and LF the LF-er is probably going to make up most if not all of the positional adjustment difference in UZR. He’s being averaged against worse defenders so he needs to have that penalty applied.

      Vote -1 Vote +1

    • Mr. Jones says:

      I’ve been looking around and can’t seem to find the run values for positional adjustments. Where might one find them?

      Vote -1 Vote +1

  16. ttnorm says:

    It is important to remember that the CF basically has to sprint on any OF fly because he is the 2nd line of defense on plays made by the RF or LF. If you take a guy without range and place him in CF, you will wear him out before May.

    Vote -1 Vote +1

  17. Jaker says:

    If we were to accept this conclusion, wouldn’t that kind of invalidate WAR, at least when comparing OFs. If you have two OFs with near identical .wOBA (and other offensive stats) but a difference in WAR of 1-2+ driven mostly by defense, isn’t this arguing that we should accept them as almost the same player?

    Vote -1 Vote +1

    • Bip says:

      The point isn’t that outfield defense doesn’t matter. The point is the value of defense in center is not that different from the value of defense in the corners. If you replace bad defender with one 1-2 wins better on defense, you’ll gain 1-2 wins. However, if they are already both on the field and you switch their outfield positions, it won’t make too much difference.

      Vote -1 Vote +1

    • Sylvan says:

      They *are* almost the same player. Batting average, our most traditional, old-school method of player valuation, is commonly calculated to three decimal places. That should indicate just how fine the differences are.

      1–2 WAR worth of defense is just one additional hit falling in per week. 1–2 WAR worth of offense is just one additional hit falling in per week. It doesn’t mean the difference isn’t “valid”.

      Vote -1 Vote +1

  18. chanelclemente says:

    Of course, there is always Alfonso Soriano to norm to.

    Vote -1 Vote +1

  19. Hank says:

    Dave an interesting article, but where is the conclusion (in terms of the #’s) You say “a handful of balls” = how many hits? (with some distribution of 1b.2B, 3B) = how many runs? – how much WAR?

    Also a couple of issues:
    – You completely ignore the arm component of defense.
    – You use BABIP to filter down the # of flyballs – the BABIP on flyballs might be .336, but it might not be that to all fields
    – You convert 29/43/29 to 3/4/3 when it is actually closer to 2/3/2. this slants the range a lot flatter. Why the heck would you round like that (and round that badly)? You went from a distribution that is closer to 50% more balls in CF to one that is 33% more.
    – OOZ and the BABIP “hits” matter as well whether it’s the ability to cut off an extra base on a hit or actually catch one of the BABIP “hits”

    This article, right down to the egregious rounding of the distribution seems to have been written with the conclusion in mind. Isn’t a few runs significant? If you can get a half win upgrade (~5runs) and get it without paying for it, that is no big deal?

    +21 Vote -1 Vote +1

    • Mrdaikon says:

      Not sure I follow your rounding estimates here.

      Dave states that the numbers might skew towards the center fielders slightly because of lazy fly balls that both the CF and a COF get to, and the CF then catches as per the dictates OF etiquette. So one can assume that the numbers are more like 43/<29…….. 30/40/30 sounds like a nice and quick cocktail napkin calculation to me.

      Interesting to me is that I didn't see any snark in the comments section about 29/43/29 actually adding up to %101. I'm glad to see that the internet appreciates these posts as intelligent/informed/thoughtful FREE SABR blog posts and not NASA shuttle trajectory calculations

      Vote -1 Vote +1

      • Mrdaikon says:

        Ugh. Typo. Sry.

        “So one can assume that the numbers are more like 43/<29……"

        Meant to type: 43/<29

        Vote -1 Vote +1

      • Hank says:

        Divide 29/43/29 by 14 instead of 10. Which one seems more accurate?

        By just rounding to the nearest 10% it skewed the corners up and CF down. It might seem intuitive to round to the nearest 10% but it led to a 50% error in his initial estimate of 200 flyballs. The problem is 29:43 became 3:4 when it is much, much closer to 2:3. Visually I would guess most people would look at it and think 3:4. And that becomes the difference between 33% more balls to CF and 50% more balls to CF

        So instead of 300 extra flyballs, he got 210… which he then just chose to round down further to 200 (more cocktail napkin rounding)

        The problem with this is when you do it multiple times and round liberally, you start injecting significant error in the #’s. At the end 100-115 became… 100. Among most of the steps along the way he chose to round down which in turn minimized the impact of the change he was trying to look at.

        Vote -1 Vote +1

  20. TX Ball Scout says:

    “Last year, the only outfielder to play 1,458 innings — 9 times 162 — was Adam Jones”

    In many baseball games, 8 defensive innings is the max. The way Jones reached that number was by playing in numerous extra-inning games. Look it up.

    Vote -1 Vote +1

  21. bluejays49 says:

    The pitch was at 82 mph and looked kinda like a slider, it wasn’t a fastball.

    Vote -1 Vote +1

  22. DQ says:

    so thorough and logical, good work

    Vote -1 Vote +1

  23. Dr. Chaleeko says:

    This is one of those things, to me, where it comes down to an individual team’s circumstances. The question comes down to at least these four factors:
    1) Attributes of the team’s home park
    2) Range of the third defender in the OF
    3) Handedness of the starting rotation
    4) Strength of the players’ respective arms.

    If the Yanks had Greg Luzinski as their third starting OF, they might want him in RF for home games (short RF porch, large LF), then Gardner in center to cover the extra room that Luzinski couldn’t get to.

    Obviously the opposite would be true in Fenway.

    The 1950s Senators, with all their lefties might want Granderson in RF.

    Vote -1 Vote +1

  24. david h says:

    Aren’t the zones for center fielders and corners different sizes? So while a guy like Logan Morrison might provide the lower limit for playable defensive acumen in a major league outfield, 84.7% is an incorrect baseline. He can get to 85% of balls BIS thinks a left fielder should get to, but probably would get to a smaller percentage of balls in the center field zone if he played there.

    Vote -1 Vote +1

  25. Joe says:

    Would the positional adjustment for WAR need to be “adjusted” then?

    Vote -1 Vote +1

  26. Ed says:

    “Last year, the overall BABIP on balls in the air for MLB hitters was .336, meaning that 1/3 of those extra balls in play are likely to go for base hits no matter what kind of alignment a team is running out there.”

    This feels really wrong. If changing the outfield alignment results in more outs, it’ll show up as a lower BABIP. I feel like expected BABIP should be the end result we’re trying to find, not the starting point.

    Vote -1 Vote +1

  27. Tim says:

    Delmon Young is clearly Chaotic Evil, and is a terrible outfielder. It’s only one datapoint, but it’s suggestive.

    Vote -1 Vote +1

  28. Zac says:

    Incidentally, Adam Jones was replaced defensively on June 27, July 16, July 25, August 1, August 22, August 29, September 4, September 17, and September 26th. The fact that he happened up at 1458 innings (because of extra innings games) is just a lucky coincidence. His team actually had 1483 innings of defense in 2012.

    Vote -1 Vote +1

    • Zac says:

      And there are team variations too. The Chicago Cubs apparently only had 1413.6 innings of defense. A team can have more or less than 1458 because of (1) games where they were losing on the road after batting in the 9th inning, (2) walkoff wins while at home in the bottom of the 9th, (3) extra innings games, and (4) complete games that did not make it 9 innings.

      Vote -1 Vote +1

  29. Sandy Kazmir says:

    A dozen marginal balls might not sound like a lot, but if they all go for doubles you’re talking about something like 15 extra runs. That’s quite a bit of difference. I would have liked if you had tied linear weights into this instead of doing some solid work, but then just falling back on your opinions without looking at the actual context. Interesting thought experiment and hopefully others will carry the torch and dig deeper here.

    Vote -1 Vote +1

  30. cs3 says:

    I would like to see a similar analysis for SS/2B/3B

    Vote -1 Vote +1

  31. Salo says:

    Then, shuldn´t we use the same positional WAR adjustment for all OF? and also compere defenders in all 3 spots the same to rate them?

    Vote -1 Vote +1

  32. Dimaggio says:

    If you have Dom, Joe and Vince then No; it does not matter. In all seriousness though this is a good article with a dumb, dumb, dumb title. Yes of course it matters. Why did the Red Sox run out an outfield of Lynn in Center, Evans in Right and Rice/Yaz against the big green wall as opposed to Dewey in left, Lynn in Right and Yaz/Rice in Center (yeah I know Yaz played out there but not next to Lynn). Of course it matters. HOW MUCH DOES OUTFIELD ALIGNMENT MATTER> that should be the title.

    Vote -1 Vote +1

  33. rory says:

    Someday, there will be readily available data on ball-off-bat trajectories. Until then, defensive rating will be hard to statistically quantify.

    It was a decent article, but you lost me at “a large part of being able to be a premium outfielder is to run down balls out of your zone.”

    Absolutely. The guy who can make the remarkable play that would’ve been ruled a hit if it would’ve dropped is the primary thing you measure when evaluating an outfielder.

    Even if it’s only, say, 10 plays that one guy would catch and another guy won’t get to, that’s 10 plays to center field, and a lot of those are going to be doubles. Statistically, a team gets about .3 of a run per base. Those 10 doubles = 20 bases which is about 6 runs.

    With Pothagorian W/L, 6 runs can easily be the difference between a win or a loss. That can be the difference between making the playoffs or not, and with millions of dollars on the line, a win or two simply from choosing the correct outfield configuration is a pretty big deal.

    Yeah, if you’ve got too identical outfielders, then there’s not a lot at stake. But, if there’s any discrepancy in skills, it becomes a very important decision.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>