Using kwERA to Project Masahiro Tanaka

Masahiro Tanaka

In just a few months Masahiro Tanaka is very likely to throw his first pitch in the major leagues. Hopefully, he will do so as a very wealthy member of the Los Angeles Dodgers.

If you’re reading this post, you probably have some interest in how Tanaka will perform.

A lot of people who know a lot more about baseball, Tanaka, and statistics have tried to answer this question. But, I’m going to throw in my two Yen anyway.

To my count there have been seven starting pitchers who have transitioned from Nippon Professional Baseball to Major League Baseball since 2007. A pretty arbitrary date that leaves out guys like Kaz Ishii, Hideki Irabu, Masato Yoshii and Hideo Nomo. I chose 2007 because between Ishii in 2002 and Dice-K and Igawa in 2007 there weren’t any starters that came from Japan that made regular starts and that five-year gap made for a convenient, if nothing else, endpoint.

The pitchers that I included in the study were: Yu Darvish, Hisashi Iwakuma, Hisanori Takahashi, Kenshin Kawakami, Hiroki Kuroda, Daisuke Matsuzaka, and Kei Igawa.

While gazing upon the stats of these pitchers I noticed that, for the most part, they were able to maintain their strikeout rates after moving to the MLB. I also noticed that most had a significant increase in walk rate.

Looking at the pitchers’ last three years in the NPB and first two season in the MLB:
K and BB rates for NBP pitchers

Overall the pitchers in the NPB K’d 22.7% of hitters and 21.4% in the MLB while walking 6.1% in the NPB and 9.5% in the MLB. Most of the pitchers in the sample followed this pattern of similar K rates and increased BB rates. The correlation between the K%’s had an r-squared of .37 and the BB%’s had an r-squared of .32.

This lead me to kwERA. kwERA uses only walks and strikeouts to project ERA. The formula is quite simple: 5.40 – 12*((K-BB)/PA).

There’s some problem with this. kwERA uses a constant (5.40) instead of a factor that changes each year. I could have found that constant, but didn’t. The correlation between K% and BB% isn’t THAT strong. But, it is what it is.

The kwERA for the combined K% and BB% of all the pitchers in the sample for their last three NPB seasons was 3.38. Their kwERA for their first two seasons in the MLB was 3.83. That’s an increase of .46 runs or 14%.

Next, I calculated Tanaka’s kwERA for his last three NPB seasons. It came out to 2.81 (24.9% K-rate and 3.3% BB-rate). Adding .46 runs to that gives an expected MLB ERA of (2.81 + 0.46) 3.27. Increasing his kwERA by 14% gives an expected MLB ERA of (2.81 * 1.14) 3.19. I then averaged these to get an expected MLB ERA of about 3.23.

I went back and looked at how well this expected kwERA compared to how the pitchers actually performed in their first two MLB seasons. I did this by comparing the average of adding 0.46 to the expected kwERA and increasing expected kwERA by 14% (xkwERA) to the average of the pitchers’ ERA, FIP, xFIP, and SIERA over their first 2 MLB seasons (EXFS AVG).

By far the biggest miss was Igawa. His MLB struggles are well documented. Also this method could not have accounted for Iwakuma performing better in the MLB than he did in the NPB. The other cases are within around half of a run difference.

If Tanaka signs with the Dodgers and pitches to something like 3.23 earned runs allowed I’ll be ecstatic.

Comments, criticisms, suggestions all welcomed.

Print This Post

Sort by:   newest | oldest | most voted
Jonah Pemstein

So you’re a Giants fan? Couldn’t quite tell.

In all seriousness, though, good job. This is interesting.


Now that he’s reportedly a Yankee, how do you think he’ll pitch? ;) Nice breakdown in any event!

Chris Mitchell

Interesting work. I did a similar analysis using K%, BB%, and HR% and came up with a 2.92 ERA. As a Yanks fan, I hope my projection’s the more accurate. 3.23 feels more reasonable though.


Interesting analysis. Nice work!