clock menu more-arrow no yes

Filed under:

Mets playoffs: Using math to identify the most exciting postseason rotations

New, 13 comments

This is nerdy. But fun! Did we say nerdy?

This dude is a super fun pitcher.
This dude is a super fun pitcher.
Gary A. Vasquez-USA TODAY Sports

The 2015 League Division Series began yesterday and the Mets begin their series with the Dodgers tonight. Suffice to say, this has been a fun ride for those of us who've been watching the Mets. For most teams, the focus of playoff discussions now centers around starting pitching matchups. Mets fans are excited to see their young hurlers in action during the playoffs, but how does this staff compare to the other seven playoff staffs in terms of its fun factor?

Carson Cistulli has attempted to measure just that; in 2010, he posited a series of quantitative formulas to determine players' and teams' "aesthetic" appeal. For a pitcher, this is done by weighing the standard deviations from the mean (also known as z-scores) of his expected fielding-independent performance, his strike percentage, and his swinging-strike percentage, and then adding those values to a luck factor (measured in terms of difference between his ERA and xFIP) and a final constant to create a statistic called "pNERD." The formula for pitchers is as follows:

pNERD = (xFIPz * 2) + (SwStrk%z / 2) + (Strike%z / 2) + LUCK + C

(Warning: Lots of math follows. Skip ahead to the data and conclusion at bottom if you need to!)

The LUCK factor is obtained by subtracting xFIP from ERA in cases where ERA is greater than xFIP. Where xFIP is greater than ERA, the number is 0. This way, "bad" luck is added to the equation (Cistulli argued that a pitcher overperforming his ERA is less interesting than one who is underperforming it), and pitchers who seem to be "improving" receive some extra benefit here. The LUCK value is capped at 2 to prevent overskewing.

Below, I've calculated pNERD values for all projected playoff starters. I've substituted Strike% with Fangraphs' updated Zone% ratings. When I calculated z-scores, I limited standard deviation calculations to only starting pitchers with at least 100 innings pitched from that player's respective league. Means for 67 qualifying NL starters were:

xFIP = 3.844 (standard deviation of 0.6249)
SwStrk% = 9.3881 (SD 2.3095)
Zone% = 45.2522 (SD 2.9945)

Means for 66 qualifying AL starters were:

xFIP = 4.0152 (SD 0.5926)
SwStrk% = 9.3121 (SD 1.8714)
Zone% = 45.5894 (SD 2.9519)

After I found my raw scores, my last step was to add a constant of 2.26 to bring the highest score to a perfect 10.

I've assumed for the sake of this article that all teams will go with a conventional four-man playoff rotation. (In reality, it may make sense for a few teams like the Mets, Dodgers, and Astros to only use three regular starters and/or piggyback with five starters.) There are still a number of additional situations to be determined, so I've done my best to forecast them. They are: Steven Matz will be healthy for the Mets, the Cardinals go with Lance Lynn over Tyler Lyons, Mark Buehrle of the Blue Jays indeed retires, and the Astros go with veteran Scott Kazmir over rookie Lance McCullers.

DODGERS ERA xFIP LUCK SwStrk% Zone% pNERD
Clayton Kershaw 2.13 2.09 0.04 15.9 49.3 10
Zack Greinke 1.66 3.22 0 12 40.4 5.63
Brett Anderson 3.69 3.51 0.18 7 45.9 4.13
Alex Wood 3.84 3.9 0 8.2 45.3 2.7
TOTAL 2.6 2.97 0 11.7 44.8 5.63 (1st)
METS ERA xFIP LUCK SwStrk% Zone% pNERD
Jacob deGrom 2.54 2.92 0 12.7 45.8 6.03
Noah Syndergaard 3.24 2.91 0.33 12.2 42.9 6.58
Matt Harvey 2.71 3.24 0 11.6 48.3 5.18
Steven Matz 2.27 3.56 0 8.3 54.9 5.02
TOTALS 2.77 3.07 0 11.9 46.4 5.47 (2nd)
CUBS ERA xFIP LUCK SwStrk% Zone% pNERD
Jake Arrieta 1.77 2.61 0 11.1 44.6 6.69
Jon Lester 3.34 3.06 0.28 10.4 40.9 5.99
Jason Hammel 3.74 3.47 0.27 10.9 45.1 4.08
Dan Haren 3.6 4.69 0 6.2 43.7 5.92
TOTAL 2.96 3.15 0 10.4 43.6 4.98 (3rd)
BLUE JAYS ERA xFIP LUCK SwStrk% Zone% pNERD
David Price 2.45 3.24 0 11.9 49.3 6.2
Marcus Stroman 1.67 3.34 0 7.2 54.2 6.56
R.A. Dickey 3.91 4.72 0 9.1 48.1 5.13
Marco Estrada 3.13 4.94 0 9.4 47 5.64
TOTAL 3.26 4.45 0 9.7 48 4.24 (4th)
RANGERS ERA xFIP LUCK SwStrk% Zone% pNERD
Cole Hamels 3.65 3.4 0.25 13.3 43.9 5.94
Yovani Gallardo 3.42 4.31 0 6.5 42.3 4.56
Derek Holland 4.91 4.45 0.46 6.9 40.1 5.76
Colby Lewis 4.66 4.62 0.04 8.2 47.1 4.89
TOTAL 4.1 4.33 0 8.1 43.9 3.93 (5th)
CARDINALS ERA xFIP LUCK SwStrk% Zone% pNERD
John Lackey 2.77 3.77 0 9.4 49.1 3.14
Michael Wacha 3.38 3.88 0 9.5 48.6 2.96
Jaime Garcia 2.43 3.36 0 9.2 48.9 4.46
Lance Lynn 3.03 3.9 0 8.9 40.4 3.36
TOTALS 2.93 3.75 0 9.3 46.5 2.79 (6th)
ASTROS ERA xFIP LUCK SwStrk% Zone% pNERD
Dallas Keuchel 2.48 2.76 0 10.3 38 8.05
Mike Fiers 3.69 4.04 0 9.8 48 2.88
Collin McHugh 3.89 3.92 0 10.3 45 2.95
Scott Kazmir 3.4 4.14 0 10.3 44.9 3.05
TOTAL 3.89 4.02 0 9.4 47.6 2.64 (7th)
ROYALS ERA xFIP LUCK SwStrk% Zone% pNERD
Johnny Cueto 3.44 3.78 0 10 42.2 3.81
Yordano Ventura 4.08 3.6 0.48 10.4 45 4.53
Edinson Volquez 3.55 4.26 0 9.5 47.7 3.49
Kris Medlen 4.01 4.31 0 8.6 46 3.51
TOTAL 3.97 4.03 0 9.6 45.4 2.42 (8th)

Conclusion

Well, that was fun! A lot of nerdy stats to tell us what we intuitively knew all this time: That among my quasi-arbitrary roster of 32 playoff starters, the Mets indeed have a pretty enjoyable-to-watch rotation. The Mets' staff scored a cumulative pNERD of 5.47, which ranks second of eight. The Dodgers have the combined staff that will be most "fun" to watch (at 5.63), while the Royals have the most "boring" (at 2.42). To nobody's surprise, Clayton Kershaw takes home the honor of being the most exciting pitcher.