Sunday, September 30, 2012

How our Japanese Projections did

So as the season in Japan and America is about to wrap up, it is time to look back at the projections I used to evaluate players coming from the NPB to the MLB and vise versa. Here is the major pitcher projection I used, and here is the major pitching projection.

As Jeff Sullivan points out, success in Japan to America isn't always linear. These projections basically assumed they were. The off-season was a long time ago, and regular readers will notice that their has been some revolution of thought on this blog. Now, instead of using the projections, I would scout more (video of most players are online) and use NPB Tracker's pitch data to evaluate pitchers. I've done this in other posts when looking at players in the two major leagues in East Asia. I just wanted to see how the projections did for fun and of course, to see if they had any predictive value.

From Japan to U.S:

Tsuyoshi Wada: After a spring where it was apparent he wasn't healthy, he had Tommy John Surgery and he didn't pitch in the regular season.

Yu Darvish:
Projection: 2.42 ERA, .995 WHIP, 192 strikeouts, ~6.5 WAR, worth 20-24 million.
Actual: 3.90 ERA, 3.30 FIP, 1.27 WHIP, 221 strikeouts, 4.9 WAR, 22 War Dollars (according to Fangraphs). I basically projected him to be around Cliff Lee statistically, and in FIP and FIP -, he was right around Cliff Lee.

Norichika Aoki
Projection: .388 OBP, .95 PPG, 2.97 PAPP, ~ 3.64 WAR, worth ~ 5-11 Million
Actual: .355 OBP, 3.2 rWAR, 3.1 fWAR, 13.7 WAR Dollars (according to Fangraphs)

Wei-Yin Chen
Projection:  1.162 WHIP 3.02 ERA, 6.072 K/9IP, ~5 million
Actual: 1.27 WHIP, 4.11 ERA, 4.43 FIP, 7.26 K/9IP, 9.5 WAR Dollars (according to Fangraphs)

Hisashi Iwakuma
Projection: 1.29 WHIP, 3.86 ERA, 77 strikeouts, ~5-10 million
Actual:  1.28 WHIP, 3.32 ERA, 4.51 FIP, 94 strikeouts, 2.1 WAR dollars (according to Fangraphs)

Munenori Kawasaki
Projection: .311 OBP, .61 PPG, 3.5 PAPP, and -1.15 WAR
Actual: .255 OBP, -.4 rWAR

From U.S. to Japan:

John Bowker
Projection: .316 OBP, .06 WAR
Actual: .261 OBP

Lastings Milledge
Projection: .361 OBP, .789 OPS, 1.43 PPG, 2.22 PAPP, 2.13 WAR
Actual: .379 OBP, .865 OPS

Jorge Sosa:
Projection (I didn't do a projection at the time, this is a back dated using the same transfer rate): 4.11 ERA, 1.396 WHIP, 7.69 K/9IP
Actual: 1.91 ERA, 1.174 WHIP, 7.0 K/9IP

Kam Micolio:
Projection (back dated again): 4.22 ERA, 1.525 WHIP, 13.01 K/9IP
Actual: 2.89 ERA, .964 WHIP, 8.7 K/9IP

Esteban German:
Projection (back dated): .392 OBP
Actual: .347 OBP

Wily Mo Pena:
Projection (back dated): .336 OBP
Actual: .333 OBP

(Victor Diaz did horrible in a short stint in Japan this year, but had been in the Mexican League for the past two seasons after stints in the KBO and Independent ball. He hasn't been in the Majors since 2007, so those statistics seem pretty irrelevant. Brad Penny made one bad start before hurting his elbow and coming back to the States where he has been pitching terrible for the Giants.)

I think a big finding is that the projections really underestimated how friendly the league is for pitchers and how devastating it is for hitters (the exception is obviously Lastings Milledge, who outperformed the projection by a lot). This has really changed over the past two seasons, as I understand it, a baseball change has made the league even more pitcher friendly. Any future projection system (not that I will design one), has to take this into serious account. However, this only applies to runs, as the projections overestimated how prevelant strikeouts are in the NPB, as the ones that came to America usually had more strikeouts and the ones that went to Japan had less strikeouts than the projections said they would. Of course, none of this adjusts for parks and that makes it at least somewhat inadequate. My metric on the relation between WAR and salary is much different than Fangraphs' (it is much more conservative, as mine is about 3 million per win, while Fangrahps is about 5 million per win).

No comments:

Post a Comment