Tuesday, November 14, 2017

Projecting Otani

With Shohei Otani coming to the major leagues in 2018, I thought it would be fun to attempt a simple projection of what his big-league stats might look like. I'll do batting stats first. There have been nine Japanese players who amassed at least 300 plate appearances their rookie years in MLB:

PlayerYearTmGPAABRH2B3BHRRBISBBAOBPSLGOPS
Ichiro Suzuki2001SEA15773869212724234886956.350.381.457.838
Hideki Matsui2003NYY16369562382179421161062.287.353.435.788
Kosuke Fukudome2008CHC15059050179129253105812.257.359.379.738
Norichika Aoki2012MIL15158852081150374105030.288.355.433.787
Tadahito Iguchi2005CHW13558251174142256157115.278.342.438.780
Akinori Iwamura2007TBD12355949182140211073412.285.359.411.770
Kenji Johjima2006SEA1445425066114725118763.291.332.451.783
Kazuo Matsui2004NYM1145094606512532274414.272.331.396.727
Tsuyoshi Shinjo2001NYM1234384004610723110564.268.320.405.725
Average14058252377151294116316.289.350.425.775

Next is a weighted average of those players' last three seasons in Japanese ball, with their final season weighted at .5, their next-to-last weighted at .3, and their third-to-last weighted at .2:

PlayerGPAABRH2B3BHRRBISBBAOBPSLGOPS
Ichiro Suzuki11048242276155262157116.367.435.544.979
Hideki Matsui139613489111162272441063.330.457.6631.120
Kosuke Fukudome1084703868812633321768.325.437.5901.028
Norichika Aoki1436465738218027395413.314.390.421.811
Tadahito Iguchi12556749594160312248826.323.396.535.930
Akinori Iwamura1436185448717027234907.312.387.555.942
Kenji Johjima1215064448314326330805.323.401.594.995
Kazuo Matsui140645579107181374328322.314.372.559.931
Tsuyoshi Shinjo12951748059125223196510.261.307.442.749
Average12956349087156283257912.318.398.542.939

And here are those numbers again, but prorated to the number of plate appearances they had their rookie seasons in MLB:

PlayerGPAABRH2B3BHRRBISBBAOBPSLGOPS
Ichiro Suzuki1697386471172384032310925.367.435.544.979
Hideki Matsui158695555126183302501214.330.457.6631.120
Kosuke Fukudome136590485110158413279511.325.437.5901.028
Norichika Aoki1305885217416424394912.314.390.421.811
Tadahito Iguchi12958250997164322249027.323.396.535.930
Akinori Iwamura1305594927815324231817.312.387.555.942
Kenji Johjima1295424758815328332855.323.401.594.995
Kazuo Matsui11050945684143293256617.314.372.559.931
Tsuyoshi Shinjo1094384065010619316559.261.307.442.749
Average13358250592162303268413.322.403.547.951

And here are Shohei Otani's batting stats and weighted average for his last three seasons with the Nippon Ham Fighters:

YearGPAABRH2B3BHRRBISBBAOBPSLGOPS
2015701191091522405171.202.252.376.628
20161043823236510418122677.322.416.5881.004
20176523120224671618310.332.403.540.942
Wt. Avg.78254220356914112392.315.395.545.939

Now, I'll move through binary component rates, projecting his stats one at a time. First, walks and HBP:

$BB
Rookies (Japan)12.4%
Rookies (MLB)8.8%
Otani (Japan)12.2%
Otani (proj. MLB)8.7%

The nine Japanese players walked or were hit by a pitch in 12.4 of their plate appearances in Japan, compared to just 8.8% of their PA their rookie seasons in MLB. So we would expect a Japanese rookie to walk or be hit by a pitch 71% as often as his established rate in Japanese baseball. Otani's weighted $BB rate is right in line with the other Japanese rookies, just a smidge lower. So his projected rookie MLB $BB rate is 8.7%, which means in a 600 PA season, he would walk or be hit by a pitch 52 times. (I know Otani won't actually get 600 PA in the majors if he's pitching, but it's more interesting to see what a full season of batting stats would look like.)

Next, I isolate HBP by finding its rate per (BB + HBP):

$BB$HBP
Rookies (Japan)12.4%10.6%
Rookies (MLB)8.8%11.7%
Otani (Japan)12.2%4.2%
Otani (proj. MLB)8.7%4.6%

So here are Otani's projected BB and HBP, in a 600 PA season:

PlayerYearLgPABBHBP
Otani2018MLB600502

Next, strikeouts per plate appearances ending in a strike (PA - BB - HBP):

$BB$HBP$SO
Rookies (Japan)12.4%10.6%16.4%
Rookies (MLB)8.8%11.7%15.5%
Otani (Japan)12.2%4.2%31.2%
Otani (proj. MLB)8.7%4.6%29.3%

Japanese rookies actually struck out a little less often in MLB than they had in Japan. Otani strikes out a lot - 31% of PA ending in a strike - which translates to a projected 29% as an MLB rookie. Which means we would expect about 161 strikeouts in a 600 PA season for Otani:

PlayerYearLgPABBSOHBP
Otani2018MLB600501612

Next is homeruns per batted ball (PA - BB - SO - HBP):

$BB$HBP$SO$HR
Rookies (Japan)12.4%10.6%16.4%6.2%
Rookies (MLB)8.8%11.7%15.5%2.5%
Otani (Japan)12.2%4.2%31.2%7.6%
Otani (proj. MLB)8.7%4.6%29.3%3.1%

Here is the biggest difference from Japanese ball: Japanese players have seen their rates of homers per batted ball plummet 60% from their established rates in Japan. And here is where I'd take this projection with the biggest grain of salt. The nine Japanese players who came before Otani are of the previous generation. The youngest is Aoki, who was 35 in 2017 (Otani was 22). These players were entering a league where they were clean, but many of the American (and Latin) players were juiced. Now, the players are clean, but the ball is likely juiced. But that means it's juiced for everybody. Also, there's the possibility that the quality of competition in Japanese baseball has improved faster in the last 15 years than it has in MLB.

But based on what Japanese players have previously done, we would expect Otani's rate of home runs per batted ball to fall from 7.6% to 3.1% his rookie year in MLB. That means in a 600 PA season, he would hit just 12 (387 * .031) home runs:

PlayerYearLgPAHRBBSOHBP
Otani2018MLB60012501612

Next is hits per balls in play (H - HR) / (PA - HR - BB - SO - HBP):

$BB$HBP$SO$HR$H
Rookies (Japan)12.4%10.6%16.4%6.2%34.0%
Rookies (MLB)8.8%11.7%15.5%2.5%32.0%
Otani (Japan)12.2%4.2%31.2%7.6%40.6%
Otani (proj. MLB)8.7%4.6%29.3%3.1%38.1%

Japanese players have seen their BAbip fall a little bit, but remain above the Major League average. This projection doesn't regress to the mean, so it assumes Otani would retain most of his absurdly-high .406 BAbip:

PlayerYearLgPAHHRBBSOHBP
Otani2018MLB60015512501612

To save time, here are the rates of extra-base hits ((2B + 3B) / (H - HR)), triples (3B / (2B + 3B)), steal attempts ((SB + CS) / (1B + BB + HBP)), and steal successes (SB / (SB + CS)):

$BB$HBP$SO$HR$H$XBH$3B$SA$SB
Rookies (Japan)12.4%10.6%16.4%6.2%34.0%23.8%7.9%9.6%76.3%
Rookies (MLB)8.8%11.7%15.5%2.5%32.0%23.8%12.0%13.9%74.7%
Otani (Japan)12.2%4.2%31.2%7.6%40.6%26.1%5.3%4.6%67.6%
Otani (proj. MLB)8.7%4.6%29.3%3.1%38.1%26.1%8.1%6.7%66.3%

To get runs scored and at-bats (and therefore, traditional rate stats), I use the following formulas:

$R = (R - HR) / (H - HR + BB + HBP)
$SF = SF / (PA - H - BB - SO - HBP)

Otani didn't have any sac bunts in the last three years, so I don't have to worry about those. Then I just subtract BB, HBP and SF from PA to get his at-bats. I didn't project RBI, but I think you get the picture: a season reminiscent of Hideki Matsui's rookie year:

PlayerYearLgPAABRH2B3BHRSBCSBBSOBAOBPSLGOPS
Otani2018MLB60054068155343127450161.287.345.428.773

This projection might be pessimistic, because it doesn't account for Otani's young age (younger than any previous Japanese MLB rookie by four years) or the allegedly tighter ball MLB now employs. But even if he homers more often than I project him to, the moniker "Japanese Babe Ruth" might be giving people unrealistic expectations. He's no Babe Ruth as a hitter, but he IS good enough to be a starting outfielder or DH on the days he's not pitching, and, given he's also an exceptional pitcher, that's exciting enough.

Just how exceptional is Otani's pitching? I'll project his pitching stats in a separate post.

1 comment: