June 19, 2019

WHAT’S ON SECOND?: The Japanese are Coming Part II REDO

November 26, 2007 by · 2 Comments 

Before moving on to Part III, there’s a clarification I need to make regarding Part II.

In addition to factoring in the difference in league parks between any two leagues when doing MLE’s, you must also factor in the difference between league SCORING environments. In the previous Part II, I ‘plugged’ a 7% difference into the formula and called it the park difference. In reality, the 7% difference is actually the ‘context’ difference – a combination of park differences AND league scoring differences. Since we can calculate the league scoring difference, we really should use it, and plug ONLY the park difference.

In terms of the final results, it won’t change anything, but in terms of understanding how to ‘properly’ do MLE’s, not just for Japanese League players but for AAA, or Negro Leagues, etc., and for understanding the interaction of parks, leagues, etc. it’s important to understand how this works.

Japanese league scoring historically is less than MLB. Japanese League managers traditionally have used more one run, ‘small ball’ strategies, and the league scoring environments definitely reflect this. For the period 1962 – 2007, NPB scoring per game was 94% of MLB scoring per game.

I’ll walk thru complete examples for a hypothetical batter and pitcher again, this time doing it completely the ‘right’ way:

Japanese Batter A:

1. Batter Runs Created in NPB = 10.0 (per game)

2. Competition Level increase moving to MLB of 10%. New Runs Created = 9.0

3. Run scoring environment increase moving to MLB of 6% (NPB 94% of MLB). New Runs Created = 9.5

4. Leagues Park Difference, moving to more pitcher-friendly MLB parks, estimated 13% (NOTE: This is a ‘plug’ estimate backed in to by the 14% negative difference in performance in batters moving to MLB vs. pitchers moving to MLB. Assuming batters and pitchers are impacted equally, that would be a -7% for batters and +7 % for pitchers. If batters are -7%, but are getting a +6% scoring environment increase in Step #3 above, then the ‘plug’ is -7% minus +6% = 13%).

New MLB MLE Runs Created = 8.3 vs. Original NPB Runs Created of 10.0

Japanese Pitcher A:

1. Pitcher ERA in NPB = 4.00

2. Competition Level increaseing moving to MLB of 10%. New ERA = 4.40

3. Runs scoring environment increase moving to MLB of 6%. New ERA = 4.66

4. Leagues Park Difference, moving to more pitcher-friendly MLB parks, 13%.

New MLB MLE ERA = 4.06 vs. Original NPB ERA of 4.00.

A few concluding thoughts for Part II:

1. This method works for analyzing or creating MLE’s between ANY two leagues, not just NPB to MLB.

2. As you can hopefully clearly see now, there are specific reasons that batters moving from NPB to MLB ‘suffer’ much more than pitchers moving from NPB to MLB, beyond some nebulous ‘position players don’t adjust as well’ theories.

3. Steps #3 and #4, that I originally compressed into one step, will almost always have some overlap. Breaking it out into two steps is really an artificial break, that I decided to add for clarity. We can only really measure the Step #3 factor, and assume the Step #4 amount, but League Run Scoring is obviously impacted by the types of parks in that league. For example, if the International League scores at 8% more runs per game per team than MLB, and we ‘plug’ in a Step #4 percentage of IL parks being 5% more batter-friendly, in order to get to the correct observed empircal results for IL batters moving to MLB, then I’d be almost certain that IL parks are really even MORE than 5% more batter-friendly vs. MLB parks, thereby partially ‘causing’ the the IL 8% league scoring difference. Regardless, the TOTAL adjustment factor for steps 3/4 would be 13%, but maybe the ‘real’ cause of the context difference is 7% due to parks and 6% other factors instead of 5% and 8%.

In Part III, we’ll finally get down to actually doing MLE’s and projections for the best Japanese players possibly moving to MLB in 2008.


2 Responses to “WHAT’S ON SECOND?: The Japanese are Coming Part II REDO”
  1. Paul Wendt says:

    Which of the three component adjustments at steps 2-3-4 are calculated and which are guestimated? You call the third adjustment a “plug” which is vague to me. It suggests a shortcut, leaving some analysis incomplete (or while some analysis in progress is still incomplete), but the basis for the first two “competition level” and “run scoring” adjustments is vague too.

    This may help. Are you using some average distance to the outfield fences as follows? The difference dR in average runs per game is one part difference dF in “fence distance” and one part everything else.
    dR = dF + other

  2. KJOK says:


    The 10% of step two is a guestimate, based on previous studies. Now that I have abundant data for U.S. Minor leaguers playing in Japan, I’ll be able to confirm a more exact figure soon.

    The ‘plug’ is really not a shortcut at all, but simply the missing variable of an A + B = C type of equation, where you know A and C, so you ‘plug’ B.

    I’m not sure where I mentioned the average fence distances, but they’re really not used at all in the calculation – they were just a comment about the fence differences not seeming to explain the total difference in HR’s between MLB and NPB. I did use average LF, CF, RF of MLB vs. NPB, which is not the most accurate comparison as 5 ft difference down the lines makes more HR difference than 5 ft to straightway center field.

Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!