.
Bill James has given SSI loose permission to turn his ideas around into Mariner articles. One thing that Dr. D is not clear about, is whether James would prefer for us to:
I know which one the average author would rather, but James isn't the average author. He isn't jealous about controlling the dissemination of his own thoughts. So which of the above approaches to use? We suspect that, when in doubt, he'd rather we be accurate. I've got an idea about another thing: many readers love to see a Jamesian insight spliced with Mariner braids.
....
Right now he's got a long series going up, "The Fielding Jones." This ain't the Free-Last-15-Posts Hey Bill; these are articles behind the paywall. If you ain't giving him a piddling $3 a month, you oughter be ashamed. :- ) Here's the link to fix that.
But we'll gingerly give a teaser-trailer. In this case especially we need to retain James' crystal-clear prose:
...
The way in which baseball statistics are generally understood is not as numbers at all, but as a peculiar kind of language. Batting and pitching statistics have a wide array of standards associated with excellence. 20 home runs means "power", 30 home runs means "real power", and 40 home runs means "unusual and unique power, possessed by only a few players at a time." 50 home runs means "historic power" or "steroid user".
There are standards like this all over the map. 20 stolen bases is a mark of some speed; 40 stolen bases, real speed, 60 stolen bases, singular and unique speed, possessed by only a few players at a time; 80 stolen bases, historic speed...
It doesn’t have anything to do with 288 of anything; a .288 batting average has no connection at all to 288 items or .288 items. The math has been buried so far beneath our conscious consumption of the data that for all practical purposes it doesn’t exist. What exists is the standard, and the relationship between the standard and the player’s number.
In hitting and pitching there are thousands of "magic numbers" or numbers which are standards of excellence—20 wins, 200 strikeouts, a sub-3.00 ERA, 30 Saves, 200 hits, 40 doubles, 100 runs scored, 100 RBI. They exist almost without limit. Because 100 RBI is a standard of excellence, 90 RBI is a meaningful standard of productivity, and 80, and 70, each number indicative of quality at a slightly different level.
... In fielding, however, there are no magic numbers—literally none. There is not one fielding statistic of any kind which has resonance in the imagination of the typical baseball fan, none which conveys any image, none which conjures up images of Del Ennis or Darrell Evans or Joe Charboneau or Marty Barrett or any other phantom from the magical cabinet of our youth.
.... The closest thing there is to a magic number in fielding statistics is a consecutive-game errorless streak. There is one thing, however, that should be a magic number for fielders: 100 double plays for a middle infielder.
Which is more common: for a hitter to drive in 100 runs, or for a middle infielder to turn 100 double plays? [He shows that the 100 DP's are much less common - jeff]
Of course, there is the problem that playing on a bad team helps a player turn more double plays, since bad teams have more runners on base against them than good teams. The winning percentage of teams who have had a second baseman with 100 double plays is .517; a shortstop, .518. I don’t know what it is for 100-RBI men; I’m sure it is higher than that, since playing on a good team helps a hitter pile up RBI.
I’m just saying. . .did you ever see a press note, "Johnny Fasthands is just six double plays away from turning 100 double plays for the second time in his career." We see those notes every day, late in the season. . .Jack Brewer is just 3 doubles away from hitting 30 doubles for the fifth consecutive year, etc." There are no similar notes about fielding. But there should be this one.
...
When we talk about "language" we talk about a complex system of communication. Body language is consistent enough that we can legitimately call it a language, like Morse Code, smoke signals, etc. Baseball stats aren't a very robust language -- more of a mini-language -- but it's much more of a language than exist in any other sport.
For fans, that is. Do you understand Trent Dilfer's football language when he lapses into it on TV?
I think that's pretty cool, to compare a middle infielder's DP's to a middle-of-the-order hitter's RBI total.
Of course RBI have their limitations, serious limitations; you and I are aware of that. What the sabermetrician is usually not aware of, is that it's a whale of a valuable stat even so. What if Biff Roberts told you that D.J. Peterson was going to rack up 117 RBI next season? That would be meaningless?
DP's depend, somewhat, on circumstances. So do RBI. But like we say. If we told you Nelson Cruz would play 155 games next year but manage only 61 RBI, then ....
Where is Robinson Cano? Here are his DP's, tabled at Fangraphs. Click on over. It's a fun scan; looks just like an RBI column. He used to be a 110-DP guy; then he became an 85-DP guy. Last year he was back up to 93.
...
You might use this idea with profit, when comparing Brad Miller and Chris Taylor. I averaged it out; Miller got the equivalent of 105 games at SS and Taylor the equivalent of 42 games.
| Shortstop | Adj. Games | Actual DP | DP/155 games |
| Brad Miller | 105 | 43 | 63 |
| Chris Taylor | 42 | 19 | 71 |
| Willie Bloomquist | 13 | 4 | 47 |
| Robinson Cano* | 150 | 93 | 96 |
.
I imagine that 2B's naturally turn more double plays than shortstops, because of the 5-4-3's. James gives "a little less than 400 second basemen" and "a little over 300 shortstops" who have hit 100 fielding RBI.
In this case, the above "Middle Infielder RBI" stats do loosely align with UZR, and with my own eye. I'd have called it like this:
Overall, the M's defensive infield is quality, quality in the sense of being "at least average." Kyle Seager has become an excellent defender, and the other two positions aren't giving it all back. Considering that your two franchise bats are playing these glove positions, you couldn't ask for any more on defense. Outfield is the issue in this town, not infield.
A takeway for me having noodled this article, is that Brad Miller might indeed have some work to do, before we certify him as "average-solid" at short. We don't say it's the gospel truth. We're just chatting baseball, kid.
Cheers,
Dr D
.
Once again you turn a gold nugget from James i to jewelry. The counting stats may not have the same predictive power as SABR stats, but boy do they tell a story. That's the first defensive stat that really tells a story. Love it!
...to come and tell how this stat "lacks context", proceed to tweak into some super-de-duper complex proprietary formula and claim how smart they are....
Only to find out that the formula doesn't really tell them anything different....
...I don't think that 100 DPs tells a story that has significant meaning. Not because it lacks context - not because I can make it better with a formula - but because if you look at the list of MIFs who have done it...it doesn't select for excellence very well. Certainly, nowhere near as well as 100 RBI selects for excellent hitting.
...I would choose:
400 OF putouts in CF
10 OF assists (any position)
Errorless game streaks for infielders
40 CS for catchers
Maaaybe DP / Hit for MIFs (we have that data to the beginning of time...that language can be taught the same way that batting average has been taught and selects a LOT better than 100 DPs does)
...I am not criticizing James' larger point. The reason fielding continues to be so misunderstood and incorrectly evaluated by casual fans is that there does not exist the same sort of subconscious language about it that there does for pitching and hitting...and he's right that this language is what drew me to the game in the first place...it was like learning the code words for a secret, exclusive club. No other sport can say that. Basketball has things like triple-doubles and 40 pt games and such, but those just don't carry the same stable meanings that baseball stats do. And forget about Hockey - a sport that used to routinely boast 200 point scorers every year and now rarely sees 100 points from more than a handful of stars. Even more, this is true of football, where it would be nearly impossible to read the bubble gum card for about 2/3 of the roster and know if they were good or bad players (you can do it for QBs, kickers, and RBs, perhaps...maybe top wide receivers...but not for anyone on the lines or anyone playing defense).
Is this any less the case for sabremetricians, as well?
It's always felt to me that the rush to the unified theory of WAR has overlooked the inherent flaws in current defense statistics. It's like taking two glasses of 'clean' water (hitting and baserunning)...mixing in one 'dirty' glass (defense)...and calling it all clean.
Could it be the case that the lack of 'magic numbers' for fielders is because the skill doesn't relate to counting...or because we simply haven't developed the right formulae yet...or both? Will access to full batted ball data solve the problem?
... and I appreciate that you were one of the early super sabes to recognize that the season, and lifetime, RBI / R leader lists are not filled by random players in favorable contexts :- )
That said, do you have the tables at hand - can you expand on your reason for believing that the 100-DP guys were not MI's playing well that year? James didn't release the lists, either, but he had them in front of him, and concluded that 100 DP's is a (largely) skill-based achievement.
:: daps ::
And I'm pleasantly surprised in the change of tone over the last few years. *Most* sabermetricians will now qualify it, like, "to the extent that UZR is accurate here..." This trickles down to the use of WAR itself - the use of WAR is a lot more nuanced than it was three years ago.
.......
Fact is, though, there's a layer of separation between the fans at the stadium, and the fans on the 'net, as to their analytical sharpness... as James notes, there are fans at the stadium who still go by simple batting average, and are annoyed by the attempt to "diminish" AVG. A good part of his article discusses.
Good stuff Diderot.
Here's a decent link to Baseball Almanac's page on SS Double Play records: www.baseball-almanac.com/rb_ssdp.shtml
Interesting stuff on that page: Like Cal Ripken led the AL in SS DP's 8 times...a MLB record. So he had a glove, it seems.
Bobby Wine (Doc, Gordon, the other dinosaurs and I remember him, a blade of a guy with terrific range and an automatic arm.) holds the NL record with 137 DP's in 1970. Rick Burlson holds the AL and MLB record with 147 in 1980. Interesting that, I don't think of Burlson as a glove first guy.
For 2B's, go to www.baseball-almanac.com/recbooks/rb_2bdp.shtml
Bill Mazeroski (of course) holds the MLB season record with 161 in '66. Jerry Priddy holds the AL record with 150 in '50.
Maz (of course) led the NL 8 times. Nellie Fox and Bobby Doerr, both renowned glove guys, led the AL 5 times (Collins and Lajoie, too)
I must admit to paying little attention to UZR. I trust the "eye test" much more. The problem is I don't get to see many guys many times, so it is hard to say who really flashes the glove. UZR gives you and idea, if you look over a run of seasons. Any one season is too dependent on the guy determining what zone the ball was in.