Down and Away: The Best Location

Yesterday, David Appelman unveiled our new heat maps, and I love them. As I noted yesterday, the linear weights graph is especially awesome, as it allows you to see where a particular hitter or pitcher is having success, and not just settling for incomplete answers like outcomes on balls in play, which ignores all the variable outcomes that happen on takes, or on swings that don’t put a ball in play. Being able to highlight a player’s overall performance on all pitches is a big step forward for the heat map concept in general.

But beyond even just individual players, there is some really fascinating information available. Based on his discussion with Chris Young, Eno got David Appelman to generate some great charts of league-wide averages for groundballs and home runs. And now, with the release of the heat map tool for all pitches on a league-wide basis, we can begin to say some pretty interesting things about pitching to specific parts of the strike zone.

When given access to the league-wide heat map, my first instinct was to look up the linear weight value of pitch locations. For the 2014 season, that looks like this.

Screen Shot 2014-05-29 at 2.13.26 PM

Nothing too revolutionary here, though. Pitches in the zone are better for the pitcher than they are for the hitter, with pitches in the very middle of the zone being not quite as good for the pitcher as pitches on the edges. But, yeah, anyone who has watched baseball knows this. Pitchers should throw strikes, and when they don’t, it works out poorly for them: News at 11.

So, instead, we have to drill down a bit further to find anything interesting. For starters, instead of looking at all batters versus all pitchers, we want to just look at batter handedness, so that one chart isn’t representing both inside pitches and outside pitches in the same location. Because the heat map tool lets you choose what kind of batter/pitcher handedness match-ups you want, we can go down to specific match-ups and look for location information based on those match-ups.

For now, let’s start with RHP vs RHB, since this is the most common match-up in baseball. And because larger samples are better, let’s expand from just 2014 data to include the last few years as well, so we’re now looking at all pitches from 2012 to 2014 thrown by a right-handed pitcher to a right-handed batter. Here’s what that map looks like.

Screen Shot 2014-05-29 at 2.19.20 PM

This is much more interesting. Here, we see that the pitchers have far more success pitching on the outer third than they do when they try to come inside, with the down-and-in area being even better for the hitter than elevated pitches over the heart of the plate. This is consistent with what we saw when we looked at Mike Trout‘s heat map yesterday; you do not want to throw Trout anything down-and-in if you can help it. While he’s exceptionally dangerous there, this data does suggest that pitching inside is, as a general rule, not as effective as pitching away, though of course one almost certainly needs to do both in order to avoid being predictable.

But this doesn’t help us answer the question of why the outer third, and specifically the lower part of the outer third of the zone, is so strongly positive for the pitchers. The linear weights chart gives us the result but not the process, so instead, let’s look at the charts that do measure the process. First off, here’s the rate at which right-handed hitter’s swing at pitches from right-handed pitchers, and we’ll switch to the 5×5 grid in order to show the differences a little more clearly.

Screen Shot 2014-05-29 at 2.24.26 PM

The down-and-away area of the strike zone is the only one where batters take more often than they swing. Hitters are much more willing to go after pitches on the inner-third than they are on the outer-third, and the down-and-in swing rate is 15 percentage points higher than the down-and-away swing rate. Hitters aren’t stupid, though; they aren’t going to pass up on pitches they think they can drive, so if they are taking a majority of pitches in that part of the zone, it follows that either they think down-and-away pitches are likely to be called balls, or that they just can’t really do anything with pitches in that area to begin with, so they keep the bat on their shoulder and hope for the best.

The charts seemingly suggest that the latter is true. Here’s their contact rate by swing locations:

Screen Shot 2014-05-29 at 2.41.05 PM

Down-and-away is the only in-zone area where RHBs make contact less than 80% of the time when they swing at a pitch from an RHP. There’s no other part of the zone that is even close to having that low of a contact rate, and additionally, the contact rate drops even more dramatically if you miss further down, further away, or further down-and-away. There is no location, in-zone or out-of-zone, that offers lower contact rates than pitches down and away.

And what about when they do hit the ball? Again, no area offers a more meager return on investment than down-and-away pitches. Here is the league’s RHB/RHP Isolated Slugging Per Pitch:

Screen Shot 2014-05-29 at 2.45.25 PM

Hitters drive the ball on pitches middle-in, but on pitches down-and-away, there’s basically no power whatsoever. So on pitches down-and-away, hitters don’t make contact when they swing, and when they do make contact, they don’t drive the ball with any kind of authority. No wonder they don’t swing at those pitches; there’s no real reward for doing so.

Now, we have to keep in mind that we’re only looking at one variable here. A pitches location is certainly important, but we can’t just extrapolate from this that every pitcher should just throw every pitch down-and-away. Realistically, the down-and-away area is probably disproportionately represented by breaking balls, as RHPs want to bury their sliders in that area when facing an RHB. Pitchers can’t live on a steady diet of only down-and-away breaking balls or else hitters will simply stand closer to the plate and sit on a pitch they know is coming. Also, their elows would explode, even more than they are now. Pitchers have to pitch to different parts of the zone, and with different pitches, in order to maintain some element of surprise.

But perhaps the fact that the down-and-away area — and just pitching away in general — is so strongly positive for the pitcher suggests that pitchers are still not going there enough. In a perfect world where the hitters and pitchers keep adjusting to each other’s adjustments, we would expect these adjustments to eventually find an equilibrium. That doesn’t exist right now. Right now, hitters basically can’t hit pitches down-and-away, and they’re much worse on pitches away than pitches in.

At the very least, the data strongly suggests pitchers should not pitch inside nearly as often as television announcers tell us they should. Pitch inside to keep hitters honest, but if you want to get them out and avoid giving up hard contact, you want to pitch outside, not inside.




Print This Post



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


40 Responses to “Down and Away: The Best Location”

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

    Good info to know..but it matters more to understand what each individual hitter is doing. The Cards are very good at game planning for each individual hitter.

    Vote -1 Vote +1

  2. Mike Green says:

    More data is required. It looks from the data like up and in is a poorer location than belt high and outside, as the contact rate is higher and the IsoP is higher. However, pitchers generate more pop-ups and other forms of poor contact on the pitch up and in. You need to look at wOBA, LD rate and so on to capture it.

    It is well-known that the pitch at the knees and on the outside corner from R to R is easily the best location for the pitcher. For the more general statements about pitching outside vs. pitching inside, you need more data.

    Vote -1 Vote +1

    • Dave Cameron says:

      Numbers like wOBA or LD% don’t really tell you anything, actually, because they only measure the outcomes on swings that result in balls in play. Swing-and-misses don’t count against your wOBA or LD%, and neither do called strikes, both of which are strongly negative for a hitter.

      That’s why the linear weight map — “RAA/100P” — is so powerful. It includes not only the linear weight values of batted balls, but also of pitches that don’t result in balls in play. And it very clearly shows that the best results for the pitcher come on pitches down and away.

      Vote -1 Vote +1

      • Eric Feczko says:

        Unfortunately, interpretations need to be constrained when examining a linear weight map by the actual outcomes that occurred.

        For example, according to the linear weight map, down and in is more valuable to a hitter than over the heart of the plate. However, hitters swinging over the heart of the plate hit for more power, generate more contact, and swing more frequently. I suspect the difference in the weight map here relates to the fact that pitches down and in are more likely to be called a ball than pitches over the heart of the plate.

        Vote -1 Vote +1

        • Park Chan ho's Beard says:

          Dave Cameron specifically addresses this point in the article. And the fact is that over the last 2+ years, the results of pitches down and in have been more positive than the results of pitches over the heart of the plate. It may seem counter intuitive, and there are more ways to look at the data further, as Cameron does later on, but that’s the actual data from real baseball that’s been played.

          Vote -1 Vote +1

      • Mike Green says:

        If I read the linear weights chart correctly, waist high and outside is slightly more effective than low and outside. I wonder if that is because they are more likely to be called strikes. It is weird that waist high and outside is so effective but up and away is absolutely the worst place for a pitch righty on righty.

        More nuance with pitches in each of the locations is definitely required.

        Vote -1 Vote +1

        • John says:

          My guess is that as you get to the corners you are more likely to see balls called?

          Vote -1 Vote +1

  3. JCougar says:

    This reminds me of why Joaquin Benoit is so successful. He doesn’t have ridiculous stuff, but nearly every single pitch is on the outer third of the plate, except maybe once or twice an inning, he goes inside to keep guys honest.

    Vote -1 Vote +1

  4. Will says:

    Leo Mazzone is rockin away…. unemployed… but rockin…

    Vote -1 Vote +1

  5. B-Race Miles says:

    Two pitchers that immediately come to mind in regards to pitching low and away are Josh Beckett and Rafael Betancourt. Did it better than any of their peers, it seemed.

    Vote -1 Vote +1

  6. Matt says:

    Really interesting, thanks.

    One thing that really jumps out at me is the LW numbers all the way across the top of the zone. Much worse for the hitter than anything low that’s middle to middle in.

    Kind of goes against the whole “don’t elevate your fastball, it leads to homers; keep the ball down and try to get grounders” thing. That should probably be amended to: “keep the ball down as long as you don’t miss to the inside”.

    So maybe Bauer wasn’t completely off base when he talked about the potential for success pitching up in the zone?

    Vote -1 Vote +1

  7. hefe says:

    “Down and away: the best location”

    This article presents an incredibly facile analysis of location. It fails to account for context (count/base-out situation) or sequencing or pitch type or even the hitter(s) (!). To then take the observed linear weight results–without adjusting whatsoever for any of these factors–and claim that, as a result, “At the very least, the data strongly suggests pitchers should not pitch inside nearly as often as television announcers tell us they should.” is silly, and an embarrassment to much of the quality content that calls Fangraphs home. I don’t object to the graphs or evidence itself–it is a first pass analysis, and certainly, the data is complex–but the conclusions given would require a much more nuanced and deep analysis than DC has given.

    -9 Vote -1 Vote +1

    • cmg8462 says:

      hefe,

      Did you read the entire article?

      It’s a 579,930 pitch RHP vs RHB sample from 2012-2014. Regarding count/base-out/sequencing/pitcher/batter pretty safe to assume that those should all be considered “average” across that sample size. Dave also explicitly mentioned down and away being a disproportionate amount of breaking balls (cites sliders).

      Anyways, seems that you wanted to use some large words and (attempt to) discredit an article that someone took a lot of time and effort to produce for free.

      Cool,
      Chris

      Vote -1 Vote +1

      • Anon21 says:

        It’s a 579,930 pitch RHP vs RHB sample from 2012-2014. Regarding count/base-out/sequencing/pitcher/batter pretty safe to assume that those should all be considered “average” across that sample size.

        No, that doesn’t seem like a safe assumption at all if pitchers tend to prefer certain locations in certain count/base-out, or if certain pitcher types tend to prefer pitching to certain parts of the plate. (We know the latter is true; see guys like Tim Hudson, who are virtually always attempting to attack the bottom of the strike zone.)

        Vote -1 Vote +1

      • hefe says:

        I did. I didn’t like it. I thought the conclusions weren’t warranted.
        1. Just cause you collect a lot of data doesn’t mean everything evens out! Maybe pitchers throw down and away disproportionately when ahead in the count. Maybe they throw it disproportionately to bad hitters. Prove to me that’s not the case before making a sweeping claim.
        2. A token reference to breaking balls doesn’t do much to counteract my concerns.

        Vote -1 Vote +1

        • cmg8462 says:

          I guess when I read this I didn’t immediately say “Hey, what about guys who hit low and outside really well!” OR “Hey, what about if you are in a 3-2 count!” Sure against your average pitcher in a 3 ball count you probably don’t need to risk throwing down and away. Pretty sure Dave was not making the sweeping claim that under all circumstances you should throw down and away. In fact, look a quote directly from his concluding paragraphs:

          “Now, we have to keep in mind that we’re only looking at one variable here. A pitches location is certainly important, but we can’t just extrapolate from this that every pitcher should just throw every pitch down-and-away.”

          Vote -1 Vote +1

        • hefe says:

          “Pretty sure Dave was not making the sweeping claim that under all circumstances you should throw down and away.”

          This is, literally, the title of the article: “Down and away: the best location.”

          Later, DC writes:
          “Right now, hitters basically can’t hit pitches down-and-away, and they’re much worse on pitches away than pitches in.”
          then…
          “At the very least, the data strongly suggests pitchers should not pitch inside nearly as often as television announcers tell us they should.”

          There is no way–no way–you could make that claim, without going through at least some of the complex analysis related to count/sequencing/batter/situation/etc., etc., etc. So it’s kind of lazy bullshit that he did try to make that claim, given that it’s not well-supported. That’s all I’m saying.

          Vote -1 Vote +1

      • Eric Feczko says:

        That certainly explains why the results are so minor then.
        +/- 1 RAA/100P is an extremely small effect. Perhaps effects such as hitter, and pitch sequencing are much more important than average pitch location.

        Vote -1 Vote +1

      • hefe says:

        and another thing, too: just ’cause this stuff was produced for free doesn’t make it immune to criticism. I like Fangraphs a lot, and I think one of the things that makes the site great is the active, participatory community, which critiques and suggests improvements on the authors’ articles (some of which are responded to). I am writing in that vein, hoping that Cameron 1) won’t try this dull trick again and 2) will do the necessary follow-up work to show that this pattern is actually a real thing, and not some artifact of some of the potential factors I mentioned above.

        Vote -1 Vote +1

        • cmg8462 says:

          Welp, touche sir. MGL and you have the same critique, although, you expressed them slightly more aggresively. I do not think this article was meant to “trick” anyone however.

          Vote -1 Vote +1

        • hefe says:

          Actually, nevermind everything I just said. I just realized that I am a flaming bag of feces and not able to understand simple english.

          Sorry for wasting all of your time.

          Vote -1 Vote +1

        • cmg8462 says:

          I understood your stance, I did not agree that your listed factors could make a significant difference on his very general conclusion. However, when MGL had the exact same argument I admitted that you were correct.

          (Turning point in admitting you were correct: I did not give “hefe” the fangraphs commentor much credence UNTIL the guy who co-wrote The Book confirmed his same sentiments)

          Vote -1 Vote +1

        • hefe says:

          no no no, the self-described “flaming pile of feces” who keeps stealing names in order to criticize other commenters is not me.

          I appreciate you coming around, and don’t blame you for not listening until MGL said something similar. That guy is smart, smarter than me. I hope I wasn’t too aggressive, but it seems my comments have gone unanswered and unread, since there are a couple of new articles today doing the same sort of … inadequate… analysis.

          Vote -1 Vote +1

    • Baltar says:

      Hefe, it’s an extremely good and fascinating post on a great baseball blog.
      It is not, however, a doctoral thesis.
      You are expecting too much.

      Vote -1 Vote +1

      • hefe says:

        Like I said, I don’t mind him not writing a doctoral dissertation. I mind him writing conclusions as though he did do the doctoral dissertation, when he didn’t. There’s way too much contextual complexity to arrive at the idea that because of this one graph, pitchers should be throwing down and away more.

        Vote -1 Vote +1

  8. waynetolleson says:

    Yup, corroborates with what my coaches taught me in Little League.

    Vote -1 Vote +1

  9. Elias says:

    These data are really incredible. Just pick a player at random and you learn a ton. I think Dave recognizes this, but is pretty clear that the best place to pitch can vary a lot player-to-player. The game theory stuff makes this complicated, though.

    So many questions! For example, why does Brandon Moss see so many pitches in what appears to be his wheelhouse (down and away) when he seems to have a pretty big hole in his swing up and away?

    But I’m pretty lost when looking at the RAA heatmaps for individual pitchers. I expected to see red on way outside pitches, but they are blue. What am I missing?

    Vote -1 Vote +1

  10. MGL says:

    Game theory equilibrium only applies for the exact same batter/pitcher, outs, runners, score, inning, etc. You cannot tell from data that does not control for everything whether pitchers or batters are using optimal strategies (i.e., whether pitchers should or should not throw more to certain locations).

    In other words, it is exactly true that in any given situation, the expected result for every pitch type in every location should be exactly the same. However, you cannot determine whether that is in fact occurring when looking at data in the aggregate.

    Here is an example:

    Let’s say that I am facing a certain batter leading off a game, and we ignore the umpire, park, weather, the fact that my stuff might be different from game to game, and the fact that the batter’s talent and approach might differ from game to game. In other words, we are controlling for batter, pitcher, game situation, and assuming everything else remains constant from game to game.

    It is true that if we have a large enough sample size, the lwts value of all of a pitcher’s pitches when we split them up into type and location and count, should be exactly the same. If that were not the case, then the pitcher is clearly acting suboptimally, since the pitches and locations that are more effective than others should have been thrown more often.

    However, when we look at just about any aggregate data, whether it be one pitcher, one batter, or everyone league-wide, that will NOT be the case because each location and pitch type is being thrown with different frequencies depending on all those variables I mentioned. Perhaps the down and away pitches are mostly thrown in 0-2 counts and the middle of the plate pitches are mostly thrown in 3-0 counts. Clearly they will NOT have the same value and they shouldn’t have the same value.

    So we cannot infer from this data, as Dave suggests, that pitchers should be throwing more pitches down and away, and more than we can infer that pitchers should be throwing fewer pitches down the middle.

    In addition, we have the problem of intent. For example, pitches thrown down the middle are not necessarily intended to be throw there – the same for any location really – and that really throws a monkey wrench into any kind of analysis where we are trying to determine optimal strategies.

    Vote -1 Vote +1

  11. Swfcdan says:

    When its low let it go, when its high let it fly.

    Down and away has always been the best location, Cliff Lee’s made a career out of it. Hit your spots there and it hardly ever is going out.

    Vote -1 Vote +1

  12. Swingdoc says:

    Another way to think of the effectiveness: To hit a square line drive, low pitches have to be hit slightly out front but outside pitches have to get deep. The inherent conflict here is solved by adjusting the contact point on the ball so that it is “cut” slightly to get the ball on the optimal trajectory and to the opposite side of the field. Less margin for error in that these balls are not really “squared up”. No other pitch location requires such a high degree of precision.

    Vote -1 Vote +1

  13. Scott says:

    “People think I’m smart. You know what makes you smart? Locate that fastball down and away.”–Gregory Alan Maddux

    Vote -1 Vote +1

  14. Syd says:

    I like the fact that the analysis reflects reality. Anyone who has played this game at anything other than a social level will tell you… low and away. Not exactly rocket science.

    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>