The Unsung Hero: Defense

Hard to believe, the Astros are almost one third through the 2018 baseball season; the season of the World Series hangover, so called. The hangover that hasn’t happened.

At 48 games last season the Astros were 32-16. They are slightly off that pace this year at 30-18. But their pythagorean record at 36-12 is tops in MLB.

And although the personnel on this team is largely the same as last year, this is a very different team. Last year’s team had historically good hitting with enough pitching to get through seven game victories in the ALCS and World Series. This year’s team is very pitching dominant, with enough hitting to stay just ahead in a very competitive AL West.

Last year the fWAR hitting season total was 32.7. The fWAR pitching was 20.8, a ratio of 1.57 hitting WAR to 1 pitching WAR.

This year the hitting has accounted for 7,6 fWAR. When extended to 162 games this comes to 25.56 fWAR. The pitching fWAR is currently 11. When extended to 162 games this is 37.13 WAR. In 2018 the ratio of hitting WAR to pitching WAR is .69 to 1, almost a reversal of the 2017 ratio.

In short, the 2017 Astros won mostly on hitting. The 2018 Astros rely mainly on pitching.

Or is it just pitching? In all the discussion this year of the record breaking pitching performances, or the somewhat disappointing offensive production, overlooked has been a remarkably improved defense, a defense which, by some basic metrics, is historically excellent.

The following is a chart with key defensive statistics, some basic, some advanced, comparing the Astros fielding performance from 2017 to the current year, approaching one third complete. I will follow that with an interpretation of the data.

One reason why I started writing this article is because as I have been watching the games, it has seemed to my naked eyes that this team is playing very good defense. Of course, the eyes of a fan can deceive, so I wanted to prove or disprove my impression. I believe the data proves that mine were not lying eyes.

Let’s start with the obvious, the stuff that forms impressions. In 48 games the Astros have only 15 errors, second in the league behind the Diamondbacks, who have played 2 fewer games. Yuli Gurriel, rookie first baseman last year and truly bad much of the time, has not made an error this season.

Likewise, Carlos Correa, who had 9 errors last year, is errorless this year. I don’t believe in jinxes, so I don’t care if you curse me for suggesting that our Captain could be chasing the season record for fewest errors (3)for a shortstop, or even the consecutive games without an error record. (110) Early, I know, but I like to think positive. A positive jinx if you will.

And then there are the amazing outfield assists from Josh Reddick, a league leading (tied) 6, as of the time of this writing.

Except for some errant throws from time to time by Alex Bregman, the Astros seem to play a very clean game of baseball this year, and much better than last year.

The first two sets of statistics bear this out. FPCT, Fielding percentage, measures the ratio of successful defensive plays to errors, or (putouts + assists)/ (putouts + assists + errors). Of course, the flaw in this statistic is that a player with limited range may not make an error, thus his percentage remains high, but he didn’t make a play that a better player would have. Nonetheless, the Astros not only reduced their percentage of errors significantly, but increased their league rank to second from 21st, just .001 behind the Diamondbacks, at .991. As recently as May 2^nd they led the league, and if they can get back to that .993 level at the end of the season they will hold the all-time lead in that category.

The second statistic is DER (Defensive Efficiency Ratio). It measures the rate of times batters reach base on balls put in play. Basically, for every ball hit into the field of play, how likely is the defense to convert that into an out? This is better than FPCT because it takes into account the good plays made to create outs, not just how many bad plays that were not made that allowed runners.

In this category the Astros have increased dramatically, not just in absolute terms, (.731) but in rank, where they rest comfortably at #1, up from 24th. In fact, at this rate they will easily break the all-time record since the statistic was kept (1999).

The last two statistics on the chart are measures I devised. (Probably someone else has done this. If so let me know. Or let me know why this method is flawed) ERA on batted balls and WOBA on batted balls tend to reinforce the concept of DER. ERA on batted balls measures how well the defense kept batted balls from resulting in runs. WOBA on batted balls measures how well the defense limited the number of bases occupied by each batted ball. The two should correlate closely.

The leaderboard function on Fangraphs did not allow me to rank the teams in ERA on batted balls. But from WOBA on batted balls the Astros improved from 24^th to 1^st. And of course, ERA on batted balls was nearly 100% better. Both these statistics confirm and validate the so far record breaking Defensive Efficiency Ratio in 2018. They show that Astros defenders are keeping batted balls from resulting in baserunners, and/or keeping those runners from advancing into runs at a very high rate.

Let me anticipate this objection: that the excellent Astros pitchers are inducing soft contact more often, thus making it easier on the Astros fielders. There is very little difference in the soft to medium to hard contact rates between 2017 and 2018. Soft contact 21% +- .5. Medium contact 49.5% +- .25. Hard contact 29.5 +-.25. At all forms of contact the 2018 Astros have significantly lower WOBA and ERA’s allowed than the 2017 Astros, especially in the soft and hard contact ranges. If someone wants to see the exact numbers I will post them in comments.

So far in 2018 we have 1054 batted balls per Fangraphs, May 21, 2018, so the sample is not insignificant, anticipating another objection. Of course, a larger sample is better and the season is not even close to half over. So we’ll see.

There is another way to use pitching statistics to validate the excellence of Astros fielding. By comparing team ERA to team FIP ratings we can show show the possible effect of defense on pitching. FIP (fielding independent pitching) is designed on average to have approximately the same numerical value as ERA. If team ERA is above FIP, it could tend to indicate that fielding did not support pitching. If ERA is below FIP, this could tend to indicate good fielding support. Think of ERA as fielding DEPENDENT pitching. If the fielding is above average, the ERA will be below the Fielding Independent Pitching.

Of course there are many other factors, but for an entire pitching staff in the long run I believe this correlation could have some validity in conjunction with other supporting data.

In 2017 the Astros ERA for starting pitchers was 4.03, slightly higher than the FIP at 3.95. The bullpen had a similar pattern. In 2018 the starting ERA for starters is currently 2.25, the FIP is 2.90. For the bullpen the FIP and ERA are nearly equal, but that is a very small sample with so many long starts this year.

Here one might object that the higher FIP in 2018 just means that we should expect regression from the pitching. Perhaps. But the entire body of data here suggests that the fielding is part of why the pitching is so good. I wouldn’t say that the higher FIP to ERA would prove good defense by itself, but it does lend some validity to the other defensive stats we’re looking at.

FYI: The other two teams in MLB with similar splits as the Astros between ERA and FIP, the Diamondbacks and the Rockies, are the two teams right behind the Astros in Defense Efficiency Rating.

The middle three metrics on the chart are considered advanced metrics. The UZR (Ultimate Zone Rating) and DRS (Defensive Runs Saved) are different ways of quantifying a player’s entire defensive performance by attempting to measure how many runs a defender saved. They take into account errors, range, outfield arm and double-play ability. Both measures take into account the probability of making a given play, a kind of difficulty factor. These are both published by Fangraphs.

Baseball Reference uses a similar metric, Rtot/yr (Total Zone Total Fielding Runs Above Average per 1,200 Innings) which measures the number of runs above or below average the fielder was worth / 1,200 innings.

As you can see by all three measures the Astros improved by varying degrees in 2018, and in each case from a situation where the defense was credited with costing the team runs in 2017, to a situation where the defense is credited with saving runs in 2018.

In Rtot/yr the Astros improved from 19^th to 6^th rank. Unfortunately I could not find UZR and DRS team rankings and if anyone knows a convenient way to find these I would greatly appreciate it.

How or why the Astros as a team is better is beyond the scope of my powers of analysis. This is my first attempt at delving into defensive metrics so any constructive criticism or advice is greatly appreciated. Ya gotta start somewhere.

I will posit a few hypotheses for further analysis.

1. Perhaps the advanced shifting that the Astros so notoriously use is putting players in better position to make plays, or perhaps confusing statcast.

2. From anecdotal sources I have heard that improving fielding was a major priority this year in Spring Training. Perhaps the extra work paid off.

I can identify some individual players whose improvement account for much of the team’s improvement.

Yuli Gurriel’s UZR/150 rating is “only” 4.5, much better than last year’s -9.1. Using Rtot/yr he went from -2 runs last year to +3 this year. Carlos Correa has gone from a -3.8 to a positive 5.1 UZR rating, and according to Rtot/yr he has gone from a +10 rating last year to a +39 rating this year. Josh Reddick’s UZR/150 rating went from +4.7 last year to +27.3 this year. And although UZR ratings are not used for catchers, no doubt Max Stassi’s errorless performance behind the plate has improved upon Evan Gattis’ 9 errors through the course of 2017.

Some other big improvers are: George Springer in center field, whose 3.8 in 2017 is currently sitting at 17.2, Jake Marisnick, whose NEGATIVE rating in 2017 of 12.7 has jumped to a positive 22.4. And surely the biggest surprise of all; Marwin Gonzalez, whose 7.1 UZR/150 in left field last year (yes positive) is 25.3 this year.

One player whose performance has reduced the overall defensive statistics for the Astros has been Alex Bregman, whose -8.7 UZR/150 rating from last year is down to -18.9 this year. This does not pass the eyeball test to many, for we have seen some great plays, but he has 8 errors and a fielding pct of .938, down from .968 last year.

In my opinion, Alex Bregman will not allow himself to continue to make throwing errors on routine throws. This should improve.

One important takeaway from all this is that at least some small part of the amazing success of the pitching staff this year is the support of the fielding. David Barron of the Houston Chronicle wrote a recent article about the Astros fielding, and he started with an anecdote about how important the fielding has been to the pitching, describing an acrobatic play by Carlos Correa. He then quoted Justin Verlander on the importance of that kind of performance to the team. “That’s a perfect example of never taking a play off. I don’t think Carlos had gotten but one ball all night, but the play he made was absolutely huge for me and for the team. It dictated the whole inning by getting the lead off guy.”

Yeah, that was the game when the Astros beat the Yankees 1-0. Was that play the difference between a win and a loss? Could’ve been.

Or as Lance McCullers put it: “ The guys here care as much about what they do on defense as they do on offense, which is why we (the pitchers) have been able to accomplish what we have.”

Note: The Astros were fourth from last in UZR/150 last year at -3.6. As of today, 5/23/18 they are third at +9.5 UZR/150. Thank you Synchsi.

(editor’s note: statistics in this article have been compiled over the course of this week and some may be slightly out of date.)