All Fly Balls Are Not Created Equal

by Jeff Zimmerman
February 6, 2014

When Juan Lagares took over in center, the fly ball outs started coming in bunches (via slgckgc).

The longer a ball is in the air, the more time a fielder has to catch it. Sounds simple, doesn’t it? Sometimes, in this world of advanced fielding stats, we forget the simple things. While we’re busy measuring things like pitch velocity and deflection, or angle of the ball off the bat, or the exact zone a ball lands in, why not also talk about hang time?

The first research on this subject was published in the 2004 Hardball Times Annual when Robert Dudek manually tracked the fly ball hang times of eight pitchers and tabulated the results. Nowadays, thanks to Inside Edge, the data are available on every batted ball that is in the air for at least 1.5 seconds. (This includes line drives, flies and pop-ups—but we’re going to use the generic term “fly ball.”) There is a lot we can do with this data.

First, here is a comparison of Dudek’s original data with the Inside Edge 2013 data. The next table shows the percentage of the time a fly ball was caught for an out, broken out by how many seconds it was in the air. The data include all fair (not foul) fly balls (balls in the air at least 1.5 seconds) except for home runs.

	2004 Dudek Data		2013 Inside Edge Data
Hang Time	#	Out%	#	Out%
1.5 to 3.0	449	12.7%	15,952	19.5%
3.0 to 4.0	354	48.9%	11,749	58.3%
4.0 to 5.0	368	73.6%	14,719	71.3%
5.0 to 6.0	324	90.7%	16,717	84.7%
6.0 plus	138	97.1%	6,281	93.3%

Despite some differences in the percentages, the pattern remains the same: the longer the air time, the more often the ball is caught.

There’s something else we can examine…how often does a fly ball land for an extra-base hit (in this case, just doubles and triples)? Does this depend on hang time?

Yes, it does. Balls that are in the air between 1.5 and 3 seconds land for an extra-base hit 15 percent of the time (think of line drives down the line or in the gap). If the ball is in the air three to four seconds, the percentage rises to 19 percent. After that, however, the odds of a fly ball landing for an extra-base hit decline the longer it is in the air. The data from this past season:

Frequency of Fly balls Landing for Extra-Base Hits, 2013
Hang Time	#	XBH%
1.5 to 3.0	15,952	15.0%
3.0 to 4.0	11,749	18.6%
4.0 to 5.0	14,719	11.0%
5.0 to 6.0	16,717	2.3%
6.0 plus	6,281	0.8%

The last figure amounts to essentially 50 extra-base hits on balls that stay in the air that long. We’re talking those high wall scrapers, or that random ball that gets lost in the lights/sun, or perhaps even those bloopers that land in that Bermuda Triangle between first-second-right or third-short-left. Those plays happened fewer than 10 times a month last season, so when you see one, know that it’s pretty random.

Of course, where the ball is hit matters, too. For instance, a ball that is in the air for three to four seconds is fielded for an out…

44 percent of the time if the hit is shallow
75 percent of the time is the ball is hit a routine depth
55 percent of the time if the ball is hit slightly deep, and
21 percent of the time if the ball is hit very deep.

These are somewhat arbitrary distinctions, of course, but I didn’t want you to think that I had forgotten how important location is. There are two essential components to every batted ball: how long it is in the air and where it lands. Most people pay a lot of attention to the latter, including Inside Edge — you can see its data in the fabulous new spray charts over at FanGraphs. Today though, we’re going to talk about the former — hang time. Let me slap some data on you:

Fly Ball Count, 2013
Hang Time	Infield	Shallow	Routine	Slightly Deep	Deep
1.5 to 3.0	2,349	20,732	6,926	1,438	128
3.0 to 4.0	676	6,073	7,859	7,264	1,536
4.0 to 5.0	2,104	5,444	6,472	8,580	6,982
5.0 to 6.0	5,660	7,250	6,502	7,254	6,912
6.0 plus	3,036	3,772	2,411	2,162	1,291

Fly Ball Out%, 2013
Hang Time	Infield	Shallow	Routine	Slightly Deep	Deep
1.5 to 3.0	80.5%	8.2%	31.6%	26.0%	3.1%
3.0 to 4.0	95.8%	44.3%	74.8%	54.8%	21.5%
4.0 to 5.0	99.6%	83.1%	95.7%	75.6%	25.4%
5.0 to 6.0	99.7%	96.6%	99.2%	90.6%	37.9%
6.0 plus	99.3%	98.3%	99.2%	94.6%	49.0%

Fly Ball XBH%, 2013
Hang Time	Infield	Shallow	Routine	Slightly Deep	Deep
1.5 to 3.0	2.8%	11.1%	22.4%	51.9%	69.5%
3.0 to 4.0	1.1%	7.1%	12.5%	33.8%	36.6%
4.0 to 5.0	0.0%	3.2%	2.4%	15.7%	21.2%
5.0 to 6.0	0.0%	1.1%	0.4%	3.3%	6.0%
6.0 plus	0.1%	0.6%	0.4%	1.3%	2.1%

To make some categorical observations about these data, I concentrated on two categories of batted balls:

All batted balls (in the air for at least 1.5 seconds)
Batted balls in the air between 2.5 and 4.0 seconds.

If a ball is in the air 2.5 to 4.0 seconds, it is fielded for an out nearly 50 percent of the time. Also, if the ball is not caught, it is more likely than other fly balls to end up as an extra-base hit. Plays in this time range are the most critical for determining which teams make the most important run-preventing plays and which ones don’t.

Hang Time Comparison, 2013
Hang Time	Out%	XBH%
2.5 to 4	49.8%	17.9%
All	62.0%	10.0%

Going forward, I’m going to call balls that hung in the air between 2.5 and 4.0 seconds “sharp fly balls.” It’s just easier that way.

A Case Study

During the first couple of months of 2013, the Mets primarily played Lucas Duda in left field, a series of players in center—including Rick Ankiel for a while—and Marlon Byrd in right. The Mets needed Duda’s bat in the lineup, but he was a liability in left. Ankiel was a decent center fielder but didn’t hit a lick and only Byrd managed to put together a solid year.

In mid-June, Duda was placed on the disabled list, Ankiel was let go, and the Mets picked up Eric Young from the Rockies and handed the center field job to youngster Juan Lagares. Voila, new outfield defense. Young and Lagares did an excellent job covering ground with their speed and instincts and Byrd continued his solid play before being traded to the Pirates in August.

In April and May, the Mets turned just 41 percent and 42 percent of sharp fly balls into outs, a number far below the major league 50 percent average. In June, as their personnel changed over, their out rate climbed to 50 percent. It was 47 percent in July, but then shot up to a remarkable 70 percent in August! Young and Lagares fielded everything batted their way, and I’m pretty sure the Mets’ pitchers took notice.

Their percentage dipped back down to 55 percent in September, which was still the second-highest month for them. The outfield play should stay at a high level in 2014. Lagares and Eric Young will be joined by new additions Curtis Granderson and Chris Young, and even though Granderson’s best defensive days may be behind him, the additions should permanently move Duda to first base or the bench, and that can only be a good thing.

Mets Fly Ball Percentages By Month, 2013
Type	April	May	June	July	August	Sept.
All out%	60%	59%	61%	64%	65%	63%
All XBH%	11%	10%	12%	9%	8%	10%
Sharp fly ball Out%	42%	41%	50%	47%	70%	55%
Sharp fly ball XBH%	15%	15%	23%	16%	11%	15%

Ballparkin’ It

Certain ballparks will affect the fielding values differently. Here is how each team did at home and on the road, and how the road team did at the home team’s park in 2013. The first table is looking at the out percentage and the second table is looking at the extra-base hit percentage for each team. (The major league average out percentage was 62 percent for all fly balls and 49.8 percent for sharp fly balls. The league average of extra-base hits is 10% for all fly balls and 17.9% for sharp fly balls.)

Out%, by Team, Ballpark and Fly Ball Type, 2013
	At home		On the road		Away team
Team	All	Sharp Fly	All	Sharp Fly	All	Sharp Fly
Angels	63.8%	48.6%	60.8%	48.4%	62.1%	48.9%
Astros	60.3%	46.8%	61.6%	49.2%	59.7%	45.9%
Athletics	66.9%	53.3%	65.6%	49.4%	65.2%	46.0%
Blue Jays	61.0%	48.9%	64.3%	49.4%	61.3%	49.1%
Braves	63.0%	52.5%	59.9%	43.7%	62.0%	52.3%
Brewers	59.7%	48.9%	61.8%	46.4%	62.8%	51.2%
Cardinals	62.0%	51.0%	60.5%	53.1%	60.4%	49.0%
Cubs	63.1%	48.0%	60.9%	49.5%	62.9%	52.3%
Diamondbacks	60.3%	48.0%	60.7%	47.0%	61.7%	49.8%
Dodgers	61.5%	50.0%	64.2%	52.6%	62.2%	50.2%
Giants	64.2%	52.7%	63.8%	51.6%	61.8%	47.1%
Indians	61.2%	52.7%	61.8%	49.2%	60.0%	44.5%
Mariners	63.4%	49.1%	66.5%	50.4%	60.3%	45.3%
Marlins	63.3%	53.4%	63.7%	49.7%	62.6%	52.2%
Mets	63.8%	52.5%	64.4%	53.6%	61.0%	47.9%
Nationals	62.1%	49.0%	62.9%	52.4%	63.3%	51.2%
Orioles	61.2%	51.8%	61.0%	45.9%	62.6%	46.2%
Padres	62.4%	50.6%	61.9%	51.1%	61.4%	50.4%
Phillies	60.7%	47.5%	61.7%	50.7%	61.2%	48.2%
Pirates	63.4%	54.4%	63.4%	50.7%	60.3%	48.4%
Rangers	62.2%	51.4%	61.4%	48.0%	64.0%	52.1%
Rays	64.2%	54.6%	64.7%	46.5%	62.3%	49.0%
Red Sox	59.4%	47.7%	57.1%	46.1%	61.9%	47.6%
Reds	62.6%	52.4%	61.8%	49.0%	64.2%	52.6%
Rockies	52.8%	44.9%	54.5%	44.3%	60.7%	48.9%
Royals	63.4%	46.9%	62.5%	49.8%	64.0%	50.5%
Tigers	62.7%	50.1%	61.2%	48.4%	61.6%	48.2%
Twins	61.4%	50.8%	61.8%	54.3%	61.5%	48.1%
White Sox	63.1%	49.5%	62.6%	51.0%	64.1%	52.8%
Yankees	62.0%	51.9%	62.3%	45.4%	61.5%	51.3%

XBH%, by Team, Ballpark and Fly Ball Type, 2013
	At home		On the road		Away team
Team	All	Sharp Fly	All	Sharp Fly	All	Sharp Fly
Angels	9.3%	17.3%	10.7%	18.6%	10.9%	21.1%
Astros	11.7%	19.6%	11.0%	19.4%	12.6%	21.9%
Athletics	8.8%	16.9%	10.5%	20.6%	8.6%	16.2%
Blue Jays	13.3%	23.7%	11.5%	24.5%	11.6%	18.7%
Braves	10.6%	18.0%	11.6%	24.2%	10.9%	18.1%
Brewers	9.6%	17.6%	11.6%	22.8%	10.0%	17.0%
Cardinals	11.1%	18.7%	12.0%	19.7%	13.1%	23.3%
Cubs	11.5%	22.1%	12.2%	19.2%	11.2%	20.6%
Diamondbacks	12.2%	19.9%	13.8%	24.6%	11.9%	20.3%
Dodgers	10.5%	18.5%	9.3%	16.4%	11.3%	20.6%
Giants	11.1%	19.8%	12.0%	19.6%	11.1%	20.2%
Indians	11.8%	19.8%	11.5%	21.3%	12.1%	22.2%
Mariners	10.7%	21.1%	8.7%	17.5%	12.0%	23.0%
Marlins	12.0%	19.7%	10.7%	18.6%	10.8%	18.2%
Mets	9.3%	15.3%	10.5%	17.6%	11.6%	21.0%
Nationals	12.0%	20.5%	10.7%	20.5%	10.5%	17.8%
Orioles	10.1%	16.9%	10.5%	20.3%	11.2%	23.0%
Padres	10.6%	17.6%	11.5%	17.9%	12.6%	23.4%
Phillies	12.3%	21.5%	10.9%	20.0%	12.0%	23.0%
Pirates	10.1%	15.7%	12.2%	21.4%	11.1%	20.4%
Rangers	11.0%	18.2%	10.4%	21.0%	10.5%	19.0%
Rays	10.4%	18.1%	9.6%	18.3%	10.6%	18.5%
Red Sox	12.7%	19.5%	16.4%	24.4%	10.9%	19.6%
Reds	10.5%	21.5%	11.3%	22.3%	11.1%	17.4%
Rockies	13.7%	20.8%	14.2%	21.4%	12.8%	20.7%
Royals	10.1%	19.7%	11.6%	18.7%	9.6%	17.4%
Tigers	11.1%	20.0%	10.6%	17.9%	11.2%	19.3%
Twins	12.0%	19.9%	11.4%	18.7%	10.7%	19.3%
White Sox	8.6%	17.3%	8.2%	15.8%	10.2%	18.8%
Yankees	11.8%	19.1%	9.2%	17.8%	12.0%	21.6%

The Rockies, who play at the expansive Coors Field (the second most spacious field in the game) are easily the worst fielding team. At Coors Field, the home and away teams field 53 percent and 55 percent of fly balls, respectively, which is well under the league average of 62 percent. It’s not just Coors Field, though. Over the past two seasons, the Rockies have been below average on the road. There are some who believe that by the time the Rockies adjusts when they are on the road, the road trip is over, but that probably doesn’t account for the entire discrepancy.

And at home, their fielding rate is actually a few percentage points less than the average away team at Coors. With this information, some defensive metrics, like UZR, have the Rockies’ outfield defense near the bottom of the league. Considering their fly ball fielding percentage, the Rockies may be getting short changed in some defensive metrics, but they’re still not very good.

The second notable item is Boston’s low fielding rate.

Red Sox Fielding Rates, 2013
Location and Team	Sharp Fly ball Out%	Sharp Fly ball XBH%
At home	47.70%	19.50%
Away team at Fenway	46.10%	24.40%
On the road	47.60%	19.60%

This seems to indicate that Red Sox fielders have an ability to limit the amount of hits, especially extra-base hits at Fenway. Boston allows an extra-base percentage of 19.5 at home; the away team stands at 24.4 percent. The difference in extra-base hits hits allowed is almost twice the value of the next largest difference (Pirates at PNC Park). The difference can probably be attributed to fielding balls ricocheted off the Green Monster. That is something that we have heard for years, but now we can see that there is something to it.

Besides just looking at the data at the team level, here are the hang times for some individual pitchers. Each pitcher has what seems to be his own unique mix of batted balls. By grouping pitchers by their groundball and fly ball tendencies, clear baselines of pitcher batted ball types can be determined. Here are the top 10 2013 groundball and fly ball starters:

Top Groundball Pitchers, 2013
				Out%			Fly ball%
Name	BABIP	GB%	FB%	1.5-2.5	2.5-4.0	4.0+	1.5-2.5	2.5-4.0	4.0+
Justin Masterson	0.285	58%	24%	14%	62%	75%	22%	37%	51%
A.J. Burnett	0.305	57%	24%	18%	46%	84%	20%	42%	47%
Rick Porcello	0.315	55%	24%	20%	56%	74%	22%	40%	50%
Doug Fister	0.332	54%	24%	13%	52%	80%	21%	37%	54%
Jeff Locke	0.278	53%	26%	17%	59%	85%	15%	45%	49%
Andrew Cashner	0.269	53%	29%	14%	46%	84%	16%	34%	58%
Wade Miley	0.296	52%	27%	7%	61%	76%	16%	37%	57%
Stephen Strasburg	0.263	52%	31%	15%	54%	83%	13%	31%	61%
Felix Hernandez	0.314	51%	27%	12%	43%	83%	21%	38%	51%
Edwin Jackson	0.322	51%	28%	6%	50%	70%	18%	40%	53%
Average	0.298	54%	26%	14%	53%	79%	18%	38%	53%

Top Fly ball Pitchers, 2013
				Out%			Fly ball%
Name	BABIP	GB%	FB%	1.5-2.5	2.5-4.0	4.0+	1.5-2.5	2.5-4.0	4.0+
A.J. Griffin	0.242	32%	50%	5%	53%	82%	10%	30%	66%
Max Scherzer	0.259	36%	45%	9%	55%	87%	10%	28%	67%
Travis Wood	0.248	33%	45%	24%	57%	88%	13%	33%	61%
Mike Minor	0.272	35%	43%	12%	58%	83%	14%	33%	60%
Dan Haren	0.302	36%	42%	16%	48%	81%	11%	35%	60%
Shelby Miller	0.280	38%	41%	0%	48%	83%	11%	30%	65%
Julio Teheran	0.288	38%	41%	8%	40%	83%	12%	31%	64%
R.A. Dickey	0.265	40%	41%	12%	48%	79%	14%	25%	67%
Jeremy Hellickson	0.307	40%	40%	6%	64%	80%	16%	29%	62%
Miguel Gonzalez	0.260	39%	40%	12%	55%	80%	14%	28%	65%
Average	0.272	37%	43%	10%	52%	83%	12%	30%	64%

Fly ball pitchers will have a lower BABIP because hitters put more air under the ball. The higher the ball goes, the more likely it will be an out. The issue with being a fly ball pitcher is that some of those balls will leave the yard as home runs.

One interesting case is Jeremy Hellickson. Hellickson had been known for dancing around his poor marks in ERA-predictive stats such as xFIP, FIP and SIERA. In fact, he became somewhat of a punching bag in sabermetric circles as a player you should not trust. From 2010 to 2012 he posted a FIP of 4.46 with an ERA of 3.06. In 2013, many analysts claimed victory, when Hellickson’s ERA (5.17) finally rose as expected while his FIP (4.22) stayed about the same.

Hellickson maintained a similar fly ball mix from 2012 to 2013. The main issue was that more batted balls which were in the air for a short time went for hits, especially extra-base hits.

Jeremy Hellickson Fly Ball Rates, 2012-2013
Season	2012	2013
%Total – 1.5 to 2.5	15%	16%
%Total – 2.5 to 4	33%	29%
%Total – 4 or more	59%	62%
Out% – 1.5 to 2.5	14%	6%
Out% – 2.5 to 4	67%	64%
Out% – 4 or more	78%	80%
XBH% – 1.5 to 2.5	11%	13%
XBH% – 2.5 to 4	8%	18%
XBH% – 4 or more	7%	7%

In the shortest two air time categories, he his out percentage declined and his extra-base percentage increased. Now, part of that could be that the Rays defense did not perform as well as it had in past seasons. The Rays struggled to make outs in 2013 (49 percent) compared to 2012 (55 percent) in the sharp fly ball category. It will be interesting to see how Hellickson fares in 2014 with Wil Myers spending more time in the outfield instead of Matt Joyce.

Wrapping Up and Next Steps

A fly ball’s hang time helps determine a fielder’s ability to catch it for an out. In 2004, Robert Dudek looked at a small sample of data to get the results of a fly ball depending on its air time. By using Inside Edge’s data, Dudek’s data is varied using an entire season’s data. The results were very similar. If a batted ball has too little hang time, it will be a hit. If it has too much hang time, it will likely be an out — if it stays in the field of play.

With Inside Edge, much more data can be examined, such as how each team performs at home and away. Different teams and players will be able to field these batted balls for an out at different rates. The key for some teams is to make sure their outfield defense matches the type of pitchers they plan on using.

The next step is to look at the data in a more detailed manner. For example, how many runs do the Red Sox save because of their familiarity with the Green Monster? I feel I created more questions by writing this article than I actually answered. Time to get back to work.

Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR once, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.

29 Comments

Oldest

Newest Most Voted

Inline Feedbacks

View all comments

Brad Johnson

8 years ago

I wonder if we can easily use this to inform xBABIP.

8 years ago

Thanks Jeff for ruining any chance I had of being productive today at work.

Excellent stuff!

Carl

8 years ago

Great article Jeff. Thank you very much.

In September the Mets, having traded Byrd to the Pirates just before the Aug 31 trade deadline, moved Lagares to RF from CF and promoted VanDecker, which likely explains the drop from 70% to 55%.

Dave Studemanmember

8 years ago

Reply to Carl

For half the games, at most. I think the 55% in September is mostly random variation. Byrd wasn’t a big contributor to the Mets’ outfield performance.

David Gassko

8 years ago

Jeff – great stuff!

Max Weinstein

8 years ago

Jeff, great work!

I would love to see some WOWY work done with FB_out% (like you did with Lagares) … Could be a good addition to any fielding metric.

Alan Nathan

8 years ago

Last summer I gave a talk at the Saberseminar (http://saberseminar.com/) entitled “Modern Techniques for Evaluating Hitting”. One of the things that interests me is how well the initial velocity vector (speed and direction) determines the landing point and hang time. The Trackman tracking system is a great tool for this kind of study, since all the quantities are measured. I found that, while the initial velocity vector is an approximate predictor of landing point and hang time, it is not perfect. One of my ongoing research projects is to understand in detail why it is not perfect.

I also investigated (as others, particular Mike Fast and Peter Jensen, have done before me) how BABIP and HR probability depends on the initial velocity vector, using the April 2009 public release of HITf/x data. Although I did not have enough Trackman data to do a meaningful study, it would be very interesting to investigate how these quantities depend on hang time, given that we have a precise measure from Trackman.

The slides for the talk can be viewed and/or downloaded here: http://baseball.physics.illinois.edu/Saberseminar2013-v2.pptx.

Jon Roegele

8 years ago

Nice work Jeff.

Yeah there is so much that can be examined with this data, you must be having fun! One that I would be interested in seeing at some point is looking at how pitch characteristics (i.e. velocity, movement, location, sequence) affect hang time on batted balls.

olethros

8 years ago

I want some sweet giffage of the ~3 infield flies with +6 second hang times that went for extra bases.

Larry Tang

8 years ago

Interesting read.

The Red Sox section stood out to me. I think part of the difference is certainly the level of familiarity the fielders have with the park, but isn’t it also reasonably likely that the Red Sox hitters are more adept at knowing when they can take extra bases? Also did it take into account how good the Red Sox offense is, relative to their average visitor? How significant do you think the actual effect is?

Vil Blekaitis

8 years ago

Reply to Larry Tang

I was thinking the same thing. Over the years, even OFs from the other AL East teams have trouble with the peculiar dimensions and characteristics of Fenway Park.
In addition to the Green Monster, you have that bullpen jutting out into the OF in right center, that chasm between Pesky’s Pole and the bullpen and the cavernous area in CF.
It’s like something devised by a maker of pinball machines.

Regardless, it was a great piece.

channelclemente

8 years ago

Just a great read. Thanks. Maybe you can wax eloquent on ‘backspin’ as a hitting goal one day, it certainly has a bearing on the loiter time of a ball in the air. I suspect, the outs vs hits (HR) tradeoff is where the issue will be decided.

Nathan Aderhold

8 years ago

This is fantastic work, Jeff. Thanks for sharing.

I’m surprised to not see Jered Weaver among your list of Top Fly Ball Pitchers. Was there an innings threshold for the guys you included on the list?

Jeff Zimmermanmember

8 years ago

Thanks everyone for the compliments.

Here are some answers to your questions:

Brad – I will get you in the discussion.

Jon – Truthfully, this article was far from fun while writing. It was in a rough, but finished stage this fall. It supposed to go in the THT annual along with the shift article I wrote. I was never more relieved when Paul and Studes pulled the plug on this one at the time. For the fastball speed, good idea.

olethros – I will have to get the other Jeff involved.

Larry – There is so much to do at the team level. Looking at the Green Monster effects if definitely one.

Peter Jensen

8 years ago

A fly ball’s hang time helps determine a fielder’s ability to catch it for an out.

While the truth of this statement is pretty obvious, that the inverse is true for line drives and some hard hit fly balls is less obvious, but important to note. A hard hit line drive that is caught well be given less hang time than an identical line drive that is not caught and continues to add hang time until it hits the ground.

Alan Nathan

8 years ago

Commenting on Peter Jensen’s comment: The truth of your statement is also obvious. However, it does depend on how “hang time” is defined. I define the same way Trackman defines it, namely as the time it would take the ball to hit the ground regardless of whether it is caught or hits some obstruction prior to that. To me, that definition makes the most sense from an analysis point of view in that it gets around the problem that you raise.

Commenting on cc: For sure, backspin/topspin matters for hang time. I have always thought that a line drive hit at 10-12 degrees with topspin is ideal for getting on base, since the ball will just clear the infield but hit ground prior to the outfielder catching up with it. Obviously, this is not close to idea for hitting home runs. One of the things I talked about in the talk I linked to upthread is whether the ability of a batter to put spin on the ball (whether topspin to get on base or backspin to get HR) is a skill. It is a question I posed in the talk and showed one way one might go about trying to answer the question. However, I did not (nor have not since) answered the question. It is a hard problem (at least, for me) and I am still working on it.

Jfree

8 years ago

I do not think you can possibly judge the Rockies OF – and especially the CF – until you explicitly adjust for the unique hang time there – and the consequences of that. That is the effect of altitude. The ball comes off the bat faster – and as long as it is in the air, it isn’t slowed down either because of the air pressure. That doesn’t affect fly balls as much (except re distance) – but it seriously affects line drives and fliners. OF have to position themselves way back in order to ensure that they can field the LD on the bounce and reduce the extra bases. Which also means that they have to run a long ways (combo of out of position and big outfield) to get the FB’s – which will mess with ‘fielding range’ stuff and all pure ‘outcome’ stuff.

It may well be that the Rockies overadjust for this – and that they also fail to readjust position once on the road. That wouldn’t surprise me in the slightest. Once I saw ‘hang time’ I was really looking forward to seeing this effect re Coors. But instead, the usual superficial conclusion.

Alan Nathan

8 years ago

Jfree: You make some good points, although I don’t agree that the ball necessarily comes off the bat faster at Coors. Moreover, I think that the major effect of altitude on hang time is largely due to the Magnus force, not the air resistance. And whether the hang time increases or decreases with altitude depends on whether the ball is hit with backspin or topspin. A line drive hit with a low-ish launch angle could as well have topspin as backspin, depending on the details of how it is hit. For a ball hit with backspin, the hang time is lower at Coors due to the lower upward Magnus force; for a ball hit with topspin, the hang time is longer at Coors due to the lower downward Magnus force. It should be possible to see this effect.

Alan Nathan

8 years ago

Just to clarify my last comment, my remarks about the major effect being due to the Magnus force applies to line drives. For fly balls, both the Magnus force and the air resistance play a role. Assuming a fly ball is hit with backspin (as it usually is), the effects go in opposite directions and, very roughly speaking, cancel out. That is, there is not much of a difference in hang time between Coors and sea level for a fly ball. That is basically what Jfree said in his comment, which I agree with.

Jfree

8 years ago

Reply to Alan Nathan

re ball off the bat faster – much of that is what the humidor fixed. But, ceteris paribus, a pitch will still travel a bit faster at Coors – and the batters also tend to get better contact because of the other effects on pitches.

re line drives – the real phenomenon I’m thinking of is – where will a line drive be after x seconds at Coors v the same line drive after x seconds everywhere else? I don’t know the answer but 5-10 feet further per second of hang time wouldn’t surprise me. That’s what would seem to affect OF positioning. Position yourself for the line drive and you are out of position for the bloop FB and it is pure tradeoff. Everywhere else, whatever works in one park generally works the same in every other park. At Coors, you have to weigh the cost of letting the LD get past you v the cost of not catching the FB at the edge of your range – and whatever works elsewhere won’t work at Coors and vice versa.

Alan Nathan

8 years ago

Reply to Jfree

Re batted ball speed: I have quite a bit of data showing the average batted ball speed in Denver is the same as everywhere else. While you are correct about pitches being a bit faster at Coors, it is only about 1 mph, which has a pretty small effect on batted ball speed (about 0.2 mph).

Re line drives at Coors: My own take on why there are so many extra-base hits at Coors is that the fences are so far away, as a way of compensating for the better carry. Moving the fences back reduces home runs but increases extra-base hits, since the outfielders have more ground to cover. It would be interesting to see hang time data to compare Denver to other places, for given initial conditions, as a way of testing your hypothesis.

Jfree

8 years ago

Reply to Jfree

Alan – so if speed-off-the-bat is not much different – but HRs/etc travel 10% further; then are they taking 10% longer to get there? I’ve never timed HRs at Coors but, anecdotally, that doesn’t feel right. They do seem to have a more ‘efficient’ trajectory (you sense they’re gone really early – and not just from the sound) and it also seems that they don’t slow down or drop out at the end. But could that offset the same initial ball speed? That latter, translated to non-popup non-HR FB’s, might also affect fielding – less time to make final adjustments even though wind never seems to be a problem (day game glare OTOH)

Scott

8 years ago

The Oakland A’s ballpark leading in highest out% and least xbh% is interesting. Their high fly ball% pitching staff, and high fly ball % hitters have already been examined by Andrew Koo but it seems there may be more to look at on how the A’s are using the Coliseum as an advantage.

Rawson Baggs

8 years ago

” … data are …”

Thank you.

jonathan

8 years ago

It would be cool to see some graphs.

Alan Nathan

8 years ago

Jfree: Here are some interesting facts on HR’s, courtesy of hittracker, for balls hit in a narrow range of initial speed and launch angle. In 2009, the mean HR distance and hang time was as follows:
all except Coors: 400′, 4.88 sec
Coors: 425′, 4.90 sec

So, you can see that the hang time for identical initial conditions is virtually the same, despite the fact that the distance is significantly longer. The similar hang time is due to the cancellation effect I referred to earlier. The very different distances is due to the reduce air drag.

Jfree

8 years ago

Reply to Alan Nathan

That s interesting. Help me out with some physics here re non-HR. I’m still trying to get the effect on fielding. Initial ball speed is the same. Same baseball. Gravity is the same. Assume same launch angle. Lower air pressure/resistance would tend to produce a lower trajectory (80% sea level, 20% outer space/vacuum) – less vertical up/down/waste, more horizontal – overall a trajectory that is 5-8% ‘different’?? – as seen from a camera off to the side (or a fan in the stands). And at terminus, the ball speed would be higher (lower drag) and would come in at a lower angle. Is that true? I’ll assume it is for purposes of rambling.

From the fielder’s perspective then. They need to initially position themselves further back – maybe 10-15 feet? – leaving a larger gap between them and the infield (affecting defensive ‘range’ and ‘arm’ and ‘OOZ’ stats?). Visually/mentally, that lower trajectory would appear to result in the ball going higher for longer (angle of elevation keeps increasing for longer). ‘Stopping’ for longer (the sweet spot for fielding where the angle of elevation stabilizes and you know you’re in position to catch it). And then coming in faster and lower to the glove.

Risk-wise, the higher terminal ball speed and lower angle would seem to raise the cost of a type-A error (getting in position, failing to catch, letting the ball get behind you) relative to other parks because you would actually have to run farther to chase the ball down (big penalty for xbh hits, runners scoring, etc). A type-B error (failing to get into position to ensure the ball stays in front of you) might become an acceptable risk – you don’t get the batter out but you keep runners from advancing as far – but that will affect how the fielder gets scored for errors and range. This seems like a prevent defense in football. But with a very different effect on ‘Rockies players’ v ‘Coors visitors’. If a visitor doesn’t adjust, they may lose a game. If a Rockie doesn’t adjust (to Coors at least), he loses his job and possibly his career since Coors will have seriously messed with his defensive stats already.

Alan Nathan

8 years ago

Reply to Jfree

You have asked some interesting questions. I don’t have any ready answers for you. However, let me suggest that you try to investigate these issues yourself by playing around with my “trajectory calculator”: http://baseball.physics.illinois.edu/trajectory-calculator.html. You can calculate trajectories under different conditions (e.g., different elevations, launch angles, initial speed, spin, etc.) to see how these things affect where the ball ends up and how long it takes to get there.

However, I return to the point I raised earlier. One way to compensate for the better carry at Coors is to move the fences back, which is what was done. This reduces the number of home runs but means the outfielders have more ground to cover. This surely must play a role in positioning and in the ability to catch up to a fly ball. Any analysis of fielding ought to take this into account, in addition to the points you raised.

Juan Pablo Zubillaga

8 years ago

Jeff, it seems to me that in the two tables showing percentages by team, the “On the Road” and “Away Team” columns are switched. You mentioned fielders at Coors having 53 and 55 percent of outs for home and away teams, respectively. But in the table, the 55% is shown as “On the road”, while the “Away Team” percentage was 60. It ceartainly makes more sense what you wrote than what the table shows, so I assume they were mistakenly switched.