December 20, 2012
In A Pickle
The Four-Tool Teams
You're familiar, I'm quite sure, with the sacrosanct quintet of "tools" that the very best baseball players are said to possess, the running and the hitting and the power and the throwing and the fielding that only a very few athletes pull together in one package on the field of play. You're also familiar with the idea that tools are (a) not all there is the universe and (b) not equally valuable. Examine, for instance, 2012's top five position players by WARP:
Trout and Cano are not known for their arms, Posey, athletic though he may be for a catcher, doesn't have wheels, Braun is not considered a stellar glove-man (though he is certainly better in left field than his brutal play at third base might have indicated he was capable of), and Cabrera ... well, that's a two-tool talent if ever there was one. And yet that two-tool talent won the MVP award! (Shh now, there there, it's all right, my statnerd friends. This too shall pass.) The reason, of course, is because at the top of the charts, most of a player's value comes from his bat being significantly better than the bats of others. Trout's WARP, for instance, breaks down this way:
Trout was third in all of baseball in baserunning runs, to give you a sense of scale. This illustrates the fact that there aren't enough opportunities to make a real dent on the bases and further the dents that excellent runners do make are small and easily fixed with Bondo.
So you see, if you didn't know already, why top hitters get the big bucks and top fielders or baserunners get shunted to the bench until the Angels manage to trade Kendrys Morales. (That's a Peter Bourjos reference.) And yet we still obsess as a baseball culture over five-tool players. Willie Mays and Bryce Harper get us hot and bothered (when the latter isn't getting us hot under the collar) while Miguel Cabrera, respect his bat though we might, engenders jokes about his midsection. It's not a failing on our part, exactly, because at the same time that we're searching for Truth and the ultimate measurement of Value as baseball scientists, amateur or pro, we seek Beauty and Joy as baseball fans, and batting alone rarely elicits elatement the way that Andrew McCutchen running a ball down in the gap, bashing a double, and stealing third all in the space of 15 minutes does.
That relatively familiar ground having been trod, the question that interests me is where we might find the Willie Mays of baseball teams. Which squads have excelled in every phase of the game? As it happens, we keep, right here on this site, a variety of team-level statistics that can help us address this very question. Four key areas come to mind:
We have all four of these stats back to 1950, so while we don't have all of baseball history available for our exploration, we've essentially got the second half of it. Here's what I—well, here's what the indomitable Dan Turkenkopf did for me after I asked as nicely as I could:
The result of this procedure is that the very best possible team in history would have a total rank of four (a "1" in each category), the smallest total rank. The very worst would have a ... well, this is where a tiny bit of arithmetic comes in. In 1950, there were 16 teams. In 2012, there were 30. In between, the number of teams was in between. A horrible team in 1950 then would have a total rank of 64, while in 2012 it would be 120. (And the team in 2012 with a rank of 64 could be average across the board.) The simple solution, which is probably not mathematically precise but which is close enough for the toy I'm creating here, is to divide each team's total rank by the square root of the number of teams in the league. At the high end, this gives more credit for finishing first in the league to teams that did so in bigger leagues (1/sqrt(30) is less than 1/sqrt(16) and thus better). At the low end, vice versa -- our theoretical awful 1950 team would have 16, 16, 16, 16, which when divided by the square root of 16 is less than a 2012 team with 30, 30, 30, 30 divided by the square root of 30, reflecting that it takes more negative effort to finish last in a 30-team league, but not drastically less the way the unadjusted total ranks showed.
(What happens in the middle of the rankings isn't great by this method, by the way: an average team in a 16-team league would have 34 total rank divided by sqrt(16) (adjusted rank: 8.5) and an average team in a 30-team league would have 62 total rank divided by sqrt(30) (adjusted rank: 11.3). I don't see a compelling reason to have the median 1950 team come out better in the overall rankings than the median 2012 team, yet that's what we get. My response is: we'll deal with median teams some other time and by a different method. For now, I'm interested in extremes, and in particular the upper bound, which is treated more or less fairly by this process.)
You'll note some intentional parallels between this ranking methodology and a tools-aggregation way of examining a baseball player, mainly in that it wipes out a very important piece of information: how valuable a given tool/stat is compared to other tools/stats. All this adding and dividing of ranks puts baserunning on the same footing as batting, in other words, which is something we'd never do if the goal were to create a ranking of the best teams. Fortunately, that's not the goal—I'm after the most well-rounded team. I don't particularly care that bad teams can finish first in baserunning (the 2006 Royals are one example) or that it's hard to be bad while leading the league in TAv (the 1997 Indians had no pitching, baserunning, or defense, but went to Game Seven of the World Series behind the number-one offense in the game, granted that they only had to win 86 games to take their division). I just care who does everything well.
I also, to address another issue with this method, don't care how much better a team is at a particular stat than other teams. I could care, certainly. It wouldn't undermine the method to use z-scores or percentiles or what have you, but what it would do is add quite a bit of complication for what I suspect would be very little gain, particularly in terms of comprehension of the results.
It's time to pay all this methodology talk off. First, the best all-around team in baseball in 2012 is the team that inspired this examination in the first place: the [editor's note: Here Jason made a joke about the Angels' team name. It has been removed, as an act of kindness to you. Don't forget to renew your subscriptions!] Angels. Our own Sam Miller has noted that the Angels were an excellent offensive, defensive, and baserunning team, but were bad at pitching, and indeed they were—finishing first in TAv, first in BRR, and second in PADE helped drive them to 89 wins, but a 24th-place mark in FRA kept them at that level, in third place in their division (of four teams) and out of the postseason. As it turns out, Ervin Santana and Jerome Williams throwing more than 300 innings was not a recipe for maximizing the Trout-Pujols-Hunter-Kendrick core. Each of those four added positive value with the bat, glove, and legs, with the exception of Pujols on the bases (as you might imagine). The Angels had a chance to be special, but their hopes were undermined by one poor aspect of the team. Much as the Angels hope they've accomplished in fixing the back end of their rotation this offseason (Joe Blanton, Jason Vargas, and Tommy Hanson can't be as bad as Santana was, can they?), Anaheim fans should note that the second-ranked team by this method in 2012 was the division rival A's, who finished fifth in PADE, sixth in FRA, ninth in TAv, and 14th in BRR. The 2012 roster remains largely intact for 2013, though the talent level of the 2013 version of the team is very unlikely to be the same as 2012's performance. Brandon Moss's .596 slugging percentage is ... well, internet baseball nerds have stopped talking in absolutes in the way that we used to, but that slugging is just not happening again.
What's more notable about the A's place in the 2012 Most All-Around standings, though, is that Billy Beane's reputation for building teams by focusing on undervalued skills and players has created a perception that his teams sacrifice some skills in favor of others. When power hitting is available, he acquires that skill at the expense of contact hitting. When pitching is cheap, he buys pitching and eschews hitting entirely. When defense in the outfield isn't appreciated, he finds those players even when (theme!) they can't add value with the bat. (Oakland's home park's run-suppression may have aided these impressions.) The 2012 team, though, was above-average across the board, putting up the kind of performance you'd expect from a franchise that doesn't have to compromise due to payroll/ownership/venue problems.
These 2012 teams are pikers, though, in the scheme of things. Neither ranks in the top 100 of all the teams in the dataset. We can stay in the same general geographical vicinity, though, to examine the most viciously awe-inspiring all-around division of baseball ever created (well, since 1950): the 2001 American League West, featuring the Seattle Mariners and the Oakland Athletics, tied for the second-best all-around seasons that we know of. In tabular form:
Wowsers. Seattle saw a little more variation by dipping as low as sixth, but two firsts and a second in the other two categories, with the A's just one or two spots behind in each? Tremendous. Remarkable. Exhilarating. Each team had a Pythagorean record that called for more than 100 wins, and the Mariners of course put up a record 116 victories. Then each team was knocked out of the playoffs by the Yankees, who finished with an 89-71 Pythagorean record and lost the World Series on just about the stupidest bloop single there could be to a Diamondbacks team that started Tony Womack at shortstop and whose no. 3 starter was Brian Anderson. The baseball gods are assholes is what I'm saying, and I'm feeling less exhilarated.
Five teams in history have put up "raw total ranks" (by which I mean adding up the ranks in the four categories but stopping there—no dividing, no square roots) that best the 2001 A's and Mariners, but four of those (1954 Indians, 1953 Yankees, 1951 Yankees, 1959 Giants) did it in a 16-team league, not the 30-team context faced by the Oakland and Seattle powerhouses. Before moving on to that fifth team, though, two of those squads in the parentheses are notable. First, the 1954 Indians are just one of two teams to finish first in three of our four categories, as they took top marks in PADE, FRA, and BRR while rolling to a respectable fifth-place finish in TAv. Their pitching was led by three Hall of Famers, Bob Feller, Bob Lemon, and Early Wynn; their defense was anchored by the up-the-middle pair of Bobby Avila and George Strickland, each of whom rate as above-average by FRAA (with Avila in double-digits); and left fielder Al Smith, a year away from MVP candidacy, added a win of value on the bases despite getting absolutely wiped out on steal attempts. The Indians won 111 games in a 154 game season (which comes to 116.8 wins over 162 games). And then got swept right out of the World Series by the Giants, who featured Leo Durocher in the dugout, Johnny Antonelli on the mound, and a declining Monte Irvin roaming left field. Oh, and Willie Mays. They also had 23-year-old, just-back-from-Korea, MVP-winning Willie Mays. So that's rough. (And fits the Seattle-Oakland pattern, by the way. Why can't the great teams win it all? Moral failings, surely.)
On the other end of the spectrum are the '59 Giants, who by my measurement are the seventh-best all-around team ever, finishing first in PADE, second in FRA and TAv, and fourth in BRR, but who by anybody else's measure are an entirely forgettable team, having won just 83 games with a Pythagorean record of 87-71. What's most appropriate, though, is that Willie Mays was on that team. As arguably the greatest all-around player in history and yet one with just one championship on his resume, it seems fitting that a just-post-peak Mays would be at the center of the team with perhaps the biggest disparity between "coulda" and "shoulda" in the last 60 years.
Have I teased you enough? I've mentioned the two teams tied for second in the best-all-around rankings, four of the five teams with better raw ranks than those two teams, and one of two teams that led the league in three of the four categories. So what's the no. 1 overall team, the fifth team better than the 2001 A's and Mariners, the other team to lead the league in at least the three categories? You're going to slap your forehead when I tell you: "Of course! The famous powerhouse! With all those Hall of Famers! Loved that World Series!" It was the 1974 Dodgers. Third in FRA. First in TAv, PADE, and BRR. Domination, simply and purely.
You saw this coming, in a certain sense. I oversold the joke with all that "famous powerhouse" stuff and thus may have circled back around and undersold the team, which won 102 games (106 (!) Pythagorean). Unfortunately, they lost 4-1 to the A's in the World Series and fell off to just 88 wins the next year. That famous infield (Garvey-Lopes-Russell-Cey) did eventually win a ring, but it wasn't until 1981, when they were all at least 32 and baseball played a weird schedule because of a strike. More importantly, in a sense, the '74 Dodgers wound up with just one Hall of Famer, and a relatively dubious one at that, one perhaps known more for his longevity (23 years in the bigs, 5282 1/3 innings pitched) than his dominance (led his league in ERA the same number of times he led his league in ER: once)—Don Sutton. Of course, the team was full of Hall of Nearly Great types (including plenty of players who did actually appear in the book, including Steve Garvey, who could hit and field, Davey Lopes, a giant on the basepaths, Bill Russell, as stalwart a defender at shortstop as his Celtics namesake was at center, Ron Cey, who just once in his career (minimum PA: 3) had a below-average season at the plate, Jimmy Wynn, who with a little more longevity and a little less Astrodome might have had a Hall of Fame argument, Andy Messersmith, unjustly more famous for his role in labor history than his career ERA+ of 121, and Tommy John, who probably looks at Sutton's Hall of Fame plaque and scratches his head. And don't forget Mike Marshall, who threw 208 1/3 innings that year without making a single start. (If the Reds want to get more out of Aroldis Chapman ...)
This was, in short, a team. My favorite aspect of the '74 Dodgers being the best all-around team is how neatly the roster construction fits the across-the-board outcome. Jimmy Wynn did post an 8.6 WARP, but the team's excellence really came from having nine different position players post three WARP or more. Nine. In a league where the pitcher bats! (Don't tell the MVP voters that, though—they gave the award to Garvey, who was by our metric the fifth-best player on his own team, passing over Wynn, Mike Schmidt, Johnny Bench, and Joe Morgan, among others. Hell, Mike Marshall finished ahead of those four guys, too. A relief pitcher!)
So next time someone wants to get all "Oh '27 Yankees" or "'75 Reds" or whatever on you, you come right back with the '74 Dodgers. "They led the league in Park Adjusted Defensive Efficiency!" you say, though you'll want to say it quickly to prevent the person from walking away in the middle of your sentence. If you can get it all out, they'll be impressed. Let's change the world together.