Thursday, September 27, 2012

Knuckle Sandwich

R.A. Dickey should win the National League Cy Young award. That's what I thought about a week ago. Nothing since then should have changed my mind. All Dickey has done in the last two weeks was throw 15.2 innings of 5-run ball against the Miami Marlins and Pittsburgh Pirates, giving up 14 hits, 4 walks, 1 home run, and striking out 17. Essentially, he did what he's done all season. Here lies the problem: In the words of Jed Bartlett (A character on The West Wing), "Show me numbers." I can look at the Mets knucklballer and say "Wow, he has been dominant this entire season, he should win the Cy Young." I needed to prove it, and not just with any old evidence, I needed to prove it with numbers.

So, here are some numbers for you. Dickey is 20-6 with an NL leading 2.69 earned run average, 222 strikeouts, and 1.041 WHIP. These are standard statistics that baseball folk have used for years in order to evaluate pitchers. Due to the recent statistical revolution in baseball, I've learned to discount some of these statistics, substituting more accurate and descriptive ones in their place. For example, wins and losses mean very little when evaluating a pitcher alone because a lot of what determines whether a pitcher gets a win or a loss has little to do with the pitcher. Earned run average, can still tell us something about a pitcher in a grand sense, but such a statistics combines the play of a pitcher and his defense, instead of the isolating solely the contributions of the pitcher.

Alright, all of this is interesting, but sabermetricians have found a way to show how many wins a given player has contributed above a replacement level (similar to league average) player. These statistics combine most to all of a player's contributions and uses some math in order to normalize other variables like league, environment, etc... The problem is that different websites calculate WAR in different ways. The two main types are rWAR and fWAR. According to rWAR, calculated and used by Baseball-Reference, R.A. Dickey is the 3rd highest WAR amongst pitcher in the National League, trailing only the Cincinnati Reds Johnny Cueto and Los Angeles Dodgers Clayton Kershaw. On the other hand, fWAR, created by Fangraphs, has Dickey as the pitcher with the 6th highest WAR in the NL. Why the discrepancy?

I've read a number of articles about WAR, and they have told me that the major difference in the calculations of WAR is the specific defensive metric used. fWAR uses UZR, rWAR uses DRS, and WARP uses FRAA. Well that is for position players, not pitchers. So, I delved into how WAR is calculated for pitchers. Although rWAR and fWAR have some differences they are relatively minor, including the manner in which park factors are adjusted for, and whether the metric is adjusted for interleague play. Overall, the main difference is simple. The statistic each used to begin the calculation is different. Fangraphs uses FIP and Baseball-Reference uses Runs Allowed.

Let's begin with the easier of the two. Runs allowed is as simple as it sounds, it calculates the number of runs allowed by a pitcher, earned and unearned. FIP, fielding independent pitching, attempts to calculate everything a pitcher does during a game that he can control. They include, strikeouts, walks, hit-by-pitches, and home runs. Fangraphs assumes that any ball put in play that isn't a home run is not the pitchers fault. While this is not always true, as we should assume that every MLB pitcher is subject to at least reasonable defense, when attempting to calculate how much a pitcher contributes to his team in order to compare him to other pitchers, using only those statistics that are solely attributed to a pitcher is reasonable. FIP is scaled to look and act just like ERA. So, in the end, rWAR includes a pitcher's team to some extent, while fWAR looks only at what a pitcher contribues while possibly leaving out some minor contributions.

So, why then does rWAR rank R.A. Dickey higher than fWAR, to the point at which we might not consider Dickey for the Cy Young award through Fangraph's goggles, but might do so when considering the confines of Baseball-Reference's rWAR? If rWAR might consider a pitcher's team too much, than it would make sense for Baseball-Reference to consider the Mets as an above average defense, right? Interestingly, the Mets have a -33 DRS, ranking them 6th from the bottom in that category.  So, what explains the disparity between Dickey's fWAR and rWAR?

We said that fWAR uses FIP, which focuses on a pitchers strikeouts, walks, hit-by-pitches, and home runs.
Name           IP            K/9              BB/9             HR/9
Gio Gonzalez 193.1 9.36 3.4 0.42
Clayton Kershaw 211.2 8.97 2.47 0.68
Cliff Lee 198 8.86 1.27 1.05
Johnny Cueto 210 7.03 2.06 0.6
Wade Miley 187 6.45 1.78 0.63
R.A. Dickey 220 8.55 2.13 0.86

         HR/FB               FIP             WAR
Gio Gonzalez 6.10% 2.84 5.2
Clayton Kershaw 8.70% 2.99 4.8
Cliff Lee 11.40% 3.07 4.8
Johnny Cueto 7.70% 3.27 4.7
Wade Miley 6.70% 3.2 4.5
R.A. Dickey 10.60% 3.28 4.5

When a ball is put in play in the air vs. Dickey, it is a home run 10.6% of the time. This ranks second to worst amongst these top pitchers. This makes sense due to Dickey being a knuckleball pitcher. When a pitch that averages between 75mph and 85mph floats into the strike zone with little movement, as a bad knuckle ball does, MLB hitters will generally crush the ball. In addition, despite Dickey's high number of strikeouts, his BB/9 are not very good either. Gio Gonzalez overcomes a higher walk rate by allowing almost zero home runs and striking out hitters more often than Dickey, while Clayton Kershaw's more innings combined with his low HR/9 gives him the edge over R.A.

Here is something to consider. As a former pitcher, I can vouch for the fact that it is easier to pitch with a lead. Certainly, all pitchers should be able to perform just as well with or without a lead. On the other hand, the more a pitcher pitches with the lead, oftentimes the more chances he can take in order to get hitters out. The pitcher/batter battle is a chess game, with each player attempting to outthink his opponent. The pitcher, who pitches with the lead, can do more that a batter cannot predict, than the pitcher who pitches with a deficit. So, R.A. Dickey gets the 15th best run support from his team in the National League at 4.61 runs/game while Clayton Kershaw, who ranks second in fWAR and rWAR, has had the NL's 8th worst run support at 3.77 runs/game. Maybe, Dickey has it easier because the Mets give him a lead more often than the pitchers ahead of him. Whether this is true or not, it does not mean he is necessarily a worse pitcher than Kershaw.

So, should R.A. Dickey win the NL Cy Young award. Looking only at the statistics, I have to say I think Gio Gonzalez deserves the Cy Young award. Walks can get a pitcher into trouble quickly, and while Gonzalez has a 3.4 BB/9, his inability to give up homeruns means that those walks do not come back to bite him as much as other pitchers'. In addition, Gonzalez strikes hitters out. He does so to the tune of 9.36 K/9, second only to his teammate Steven Strasburg. Gonzalez even satisfies old school voters with 20 wins, 2.84 ERA, and 201 strikeouts.

Okay, Gio is good. In fact, he's been incredibly good, but what about my man R.A.? Well, Dickey has statistics fairly comparable to Gonzalez's and one thing that Gio doesn't. Dickey does it all throwing essentially one pitch. Everyone praises Yankees Closer Mariano Rivera for getting out three batters a game with one pitch, but R.A. performs a similar feat but multiplied by 7. Now, the knuckleball is different from the cutter. The cutter is merely a variation of a fastball, whereas the knuckleball is as similar to a fastball as a rocket ship. It is almost impossible to prepare for a knuckleball because event he pitcher doesn't know how the pitch will move and where it might go. While the knuckleball is difficult to hit, it is even more difficult to throw properly. When thrown improperly, a knuckleball is as straight as an arrow and looks as big as a beach ball to opposing hitters. This fact is seen by Dickey's homerun percentage.

Dickey throws the most difficult pitch known to man and does so incredibly well. Gio Gonzalez does the job a hair better than Dickey, but uses the typical pitcher's arsenal. Were Gonzalez above and beyond the better pitcher, I would award him the Cy Young without question, but due to the difficulty it takes to control a knuckleball, and the even greater difficulty it takes to dominate opposing hitter with said pitch, I think Dickey deserves the award. Phil Niekro is known as one of the best pitchers to ever throw a knuckleball, yet he never attained the success that Dickey has. The same is true of Tim Wakefield, the most recent pitcher to throw a knuckleball. Dickey's numbers show he is one of the best pitchers in the National League this year, and the way in which he has dominated NL hitters warrants him the Cy Young Award. So, here are my top 3 in the NL as of today:
1. R.A. Dickey
2. Gio Gonzalez
3. Cliff Lee

No comments:

Post a Comment