April 17, 2003
Infield DefenseI've recently written a couple of columns sketching a general measure of outfield fielding by looking at putouts the outfield turns as a percentage of team fly balls, using 2002 season data.
What about unit defense in the infield, though? Can we do the same thing there, except with ground balls instead of flies? And can that lead us to some really strange conclusions?
Yes on all counts, with some problems. While outfield putouts are context-neutral--each time a putout is recorded the batter, and only the batter, is out--infield putouts are context-heavy. A man on first means a successfully turned ground ball to short goes to the second baseman for the first out, and then (if possible) a second putout is recorded by the first baseman if the ball arrives there in time to get the batter.
For my purposes, though, I'm only interested when the infield turns any out, and only that first out. I'm going to try and isolate that by looking at infield unit POs by making a few adjustments:
So whenever I talk about IF POs from here on, remember it's not raw, it's adjusted.
Now, because infielders snag a lot of putouts on line drives, this is going to be a little shocking at first.
Ground Adjusted IF Team Balls IF PO PO/GB GB/tFB Dbacks 1968 1961 100% 1.41 Expos 2143 2151 100% 1.61 Indians 2024 2032 100% 1.44 Padres 2012 2053 102% 1.53 Orioles 2119 2171 102% 1.50 Rockies 2108 2164 103% 1.53 Reds 2108 2177 103% 1.51 Jays 2035 2106 103% 1.37 Tigers 2089 2179 104% 1.30 Pirates 2166 2265 105% 1.68 A's 2128 2233 105% 1.43 Rangers 1982 2091 105% 1.37 Royals 2067 2181 106% 1.42 Astros 1940 2075 107% 1.38 Braves 1977 2132 108% 1.40 Giants 1875 2034 108% 1.19 Cubs 1753 1903 109% 1.31 Phils 2028 2202 109% 1.55 Cards 1935 2104 109% 1.36 Mets 1887 2060 109% 1.27 Yankees 1907 2089 110% 1.26 M's 1809 1982 110% 1.09 Red Sox 1976 2167 110% 1.46 DRays 1853 2041 110% 1.10 Brewers 1866 2059 110% 1.30 Angels 1854 2051 111% 1.22 ChiSox 1892 2105 111% 1.25 Marlins 1861 2083 112% 1.26 Dodgers 1856 2111 114% 1.33 Twins 1737 1994 115% 1.00
The average team saw 1,965 ground balls, and got 2,227 adjusted putouts from their infield unit, for an average rate of 107%. The GB-FB ratio here is actually playable balls as (ground balls/[fly balls-HR]), instead of the published one, and the average adjusted GB/FB ratio is 1.36. Anyone who wants to send me detailed distributions of line-drive and fly-ball outs among fielders can drop me a line and I'll start working on some seriously cool stuff.
That's an interesting table, no matter how you look at it. While the outfield distributions clustered within 12%, there's about a 15% spread in the ability of the best-fielding infields to turn their share of playable balls into outs. And the Twins (led, as we've seen before, by Corey Koskie), would seem to have a league-best infield unit to go with (by PO/FB, anyway) a league-bottom outfield.
The problem with infield defenses is that it becomes much harder to break this down without running into the same issues that have plagued weighted events systems. For instance, I wanted to do a sidebar on pitcher put-outs, because the range of abilities initially seems huge: The Royals, for instance, sit at 7%, while the Diamondbacks were at 2%. But that figure doesn't mean all that much, because a first baseman who flips to the pitcher for the putout could swing the ranking from worst to first with 50 courteous plays over the course of the season.
What I'm more interested in is where teams are getting their outs from, particularly if that doesn't match up with their staff tendencies. For example, is there a case of a severe ground-ball staff that's still getting a large share of its outs from the outfielders? That would seem to indicate a particularly bad infield unit. At the same time, the range of OF PO/FB ran from 65% to 77%, the infield PO/GB ran from 100% to 115%.
Let's go to the graph:
You can see that there aren't any tremendous outliers. In general, the more fly balls teams induce, the more outs they get from the outfield, and vice versa. Yeah, I know, that's not the kind of dramatic insight you wanted from this column, but this is an interesting testament to a dynamic within every team: No defensive unit, infield or outfield, is allowed to get so bad that their out production is dramatically out of whack from what you'd expect, given only the team's ratio of ground balls to playable fly balls.
Even the Twins, who show up at the extreme of both my IF and OF PO ranges, appear here as the single set of dots to the far-left, but their infield production and lack of outfield production still puts them just off the line you'd draw through the other 29 teams.
This points to something strange. Teams, good defensively overall or bad defensively overall, are more or less equally good or bad in both their outfield and infield units. Now, whether this results from a managerial desire to balance their rosters by concentrating on using their bench to shore up the weaker unit with defensive replacements, or if it just shows us that organizations that value defense value it consistently through positions...I don't know. Honestly, when I started this I was sure there would be a couple teams with hugely unbalanced units, and that's all I have for possible explanations on why it's not the case.
In any event, looking at the complexity of trying to further slice infield defense down without having play-by-play data, I'm heading back to outfield unit defense again next time.