Jim Albright / the japanese insider
Most Dominant/Most Dominated Single Season Teams In Japanese Baseball History
Our webmaster, Craig Tomarkin, suggested after I submitted my article on Japanese baseballs dynasties that I consider doing an article on Japans best single season teams. He further suggested I look at his articles on the greatest teams in major league history since 1893. Its four articles: Part I Part II Part III Part IV .
I read the articles and while just like with every rating system (including my own), I wouldnt agree with all the conclusions it reached, it looked like Craigs suggestions were a very promising idea. Ive also been working on assessing sustained success and failure of Japanese teams, so the thought of using Craigs system or a mirror image of it to determine the worst teams appaealed to me. One thing I liked about Craigs last formula to do his ratings was the use of the ratio of runs scored to runs allowed. Those of you familiar with Bill James work realize there is a predictable relationship between winning percentage and this ratio. Thus, unless a team is unusually lucky or unlucky, Craigs adjustment merely reinforced the teams position. However, if it was unusually lucky or unlucky, this adjustment helped address this factor and thus more accurately assess how a team did. In fact, I believe the system does a fine job of assessing the dominance of a team.
When I tried to apply Craigs approach to Japanese baseball, I eventually learned a hard lesson. Namely, assessing a teams dominance is not quite the same thing as assessing its greatness. True, all great teams are, by definition, dominant. However, my experience in doing this article convinces me that not all dominant teams are great ones. Craig certainly understood much of what I will write below when he wrote his own articles. He acknowledges he is measuring dominance, and that if he evaluated 19th century teams as well, the rating system would place an unusually high number of those teams at the top of his list. Similarly, when I suggested in an email to him that he consider joining me in rating the worst teams that the list would be dominated by teams early in the period selected (assuming he went back to at least 1900).
I would agree with Craigs sense that the 19th century teams should not be seen as consistently the greatest teams in baseball history. However, I believe it is reasonable to assert that they were consistently the most dominant teams in baseball history. The reason is good teams in that era had one tremendous advantage in being dominant over teams from later periods. Namely, that there were often some very bad teams around also. Naturally, the good teams beat up on their hapless foes. It certainly makes them more dominant over their opposition, but it has little to do with the greatness of the team, because great teams are ones which would stack up well against quality opposition.
Getting back to Japanese baseball, there are two eras which had a bunch of exceptionally bad teams. They occurred at the very beginning of Japanese professional baseball during wartime in Japan and for the first five or six seasons after the big expansion in 1950 to two leagues from one. When I tried using Craigs system to identify the best teams and the mirror system to identify the worst teams, both lists were heavily populated by teams from those two time frames. I first thought it might be a function of the shorter schedules of Japanese baseball, but my first attempt to alleviate such an effect had only a small effect on those two eras domination of the two lists.
It finally occurred to me that the problem lay in the dominance versus greatness issue Ive been writing about here at such length. I eventually realized that it was no coincidence that the two lists were skewed toward those eras. In fact, it was common for a single season to be represented on both lists.
That last phenomenon could only logically be explained in three ways: 1) it was merely random, 2) there were a lot of great teams then, and they just beat up on the lesser opposition, or 3) there were a lot of awful teams then, and they were consistently thrashed like rented mules by teams which were good and/or great.
The explanation of random chance has two key strikes against it: one, the clustering of the teams into two eras, and two (and maybe more tellingly) is the frequency with which the same season showed up on both lists. The second choice is possible, but I really favor the bad teams explanation. I have several reasons for feeling this way. First of all, any sport which does not have a free agent system when it starts out and/or goes through a huge expansion has a couple of teams which are real stinkers. It is unlikely that Japanese baseball is any different in this regard. Second, there is the confirmed tendency in all sports for the extremes to reduce over time. Third, there is the assertion by many well-informed observers that the quality of Japanese baseball has improved over time. It is easy to fit all of these facts into a scenario where the quality of play improves, with the largest improvement coming from the worst teams. This scenario would work well with the bad teams explanation. I cannot think of a similarly good scenario to fit these facts and the good teams explanation.
As I indicated earlier, I think Craigs system does a good job of identifying dominance. However, this exercise has taught me that unless a system tries to filter out the difference between a great team and a dominant one, I cannot accept it as anywhere near the final word on a teams greatness. I am aware of one effort which at least seems to try and address this issue. Rob Neyer and Eddie Epstein wrote a book entitled Baseball Dynasties, and their method for identifying great teams relied on standard deviatons from the mean. Id have to refresh my familiarity with the fine points of such mathematics to give you a good assessment of how well theyve dealt with the distinction between dominance and greatness, but its clear they have tried to do so.
Frankly, I decided I wasnt going to take the time to bone up on my math nor to program Excel to do the standard deviation calculations Id need to try and replicate the Neyer-Epstein approach or some similar idea. I settled for 1) titling my article in terms of dominance rather than greatness, 2) manipulating the evaluation system to give teams from every era a chance to show up on either list, and 3) confessing here that such manipulations require a definite compromise in truly assessing dominance. I retained Craigs idea of using the ratio of runs scored to runs allowed in the formula, and the idea of using the mirror image of the system used to identify dominant teams to identify the teams which were most dominated. The formulas are:
((2 times wins) minus losses) times winning percentage times (runs scored divided by runs allowed), plus a postseason bonus I will describe later
For being dominated:
((2 times losses) minus wins) times (1 minus winning percentage) times (runs allowed divided by runs scored)
If a team is around .500, it should allow about as many runs as it scores, and both systems should come close to replicating wins (dominance) and losses (being dominated) plus one half of ties. The dominance formula will push high win teams which score a higher ratio of runs relative to their opponents dramatically forward, and the being dominated formula will push forward exactly the opposite kind of team.
I wound up using the two times wins figure in the dominance calculation in order to 1) give every era a chance to get a team on the list, and 2) to minimize the effect of random chance on a shorter schedule. My two lists still wind up with heavy representation from the two time frames in question, the startup in wartime era, and the big expansion era. Since I am measuring dominance, I can live with those eras dominating the lists. I do believe that the most dominant teams are quite good, but I have reservations in asserting it accurately represents the greatest teams in Japanese baseball history based on the fact the lists are both skewed toward those two eras. As you might guess, I am considerably more comfortable in asserting that the "most dominated" list is close to a correct rendering of the worst single season teams in Japanese baseball history.
There is one aspect of the dominant formula which is not mirrored by the most dominated formula. That is the playoff bonus aspect. This part of the formula exists to push forward Japan Series winners more strongly than those who lost the Japan series, but yet to push those teams forward more strongly than those who only won a split season title without winning the whole season, but yet to push those teams forward more than those who didnt even qualify for any playoff. A team which wins its league for the full season receives 7 points, and a split season champion which does not win the league for the full season or a team qualifying for a league championship series but fails to get to the Japan Series gets 3.5 points. Finally, a team which makes a playoff but fails to play for the league championship gets 1.5 points. I also awarded two points for every Japan Series win (four has always been the mark to win the series). I allowed one half a Japan Series win to those full season league champions who predated the Japan Series (the 1939-1949 league champs, in other words). These relatively small bonuses helped sort the results in a way that the list is dominated either by teams which won the Japan Series or who predated it.
Below are the two lists for your consideration, first the one for the dominating teams, then the one for the most dominated teams.
Most Dominant Teams
Most Dominated Teams
Notes: In "Playoff Status", a value of "1" means the team won the league in a full season, a value of "0.5" means the team won a split season half, but either lost in the playoffs of no playoff existed. Teams winning their league before the Japan Series started in 1950 were awarded one-half a win in that category.