Moving in the Fences: A History

Colorado might be the place that most made people aware that baseball works differently in different ballparks. It was pretty hard to deny the fact that, in Colorado, hit baseballs just took off. Since people became aware of Colorado playing in a hitter-friendly stadium, many people have also become aware of San Diego and Seattle playing in pitcher-friendly stadiums. Petco Park and Safeco Field are two of baseball’s newer parks, and to date they’ve played reasonably extreme. Because of their established pitcher-friendliness, both Petco and Safeco are having their dimensions adjusted this offseason. The idea isn’t to make the ballparks hitter-friendly — it’s to make them more hitter-friendly, or basically more neutral. You bring the fences in, and it follows that offense ought to go up.

Yet it’s interesting what we can observe in recent history. I can identify four instances in which fences were moved in somewhere with the idea of helping the hitters. Between 1994-1995, the Royals made adjustments at Kauffman Stadium. Between 2002-2003, the Tigers made adjustments at Comerica Park. Between 2005-2006, the Padres made an adjustment at Petco, which obviously wasn’t enough. And, between 2011-2012, the Mets made adjustments at Citi Field. Though simple park factors are imperfect and while in certain cases we’re working with limited data, the relevant numbers are of interest. We’ll go in order.

Kauffman Stadium

This one’s tricky, because not only were the fences brought in, but they also replaced the astroturf with actual, real-live grass. These adjustments were simultaneous so one can’t be stripped entirely from the other. But anyway, the Royals figured Kauffman was too big, too pitcher-friendly. Over three years, between 1992-1994, the home-run rate at Kauffman was 70% the home-run rate in Royals games on the road. Interestingly, over the same three years, the run-scoring rate at Kauffman was 111% the run-scoring rate in Royals games on the road. Kauffman was actually already playing hitter-friendly. Over the next three years, between 1995-1997, the home-run rate at Kauffman was 95% the home-run rate in Royals games on the road. The run-scoring rate at Kauffman, however, was 99% the run-scoring rate in Royals games on the road. After the fences came in and the turf went out, Kauffman saw more homers, but fewer runs, relatively speaking.

Comerica Park

Comerica was absolutely enormous. Still is, really, but it’s less enormous than it was. A major adjustment was made to the fence in left-center field, as the Tigers worked to make the power alley more reasonable. Over three years, between 2000-2002, the home-run rate at Comerica was 68% the home-run rate in Tigers games on the road. Over the same three years, the run-scoring rate at Comerica was 93% the run-scoring rate in Tigers games on the road. Subsequently, between 2003-2005, the home-run rate at Comerica was 88% the home-run rate in Tigers games on the road. Over the same three years, the run-scoring rate at Comerica was 92% the run-scoring rate in Tigers games on the road. The homers went up, but the runs didn’t follow.

Petco Park

After two years, it felt like Petco was playing too pitcher-friendly, so the Padres made one adjustment, which clearly wasn’t good enough, since now the Padres are making more adjustments. Over two years, between 2004-2005, the home-run rate at Petco was 74% the home-run rate in Padres games on the road. Over the same two years, the run-scoring rate at Petco was 82% the run-scoring rate in Padres games on the road. After the power-alley adjustment, between 2006-2008, the home-run rate at Petco was 81% the home-run rate in Padres games on the road. Over the same three years, the run-scoring rate at Petco was 80% the run-scoring rate in Padres games on the road. Maybe you’re seeing a pattern — the homers went up, yet the runs didn’t.

Citi Field

This one’s also tricky, because the Mets made their big adjustments between 2011-2012, meaning we have but one season of post-adjustment data. But one season is better than zero seasons, and that one season can tell us something interesting. Over three years, between 2009-2011, the home-run rate at Citi was 83% the home-run rate in Mets games on the road. Over the same three years, the run-scoring rate at Citi was 91% the run-scoring rate in Mets games on the road. Then, last year, the home-run rate at Citi was 110% the home-run rate in Mets games on the road. Meanwhile, the run-scoring rate at Citi was 87% the run-scoring rate in Mets games on the road. Once again, homers up, runs not up, based at least on one season.

All of these park adjustments, of course, were different, and they’ve only been grouped together because they involved bringing the fences in. From the four examples, we see four cases in which the home-run rate improved, and four cases in which the run-scoring rate didn’t improve. In fact, in all four cases, the run-scoring rate got lower, relatively speaking, although in Citi’s case we have one season of data and in Kauffman’s case we have the simultaneous removal of astroturf. We don’t have strong conclusions; what we have is interesting evidence.

Evidence that bringing in the fences might have both intuitive and counter-intuitive effects. Park effects are complicated, meaning park adjustments are complicated. It’s easy to understand how nearer fences can mean more home runs — fly balls have to travel less distance. But that can also have other effects, like cutting down on doubles and triples, and there isn’t a perfect correlation between homer rate and run-scoring rate. If you have a pitcher-friendly ballpark, the answer might not be as simple as moving the fences closer to home plate. These things are tough to project.

Of course, what’s happened before after different adjustments might not mean anything for Petco or Safeco in 2013. No ballpark has done exactly what those two ballparks are doing. And the adjustments are being made in large part for psychological reasons — the Padres and Mariners are tired of bumming hitters out — so run-scoring might not be the key. Even if the run-scoring rates don’t change, it could be a psychological success if the home-run rates improve, because hitters want to be rewarded with dingers for their long fly balls. A place where it’s hard to hit homers feels different from a place where it’s average to hit homers, but still hard to score runs. There’s a lot going on here.

Here’s one thing we can say: going forward, it’s probably going to be easier to homer in San Diego and Seattle. Here’s one thing we can’t say: going forward, it’s probably going to be easier to score in San Diego and Seattle. That might well end up being true, but it would be wise to wait for statistical evidence. The historical record is a curious one.

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.

34 Responses to “Moving in the Fences: A History”

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

    I took notice of Citi Field especially. My theory is simply the reduction in OF area for hits to fall in. I know that Citi Field saw a decrease in all hit types, according to ESPN’s Park Factors- aside from HRs of course.

    As for Kauffman, I would suggest that real grass slowed balls down once they made contact with the ground. Perhaps this reduced extra bases…

    Vote -1 Vote +1

    • Anon says:

      Also, deeper walls allow runners to advance extra bases.

      It would be interesting to compare the rate of sac flies, 1st to 3rd, and 2nd to home for parks of different dimensions.

      +5 Vote -1 Vote +1

  2. Anon says:

    Nice research.

    I like having some differences to ballparks but it would be nice if MLB (and the teams) would limit their magnitude. (ex. Houston CF wall should be moved in ~20-30 feet to make it a somewhat reasonable distance and remove the hill/flagpole from play.)

    Are there any cases of fences being moved farther away?

    Vote -1 Vote +1

    • rbt says:

      Sure, and it’s one of the ones mentioned above. The article fails to note that the Royals actually moved the fences back out to their original location after the 2003 season.

      Vote -1 Vote +1

  3. Steve says:

    The other side effect of moving the fences in is less ground for outfielders to cover. Double and triples get replaced by home runs, but at the same time, it would seem that singles and possibly doubles get taken away by the fielders playing a little shallower. The difference between a hit vs an out is much different than a double/triple vs a homer so the ratio of hits taken away vs upgraded hits can be very low and still come out as run scoring neutral.

    Sample size is likely and issue, but ‘what you got is what you got’.

    Vote -1 Vote +1

  4. Ned says:

    Here’s one thing we can’t say: going forward, it’s probably going to be easier to score in San Diego and Seattle.

    A little alcohol never hurt, either.

    +7 Vote -1 Vote +1

  5. J6takish says:

    One thing about all these examples, outside of the surprising Padres teams, these ballparks also were homes to historically awful teams when the changes were implemented

    Vote -1 Vote +1

    • legendaryan says:

      That’s why the comparisons were % of runs that team scored on the road vs at home. It’s the same awful team, just the stadiums they were playing in compared to their own.

      What is missing, however, is the parks they were playing in. For example, the Padres play a lot in Chase and Coors, so perhaps those stadiums allow them to score more away than at home.

      Vote -1 Vote +1

      • Nathaniel Dawson says:

        I’m guessing his emphasis was more a comment on bad teams feeling they have a need to change things.

        Vote -1 Vote +1

  6. DJG says:

    Very interesting article.

    Could it be that outfielders play too deep, in general, in big ballparks? Or is it that the moved in fences prevent enough “extra bases” (e.g., runners not scoring from first on a double, or would-be triples in a bigger ballpark turning into doubles) to more than offset the homers?

    These seem like the two likeliest explanations, if, in fact, it is true that moved fences increase runs. I tend to think it’s the latter.

    Vote -1 Vote +1

  7. jacob says:

    methodology challenge. how much of an impact do 9th inning road home runs in losing efforts have on the tally. it seems to me like every team would hit more home runs on the road just because there are more AB’s on the road.

    Vote -1 Vote +1

  8. Hurtlockertwo says:

    Interesting data. The Padres need more than closer fences.

    Vote -1 Vote +1

  9. Richie says:

    All these ‘could bes’ could be checked with additional basic research. Changes in BABIP, isolated power overall. This article is more basic than interesting. Which is fine, but only for a start.

    Vote -1 Vote +1

  10. Mawazi says:

    @jacob I’m not sure how important that information is to this study. We’re interested in roadHR-homeHR (as a percentage of league-wide numbers) and how it changes when the environment changes, not necessarily the raw numbers. Or am I missing something?

    Vote -1 Vote +1

  11. rusty says:

    These comparisons seem perfect for deployment of Error Bars!

    Vote -1 Vote +1

  12. jimbo says:

    Curious that the run metric declined in all four cases. I’d be interested in seeing raw data for that since it is possible the teams scored more runs home AND away in the following season.

    Maybe the increase in home HR rates boosted offensive morale or something…

    Vote -1 Vote +1

  13. Quong says:

    Every team’s player personnel differs from one year to the next. Isolating the park’s dimensions as a lone factor is off base.

    Vote -1 Vote +1

    • legendaryan says:

      He compared each team to itself. Each team has specific road and home stats. The biggest factor between those to splits is the stadium.

      Vote -1 Vote +1

      • Absolutely says:

        Not if the statistics are isolated for personnel. The 1995 Royals were essentially missing their best hitter from 1994.

        Vote -1 Vote +1

        • hk says:

          Yes, but they were missing him in both their home games and their road games in 1995 whereas they had him in both their home games and their road games in 1994.

          Vote -1 Vote +1

  14. thirteenthirteen says:

    This is interesting, but I have to wonder how this correlates with the overall quality of the offense playing in these parks. I mean, I don’t think the shape of Comerica Park had a lot to do with the 2003 Tigers losing 119 games.

    As a Giants fan, it’s impossible for me not to immediately compare these parks to AT&T, which is also a terrible place to hit home runs. Last year the Giants the fewest home runs of any team in baseball, but they were near the top of the rankings in pretty much every other offensive statistic. It’s certainly true that the home park reduced their scoring (they scored over a hundred more runs on the road than they did at home) but it seems to me that a good offense can overcome a tough ballpark.

    More data plz.

    Vote -1 Vote +1

  15. Baltar says:

    The number of runs a team scores correlates much better to their number of walks than to their number of home runs. Perhaps the teams looking for more scoring at home should decrease the size of the strike zone.

    Vote -1 Vote +1

  16. DrEasy says:

    How was the rate of doubles impacted by the change? That could partially explain the increase in runs scored.

    Vote -1 Vote +1

  17. Carnac says:

    The intuition: Moving in fences decreases hits, but increases home runs. The value of a good defense decreases because the capacity for a good defender to exercise plus range shrinks with the size of a ballpark. This means that more bat-oriented outfielders who don’t necessarily have passable defense can be deployed, since there’s less chance of them to subtract runs through poor defense.

    The reality?: Offense overall hasn’t really been affected. I’ve been told that the decline of home runs wasn’t because of steroids, but a decrease in great home run hitters. Not to mention poor offense-oriented players overall.

    Some teams have shown that it’s possible to play well in pitcher’s parks. It’s not the park necessarily, it’s the talent. All moving in the fences does is redistribute where the offense comes from, and needlessly if the players are poor to begin with.

    Vote -1 Vote +1

  18. MGL says:

    There is WAY too much sample size fluctuation in 1 or even 3 years of park factor data. I am not convinced of anything (other than common sense tells you that moving fences in or out changes the HR rate in a predictable fashion).

    If you want to have a larger sample size to test how moving fences changes run scoring overall, add the following changes:

    As someone mentioned above, in 04, the Royals moved the fences back 10 feet.

    In 06, the Phillies moved the LF fence back 5 feet (and raised it 2.5 feet).

    I agree that moving fences changes the number of singles, doubles, and triples (and fly outs) in the opposite direction of HR totals, but I am also pretty sure the net effect is always to increase scoring when fences are moved in.

    My regressed park run factor data show this:

    Mets went from .95 to .96 last year. (increased run scoring)

    Det went from .94 to 1.01 in 03. (increases run scoring)

    SD went from .87 to .84 in 06. (decreased run scoring)

    KC went from 1.10 to 1.01 in 04 (decreased run scoring)

    PHI went from 1.05 to 1.01 in 06 (decreased run scoring)

    If you combine all of those, you see a strong correlation between moving fences back and decreased run scoring, or moving fences in and increased run scoring.

    My park run factors adjust for the road parks played in, they adjust for changes in the overall league-wide park factor from year to year, they adjust for the pool of players in each game, they are multi-year (for example, the DET park run factor after the fences were moved in in 2003 are based on 12 years of data), and they are regressed.

    +5 Vote -1 Vote +1

    • legendaryan says:

      May we get more insight into ‘your’ park factors? Logically it makes no sense for singles, doubles and triples to increase in park with shorter fences.

      The only explanation I can imagine, is that most parks you analyzed were impacted so heavily by an increase or decrease in HRs, that the number of 2R, 3R and grand slams hit, weighted the parks towards more run friendly.

      Seattle especially suffers from drastically cooler weather and a lower relative humidity (people think the rain means it’s humid) than the other MLB parks.

      I personally, expect Safeco HR rates to rise and run rates to decrease. BABIP was already low for Safeco, and I see no logical reason as to how less outfield can increase hit totals. And the only way I can imagine shorter fences increase run factor, is by drastically increasing HR rates- which I don’t foresee being the case.

      Vote -1 Vote +1

  19. Doug M says:

    If I were a baseball executive, I would think I would want a park with some lopsided characteristics. Then I could tailor my team to my home park. If I had a park which was hard to hit homers in, I would sign hitters with good speed and decent batting skills, but not power hitters. Maybe look for more depth in the batting order. It seems that that is what the Giants last year.

    Vote -1 Vote +1

  20. Denny says:

    Maybe store the balls in Seattle and San Diego in a 0% humidity room, kind of an anti-humidor, to counter-balance the dampness of the stadiums, which also plays a huge role in keeping the ball in the park?

    Vote -1 Vote +1

  21. Ryan says:

    Announcements about moving fences always make me think of the 1990 Cleveland Indians. Alex Cole had a strong two months of leading off and stealing bases, so they moved the fences BACK that offseason to give him room to move. Sure enough Cole dropped off the next year just as Albert Belle was emerging as a power hitter who could have benefitted from the closer fences

    Vote -1 Vote +1

  22. ttnorm says:

    I wonder, assuming the sampling here held true, whether some rethinking of FIP would be in order. Doesn’t the methodology of FIP give a lot of weight to HR rate as an unambiguous predictor of run prevention. If in fact there is a poor correlation between HR rate and run prevention, then what use is FIP?

    Vote -1 Vote +1