Let's start with a quick explanation of Win Probability Added - you probably are aware of it, even if you don't know what it is, because it's what's plotted in the Fangraphs chart that's part of every recap. It's based on the idea that, at any point in a game, for a given score, inning, outs and men on base situation, you can look at the historical record for the same situation, and figure out what percentage of the time your team will win - the Win Probability (WP). Do that for every plate appearance in a game, and the result charts the ebb and flow of a game, as it goes from 50/50 at the very beginning to 100% for one side or the other after the final out.
You can figure out an impact of a player by seeing the WP before and after each at-bat in which he's involved and crediting or debiting him appropriately for the result. That's what we use when figuring out the heroes and villains for each game, listed just below the WP Graph in the recap. And, you can add up the WP for a player in his games across the season, to find out how he performed overall - how much "win" did he add or subtract over the course of the year - that's WPA, Win Probability added.
Let's take a look at the best (and worst!) Diamondbacks performers in the various categories: it is important to group them, because you can't really compare htters and pitchers, starters and relievers, etc. Couple of other things to note: Baseball-Reference.com and Fangraphs.com have different figures (sigh...). I'm using the ones from Fangraphs, though the rankings are similar, e.g. the top nine hitters are identical on both. Also, remember that each win is worth a total +50% as you go from 50% to 100%; each loss -50%, for the opposite reason. Since we won the same amount as we lost. Arizona's 81-81 record means that their total WPA will be exactly 0%:
Winner: Paul Goldschmidt, +319%
Honorable mentions: Miguel Montero, +186%, Aaron Hill, +143%; Jason Kubel, +101%
Dishonorable mentions: Ryan Roberts, -62%; Josh Bell, -68%; Jake Elmore, -90%
Loser: Chris Johnson, -95%
At the top, Goldzilla dominated, his bat adding more than three wins to out tally. That's pretty impressive: it left him just outside the top 10 in the National League (led by St. Buster at +501%), and no D-back has reached that mark since Tony Clark powered his way to +447% in 2005. It's probably no great surprise to see Montero and Hill fill up the podium, with Kubel also worth more than 100%. But after those four, there's a steep drop-off to Justin Upton at +32%, who barely beat out Wil Nieves (+29%) - it's mostly because of Upton's struggles in those dreaded "late and close" (L&C) situations, where he hit .176, and in 68 at-bats there, managed one home-run and one double.
Failure in those crucial situations often result in a bucketload of negative WPA, and that's what doomed a lot of those in negative territory. Johnson would be the poster boy, hitting 4-for-32 with a .397 OPS, while the trio of Roberts, Bell and Elmore all batted .133 or lower in those L&C plate appearances. The Diamondbacks tally overall was down at -192%, but that figure includes pitchers hitting, which was as ugly as you'd expect. They were responsible for a total of -478%, led by Trevor Cahill's -98%, but we excluded all of them from the list, since it's not their primary task.
Hitter (Single Game)
Only two hitters broke 50% in a single-game, and if it hadn't been for the massive WPA boost provided by his three-run walk-off shot in a 9-8 victory over the A's, Roberts would have easily been the single-season worst culprit, as all his other games combined for -157%! The Rockies must have been fed up of Hill by the end of the year. That +83% game came barely a week after Hill had also +47%'d them. However, it's interesting to note that neither of his cycles were that highly-ranked, both coming in at below 20% of WPA, because the hits largely didn't come in game-critical situations.
The Rockies did get their revenge on Hill the next day, however. He went 0-for-5, hit into a double-play with two men on base in the seventh, and Arizona down by one, then ended the game with a fly-out and the tying run on third in the ninth. Kubel's deficit was almost entirely the result of a single at-bat, good (or bad) enough for -46.8%. It came in the ninth inning of a back-and-forth contest against the Dodgers, where we had blown a five-run lead going into the seventh, and trailed by one with one out in the ninth, and runners on the corners. However, Kubel hit into a game-ending double-play - you don't get much more "late and close" fail than that
Pitcher: Starter (Season)
It's kinda hard for starting pitchers to do much to accumulate WPA, because they are working at the beginning of the game, when not much has a decisive effect. Take, for example, Sep. 15, and the five-run third inflicted on Tyler Skaggs by the Giants. Despite being probably the worst single inning by an Arizona starter this season, he was only charged with -38.4% WPA for the frame. And it was somewhat mitigated by the three scoreless frames he put up around it, worth +11.7, so his overall number was -26.7%. The key thing for any pitcher and WP is putting up a zero. WP does not care about how many base-runners are stranded, just the score.
It's quite interesting to see WPA favors Cahill over Miley, even though his ERA was almost half a run higher. That's because WP also takes into account the game situation - not all runs are created equal. Allowing a run in a tied game is seen (probably rightly) as a much more negative thing than doing so in a blowout. A quick eyeball check seems to back this up. Only four of Cahill's starts ended up being decided by more than five runs. For Miley, the same figure was seven (plus a mop-up relief appearance) - even if you're racking up zeroes, they don't have much impact on the result when you're up by eight or nine.
Pitcher: Reliever (Season)
It's in the late innings that the potential for big WPA - in either direction - is really present. That's clear from the fact that Ziegler had a higher WPA than Ian Kennedy, who threw three times as many innings. Ziegler also benefited from being called on to clean up, with his patented double-play ground-ball generator. Witness the appearance on August 30th: he came in for the seventh with two runners on, threw four pitches, got the double-play and left. +14.4% WPA for facing one batter, thank you very much. Twin killings are very, very good for WP, since they're not only outs, they remove a base-runner who was previously counted against you.
Probably as notable as who was listed, are the two players who aren't. Closer J.J. Putz ended up at just +17%; as we'll see, that's because blown saves are really bad for WP, to the extent that it takes four or five successful saves to counterbalance one. And Designated SnakePit Whipping Boy Mike Zagurski? He ended up third best, at +48%, because of two excellent performances on July 6th and September 4th, each worth +35%. The latter, for instance, he came on in a tied game with runners on the corners and no outs, but a strikeout and ground-ball double-play preserved the tie. Sure, he sucked
occasionally often - but at least had the grace to do so in low-leverage situations.
Pitcher (Single Game)
Winner (Starter): Wade Miley, +48.2% (vs. KCR, May 20)
Winner (Reliever): Craig Breslow, +37.5% (vs. WAS, May 2)
Loser (Starter): Ian Kennedy, -67.3% (vs. PIT, Aug 8)
Loser (Reliever): J.J. Putz, -83.7% (vs. LAD, Sep 2)
For a starter to get good WPA, they need to be pitching in a tight game - a scoreless tie is ideal. That's what Miley did on May 20th: He traded zeroes with his Royals' equivalent for four innings, then we edged ahead in the fifth, and he put down three more zeroes, as we won 2-0. Still, a long way from the best ever: perhaps surprisingly, Brian Anderson from Sep. 1998, who was his own closer in a 1-0 win over the Pirates. Breslow came out on top for his appearance in the eighth inning, protecting a one-run lead (David Hernandez had worked the previous two games) after Ziegler let the first two men reach.
Kennedy's outing wasn't just the worst of the year, it was the worst by a Diamondbacks starter in over four years, since Micah Owings coughed up -67.5% in July 2008. [In case you're wondering, the franchise worst ever belong to Russ Ortiz at -73.0%, in June 2005 - he wasted a 6-1 lead with a seven-run sixth inning]. Putz occupies the bottom four spots on this year's list: all cases where he came in with a one-run lead, blew the save, and took the loss by allowing a pair of earned runs. He only had one save crack the 20% mark on the other side; as noted above, it's a lot easier for closers to lose WP than acquire it.