Walk and Strikeout Factors, 2010-2012

One of the very most important principles in the field of baseball analysis is the concept of park factors — having an understanding that the game can play differently depending on the environment. I don’t think this is unique to baseball, but it’s most evident in baseball, where a game in old Coors Field would be very different from a game in recent Petco Park. All decent analysis has to take park factors into account. Otherwise, you’re just leaving way too much off the table.

But most people, when they think of park factors, consider what happens to the baseball once it’s put in play. This park increases doubles; this other park reduces home runs. These are the easiest park factors to understand, absolutely, and they’re generally the easiest to explain. Fenway Park has the Green Monster, which does things. Coors Field is at elevation, which does other things. I don’t need to explain this stuff to you.

But far fewer people think about parks having an effect on walks and strikeouts. These effects are less intuitive, but they exist, and they’re worth highlighting because they don’t get enough attention. You’ll find some walk and strikeout factors here, on the FanGraphs Guts page, but you won’t find them here, on the FanGraphs handedness Guts page. So below I’m just going to insert a simple table of data. You’ll find all teams, corresponding to those teams’ home stadiums. Data is for 2010-2012, for all stadiums except Marlins Park, for which we have only 2012 data. You’ll see two data columns: uBB% and SO%. The former refers to unintentional walk rate at home, divided by unintentional walk rate away. The latter refers to strikeout rate at home, divided by strikeout rate away. All home and road numbers are combined, and while this is pretty simple and could be adjusted for the sake of improving accuracy, this’ll get us most of the way there.

In theory, the table is sortable!

 

Team uBB% SO%
Angels 90% 104%
Astros 103% 105%
Athletics 105% 100%
Blue Jays 96% 100%
Braves 100% 105%
Brewers 104% 105%
Cardinals 104% 98%
Cubs 105% 101%
Diamondbacks 93% 100%
Dodgers 101% 101%
Giants 98% 102%
Indians 96% 102%
Mariners 107% 109%
Marlins* 105% 100%
Mets 102% 102%
Nationals 87% 94%
Orioles 100% 97%
Padres 105% 108%
Phillies 102% 102%
Pirates 90% 89%
Rangers 98% 93%
Rays 98% 103%
Red Sox 100% 97%
Reds 101% 103%
Rockies 96% 87%
Royals 99% 93%
Tigers 92% 90%
Twins 105% 97%
White Sox 120% 104%
Yankees 101% 100%

The most extreme effect observed is that U.S. Cellular has increased walk rate by 20%. It’s also increased strikeout rate, but by only 4%. One explanation might be that U.S. Cellular is small, and pitchers are more likely to nibble out of fear of allowing a dinger, but then Safeco Field has had the next-highest walk-rate boost so who knows? There are probably several explanations that go into each number seen.

Nationals Park, meanwhile, has greatly reduced walks, while PNC has significantly reduced both walks and strikeouts. Safeco and Petco Park have allowed for the greatest strikeout-rate boosts, and it’ll be interesting to see if these numbers change going forward, now that both stadiums are getting dimension adjustments to become less pitcher-friendly. It’s not a surprise to see that Coors Field drives strikeouts down, but it might be a surprise to see Comerica Park reducing strikeouts, too. The Marlins used to play in a strikeout-heavy stadium, but early results suggest that Marlins Park won’t play in the same way.

Nothing in here is novel, and many of you have seen stuff like this before. I was first made aware of walk and strikeout park factors years ago, myself. But these things are easy to ignore, when in truth they can be pretty meaningful. Anything can affect everything, and there can be a lot more to a ballpark than you might figure.




Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


48 Responses to “Walk and Strikeout Factors, 2010-2012”

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

    //but it might be a surprise to see Comerica Park reducing strikeouts, too.//

    It is, a bit. What could possibly be behind that? Could it be because Comerica is an offense-friendly environment?

    Vote -1 Vote +1

  2. Steve says:

    Hi Jeff. :-)

    Vote -1 Vote +1

  3. Baltar says:

    I am not a statistician, but other than that 120% these numbers have what seems to me to have a very low variation from the mean. Are you sure they mean anything?

    Vote -1 Vote +1

    • jordany valdespin says:

      I, Jordany Valdespkn, am a statistician and because of the large sample size, the results are statistically significant. Substantively significant? Read the article.

      Vote -1 Vote +1

  4. Bobby Ayala says:

    I’d love to see this expanded into swinging vs looking strike rates, maybe with some screenshots of batters’ eye views of how the ball may or may not blend in with the backdrop. Seems like that could be a big contributing factor to these numbers.

    That would only be, what, 20 hours of work? Let’s get on that!

    Vote -1 Vote +1

  5. Anon says:

    Batter’s eye and foul territory are the two contributing factors that come to mind.

    What would be other causes for BB and K park factors?

    Vote -1 Vote +1

    • I Agree Guy says:

      Maybe amount of daytime vs nighttime games or the number of plate appearances where shadows interfere with the batter.

      Vote -1 Vote +1

    • Giantsville says:

      Could the skill of the home pitching staff, 50% of the data, be a cause of some of the numbers? It would be enlightening to look at home pitching staff vs. park factors to gain an understanding of how much of the data is a direct effect of the park itself. If the White Sox staff was at 135% of league average walks overall, and road staffs at U.S. Cellular are at 105%, than the park’s effects are not 120% but at a little higher than 105%, no?

      Vote -1 Vote +1

  6. Stuck in a Slump says:

    The only reasons that I can come up with for why parks can have such variance in K% and uBB% is a combination of the batter’s eye in CF and hitters feeling the need to swing harder to try to drive the ball out of the park.

    Vote -1 Vote +1

  7. Tom says:

    George W. Bush did WTC.

    He’s always been friendly with the Bin Laden’s — they invested in his oil companies. It’s a fact. Look it up.

    -31 Vote -1 Vote +1

  8. olethros says:

    Shouldn’t this include the SO/BB rates for the opponents as well? You’re effectively throwing out half of the data otherwise.

    Vote -1 Vote +1

    • Evan says:

      I’d guess that this combines data for pitchers and hitters for the team in question, which catches everything.

      Vote -1 Vote +1

      • olethros says:

        That makes sense. The way it’s written, it seems to be referring to that team’s hitters only. Minor quibble. Based on the low variance of most teams, I’m still inclined to think that this data is mostly noise.

        Vote -1 Vote +1

      • joser says:

        Yes, I’m not sure what possessed Jeff to use team names here. It’s a park effect, so the entries should be parks: “Safeco Field” not “Mariners.” If he didn’t want to double-check what Enron-Pets.com-RandomTelCo Field is called this month, he could just use the cities and (NL/AL). Even if the year-to-year variation amounts to noise, we know from previous work there are real differences from park to park.

        Vote -1 Vote +1

  9. Matt says:

    Any chance the high walk players like Dunn who easily paced the majors in BB% (and frequent visiting players like Mauer, Santana, Willingham and Prince who also finished in the top 10 in the AL with Austin Jackson and Choo not too far behind) are the reason why the Cell receives such a high rating?

    Vote -1 Vote +1

  10. The Humber Games says:

    Interesting side note from this data.

    For example:

    Rays 98% 103%
    Angels 90% 104%

    These teams were each 11 games over .500 at home, even though in terms of your BB and SO rates they performed significantly more poorly there.

    Vote -1 Vote +1

  11. Ryan says:

    One thing that can throw these stats off are that pitchers on the teams don’t throw the same amount of innings at home and away. If a player with a low walk rate throws more at home, it will mess up the numbers.

    Vote -1 Vote +1

    • Tactical Bear says:

      This. In order to avoid this problem, we need to break things down further. We can comparting and pitching numbers and compare those totals, but if we group them all together, team composition and quality will have a significant impact.

      Example: team is made up of comparatively low k% pitching staff (especially bullpen) high k% hitting. At home, the pitching staff throws 9 innings, and the hitters play 8 innings some of the time. On the road, the hitters always play 9. This will simply give the hitters more chances to affect the K% on the road than at home, and making it appear as if the home park reduces strikeout rate (in this case, even more signifiantly so because the total number of Ks being compared is lower due to my staff being composed of low k% pitching), when reality this effect exists independently of the park.

      Also, imagine a team with extreme high k% starting pitching and a low k% bullpen. Wouldn’t their home park appear to boost K%, simply as a result of the team’s performance ACROSS THE BOARD being better at home than on the road? The starters will throw fewer innings on the road and more at home, seriously screwing up the data.

      I really want to look at this, so if anybody wants to shoot me the methodology or some raw data or something, I’ll stumble over it and eventually contribute nothing. Shimmeringwang@gmail.com

      Vote -1 Vote +1

  12. Dan says:

    Why would PetCo drive up walks? That is completely counter-intuitive.

    Vote -1 Vote +1

    • Tactical Bear says:

      Best guess: hitters know balls in play at PetCo have less relative value than at other parks, and are more likely to take a patient (ie, high K% high BB%) approach. Swinging makes less sense because the run value of a swing at PetCo is worse than at, say, Fenway.

      Possible counter-intuitive corollary: pitchers understand the absolute run value of a walk is lower at PetCo than other parks, and could be more likely to nibble against dangerous hitters.

      Vote -1 Vote +1

      • Dan says:

        That’s a stretch. First of all, do you really think each player’s game is that fine-tuned, where they’re changing things up accord to park factors? Maybe the pitchers do that to an extent at their home parks, but not the hitters, and especially not in the way that you’re talking about. Players tend to just keep doing what got them to the majors, and that’s probably the smartest approach for the vast majority of players.

        Secondly, this would show up in a number of more obvious ways in the PetCo numbers, and I don’t see any red flags.

        Best guess: Noise.

        Vote -1 Vote +1

  13. PL says:

    “The most extreme effect observed is that U.S. Cellular has increased walk rate by 20%. It’s also increased strikeout rate, but by only 4%.”

    How is this not mostly because of Adam Dunn’s resurgence?

    Vote -1 Vote +1

    • AC of DC says:

      The above is presumably a joke, but for those seriously asking about the impact of a single player or few players, understand that in a single season (such as 2012, where Dunn surged to be a roughly average player and, notably, still struck out twice as often as he walked) at his home park a guy will put in upwards of 300 PA’s (though usually fewer), whereas the data presented cover three years of play (including 2010, when Dunn played for the Nationals) for everybody that showed up, or somewhere north of 17,000 PA’s.

      Point is, no matter how extreme his performance, no lone player is going to own the majority of responsibility for any park’s factors.

      Vote -1 Vote +1

  14. Evan says:

    Great piece Jeff, both concise and enlightening. It would be interesting to see how some of these trends compare with assumed or quantified atmospheric conditions in specific parks at the field level.

    Vote -1 Vote +1

  15. Bryce says:

    Is there any way to disentangle physical effects (batter’s eye, foul territory) on BB% and K% from strategic effects (swinging for the fences, pitching aggressively)? I suspect that most of the observed difference comes from players adjusting their playing style because of the other park factors (strategic effects), but I have no idea how to investigate this.

    Vote -1 Vote +1

  16. Tactical Bear says:

    I think PetCo and Safeco are interesting examples. Could someone direct me to where I could find swing% home/road splits? It seems intuitive to me that the relative value of a swing would be much lower in a park that suppresses power, which should lead to depressed swing rates, and therefore increased K% and BB%.

    Even if pitchers pounded the zone, a significant decrease in swing rate would lead drastically bump strikeout and walk totals, right?

    Vote -1 Vote +1

  17. Nathaniel Dawson says:

    Air density would account for a great deal of the SO park factors, maybe even the majority of it. Foul ground also has a big influence, and in certain case, batter’s eye as well.

    Foul ground would affect walks the same way it does strikeouts, but I’m not sure what other factors play a role.

    Vote -1 Vote +1

  18. MGL says:

    Yes most of the variation you see in those 3-year samples is noise.

    Off the top of my head I would say that you would regress 3-year K and BB factors at least 75% toward some mean (league average if you know nothing in particular about the park that would assure you of a different mean).

    To get an intuitive sense of how much noise there is in these numbers, I suggest that Jeff or someone else put the 08-10 numbers side by side with these numbers. You will likely see lots of numbers all over the place. You will at least see a sample if the regression.

    Vote -1 Vote +1

  19. MGL says:

    07-09 numbers that is…

    Vote -1 Vote +1

  20. Hank says:

    Perhaps a stupid question – which rate is the park factor calculated off of:

    Is it based on K% and BB%’s or K/9 and BB/9?

    I assume it is based on the %’s to isolate out BABIP effects?

    Vote -1 Vote +1

  21. Scota says:

    Looks like noise……

    Vote -1 Vote +1

  22. pft says:

    I suspect umpires have a bigger effect on these rates than park, not to say park is negligible. I imagine different parks have a different distribution of umpiring crews. Perhaps those crews that worked in Chicago collectively had a tighter strike zone? This could contribute to the higher HR rates as well.

    Vote -1 Vote +1

  23. Harry says:

    Would be interesting to survey MLB players on how much of an effect each outfield backdrop plays for each park…

    Vote -1 Vote +1

  24. Tomcat says:

    This makes me wonder why in 20years of operation the Rockies have never tried moving the foul poles towards center? There is literally no downside and it would lower the run environment at least a little.

    Vote -1 Vote +1

  25. Matt Hunter says:

    I’d be interested in seeing the park factors for plate discipline numbers to measure the effect of the batter’s eye in particular. That is, in stadiums where it is more difficult to see the pitch, batters will likely swing at more balls and take more strikes. And, they will make less contact with pitches that they swing at. This would affect Ks and BBs, but we might get a better idea of this effect if we isolate the plate discipline numbers.

    Vote -1 Vote +1

    • Nathaniel Dawson says:

      And how is it that you could conclude that a difference in plate discipline between parks would be solely caused by batter’s eye? That would assume that there would be nothing else affecting it, which,…..how would you know that?

      Vote -1 Vote +1

  26. James G says:

    So if park factors impact BB and K’s and they also impact ERA, then wouldn’t regression point estimates used in FIP and XFIP be biased due to endogeneity?

    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>