The comparison of the Japanese Leagues vis-a-vis MLB is interesting in its own right, but the key section for us is about 2/3 of the way down. The author mentions similar ratings of the Federal League, the American Association, the Union Association, and the Players League, all of which have interest to us here.

I'll be quite interested in seeing Robert's work when it becomes available.

I did research along these lines last year here http://baseball.fanhome.com/forums/showthread.php?threadid=403087

There are certain assumptions that you can make that will drastically alter the results.

Please, read through the whole thread before making any comments. And if you choose to criticize, try to keep the comments to the methodology rather than the conclusions.

That piece of research was done as a starting point, and posting on fanhome was to try to get feedback from my peers.

Joe/scruff : It's funny, but I am surprised that anyone would be surprised that Japanese baseball would be that strong. (Does that make sense?)

What makes a league strong is strong players, and what makes strong players is a population of good athletes with the motivation to play baseball.

The question of getting elite athletes is first and foremost simply one of population: and yes, Japan draws its players from a population half as large (very roughly) as MLB does, but the league is half the size of MLB, and salary levels for most of history have been roughly equal in terms of where they are on the overall salary scale. So there has been roughly the same motivation for becoming players; as well, the two countries give roughly equal social prestige to elite athletes and baseball players in particular.

So to me, there's no question that the two leagues (MLB and the Japanes leagues) should be of roughly equal strength... which they are. The .94-.95 coefficient does represent a significant difference; this could represent anything from physical characteristics (which, of course, affect baseball far less than most other North American sports) to training methods (there is at least some anecdotal evidence, it seems to me, that lifting weights will help a player more than taking five million grounders) to the pyramidal nature of the development system in the U.S.

Given the existence of this information, I think there's significant reason to at least permit the nomination for the HoM of players who played their whole careers in the Japanese leagues after the widespread introduction of foreign talent.

I am currently working another research project; once I am finished that I might take a stab at extrapolating Clay Davenport's research back to the 80s and 70s and beyond.

The nice thing, of course, is that we don't have to do this right away: we can time to get the data in before these decisions will become important in the 1960s and 1970s. And if we find we're not ready then, we can wait until 1985. Or 1999. Or the 2007 election, for that matter. It's nice that nobody drops off the ballot for this reason...

I agree Craig. I just always thought that if they were that good, they'd just come over here, to play with "the best". Pretty arrogant, I realize, but still, that's what I thought.

I do have one issue with comparing player stats from year to year. Most of the players that go to Japan are past their peaks, or at level in their peak. Few are still developing. As a player ages past 27-28, his ability declines (generally). So if you take a 32 year old player and compare him to himself, generally he'll look worse at 33, no matter where he's playing. So I think there might be an age bias in these studies. That would have to be controlled for at a minimum to increase the validity of the findings.

As for the HoM, I think you are correct. When did the Japanese leagues first come into existence? Oh played in the 60's, right? I would have no issue with voting in some of their superstars. The marginal candidates would be a much tougher sell, but the top end guys should be given just as much consideration as the Negro Leaguers, IMHO. They key however will be for some "experts" the ability to present solid data arguing that the Japanese players are X% relative to the majors. Once that data is presented it shouldn't be too tough to compare the two.

This doesn't me we even need Japanese-U.S. stat line translations. If we know the league is 90% for example, we can eyeball that the stars there would have been stars here, and just use a slightly tougher standard for them. MLE's would be nice, but aren't absolutely needed.

If we are going to include Japanese players, I would want to add a few to the total inducted, our number wasn't anticipating them.

Tangotiger -- Interesting thread, I'm not through it yet, but that's some pretty good work so far. Have you refined the study since that posting?

Tangotiger,

Interesting post. I have only looked at your work for about an hour, so I can only comment on your first methodological choice (pool of players.) I think, however, that it's a very important point and should be considered before looking further.

You include all players between age 23-33 (peak years). Further down in your post you run numbers again using ages 23-29.

I ran through the National League hitters in 1945 and 1946 (I assume the AL is the same.) I marked down all players with at least 100 ABs in both 1945 and 1946, as these are the main numbers going into the calculation. Here's what I found:

Players Age 23-26 in 1945 with at least 100 ABs in both years: 7
Players Age 27-33 in 1945 with at least 100 ABs in both years: 28
Players Age 27-29 in 1945 with at least 100 ABs in both years: 14

I think you can see where I'm going here. The first group are players who are likely to IMPROVE in 1946 based on age. The second group are players who are likely, as a group, to DECLINE in 1946.

The second group is 2 TO 4 TIMES LARGER THAN THE FIRST.

I don't know offhand, but I would not be surprised in the average age of a ballplayer is well over 27.

So, when you conclude that the "playing conditions caused players to lose 5% of their hitting stats" in Step 3, I'd respond that it would be shocking if this pool of relatively old players DIDN'T decline by 5%, not because of playing conditions, but because they are OLD.

I think this age issue may taint your whole analysis.

Hmm. That was more capital letters than I normally use. My intent was emphasis, of course. Not yelling. I think your basic framework is sound, but I do think it needs some rethinking.

Sruff: no have not refined it yet, as I've since discovered more problems with the year to year comps, and that has to do with selective sampling and regression. You can find more information in a similar study I did here: http://baseball.fanhome.com/forums/showthread.php?threadid=662692&pagenumber=1

Not sure I followed your math. Why are we dividing by nine?

Generally, though, your conclusion in the link was that each year the league is 1% better than the year before. But if age considerations makes everyone seem 1% worse v. league every year, doesn't that explain the whole difference?

Two things I would like to point out here.

>> Remember, though, this is in runs, and runs are proportional to EqA to the 2.5 power.

The .941 of the article's "Davenport translation" is equivalent to a run factor of .859, if I'm interpreting this properly. This means that while Japan is close, it is not 94% of MLB. As I pointed out on another thread, a .500 MLB team could expect to play about .650 ball over there. (Or a .350 team would expect to go about .500 over there.) If it was a .94 run factor, .500 here would go .560 over there. The quality disparities between NBA East and West or between AFC and NFC have gotten worse than that at times.

The other is the age discussion. How important the age distribution of your sample is when doing these studies depends on how accurate a result you are trying to get. When trying to get a rough indication of a league's quality, it really is not an important issue. If all the players going from MLB to Japan are declining 2% per year and all the players coming from Japan to MLB are improving 3% per year, then the run factor may be .81 instead of .86. So an MLB .500 team plays .700 ball instead of .650 over there. It's not critical until we get to ranking players that are fairly close. We're then debating whether the 150 RBI guy over there is 122 or 129 RBI's over here.

For the type of study that tangotiger is doing, measuring small changes year-to-year to draw large conclusions over time, getting this bias out of the study is crucial. Small biases compounded over 50 or 100 years can make the difference between concluding that Babe Ruth would be comparable to Mark McGwire or that he'd be unable to get out of AAA these days.

jimd, you're right about small differences blowing up into large magnitudes. That is why I would be (and am) uncomfortable with making too much of an attempt to discount past _eras_ (as opposed to intra-era comparison between leagues), with the exception perhaps of the pre-organized era.

It's very unlikely that enough voters will be convinced enough to vote for anyone from Japan outside of Oh, although others are certainly a possibility. Nomura, Kaneda, possibly Nagashima,

jimd: you are absolutely right

matt: I had said 1 part pre-26, 2 parts peak, and 6 parts post-30, for a total of 9 parts. A gain of 3% in the pre-26 group and a drop of 2% in the post-30 group IN THIS EXAMPLE will show a 1% drop overall. However, this distribution of age is NOT the typical. The data group I had chosen (23 to 33 or 24 to 29 or whatever it was) did show a fairly normal distribution, such that, on average each year, the gains and losses cancelled out. At worst, there could be a 1% error. Typically the error would be less than .2% (1 part, 2 part, 2 part respectively would give you that error rate).

However, these small error rates, when compounded, would have an enormous impact. I agree that we have to make too many assumptions to come to any firm conclusion.

I would think that players should be compared to their peers when it comes to HOF discussion.

By the way, I like the "tiered" HOF. The "typical" HOF can be at level 2, the Ruths at level 1, and the Jim Rices and Kirby Pucketts at level 3.

Tangotiger:
It's nice to see somebody trying to correct the problems that plagued Dick Cramer's stab at comparing players over different eras! I'm also glad to see that you're looking at the aging distortion that derailed Cramer's original work.
With that said, I can't believe Ruth would be an average player. I have no doubt he wouldn't be as dominant, but he would still be HOF (or HOM) bound.
Have you tried to look at the standard deviation of batting for a particular age? I remember when Bill James was discussing Cramer's work that a statistical illusion will cause the average batting stats to remain constant at every age level. In other words, the batting line for a 22 year old would be comparable to a 39 year old. Obviously, this is not so because we know players age. But if the statistics remain constant, wouldn't the standard deviation change? If we can get a handle on the SD, you might be able to expand your project to use all available year-by-year comparisons so as to increase your sample size.
I haven't done any of the work yet because I'm lazy [:-)], but hopefully sometime this weekend (unless somebody beats me to it) I'll see if I can observe a change in the standard deviation. I think the methodology (as I believed so twenty years ago) is sound, but needs to be fine tuned. If we can develop a working formula, we could actually find out how strong the Negro Leagues were (as long as the statistics for the league are reliable!)

The biggest problem with my study was the headline. I learned my lesson not to attack an American icon like that. The conclusion of my study is that Babe would be somewhere between average to great, depending on your assumptions. Granted, a study that has that kind of conclusion is virtually worthless. What does have value if the framework. More effort should be spent on that, before taking another step forward.

I'm not sure what you mean with the stdev, and how I could use it. My next step was to come up with the appropriate aging patterns by age. I published those last year (see link above). But, they are "wrong" because the problem with looking at performance from year-to-year is that it is based also on the playing time that the manager gives the player (regression and selective sampling).

Whatever you choose forces you to make an assumption that would derail the whole thing once you start the "chaining" process.

Tangotiger:
I didn't take your study as an attack on Ruth. I couldn't care less if an icon is devalued in the public's eyes when one can present facts to back up the devaluing. I have no doubt the Babe would loom much smaller in our era than he did in his (or for that matter Cobb, Hornsby, Grove, Wagner, etc.) But it's hard to believe Ruth or Wagner would not be somewhere in star territory. It doesn't mean you are wrong, it's just hard to fathom. I agree that the framework is in place, so it's really just a matter of tweaking here and there.

I'll be looking forward to see what corrections (if needed) will be needed for your study. I think you are damn close!

John M: sorry, I didn't meant to suggest that you were attacking, but rather that others have attacked just because...

I know, it looks like I'm close, but then I spend time doing other research on other things, and I never got a chance to get back to this one. Maybe I'll get back to it when my time opens up...

...that is, others being others on fanhome and usenet.

I reconstructed this thread.

Not sure why you reconstructed this thread John, but since you did, I'll do a bit of self-promotion on my series of articles on Japanese MLE's here:

The Japanese Are Coming

One factor I've not seen mentioned here is that MLB draws not only from the USA but also from the DR, Venezuela and other places. If you just compared US born players vs Japan born players you might get a very different translation factor.

1. Japan also draws players from Venezuala, DR, Korea, etc.

2. The comparison is MLB players vs. NPB players, not really Japanese-born vs. U.S.-born....

Good time to mention this. Among the newbies for the next election is Kazuhiro Sasaki. In his 4-year career in MLB he had 129 saves and a ERA+ of 138. In Japan he had 252 saves with an ERA of 2.41. Any chance he's a viable candidate?

How long did he play in Japan? Was he the dominant player at his position while over there (or among the dominant players?).

Just four years in MLB probably aren't enough for me, but it's close. For guys like Fielder, Reggie Smith, Soriano, Ichiro!, Nomo, Hideki Matsui, I'd definitely consider both continents. Sasaski might be a little short in terms of MLB contribution.

I'm curious as to what others think.

Sasaki didn't retire until after the 2005 season, so not sure he's technically eligible anytime soon.

He had 9 'Mariano Rivera' type years in NPB before coming to MLB.

I guess I should do some MLE's for him....

Sasaski might be a little short in terms of MLB contribution.

You've got to be kidding.

The only reason we excluded Japanese players was because they never played in North America.

Now, when we get to a point where we have Japanese players who played in America, you're going to say they didn't play here enough??

I don't understand this. I never have. The arbitrary exclusion of Asian baseball, in my opinion, our biggest mistake as an institution.

Sasaki didn't retire until after the 2005 season, so not sure he's technically eligible anytime soon.

His last year in MLB was 2003. This is what we have always gone by. That he played back in Japan in 2004 and briefly in 2005 shouldn't delay his eligibility.

That precedent aside, it is conceivable that in some future year we may make a different determination. For instance, suppose Ichiro went back to Japan in a couple years and starred into his 40's. The determination then might be different. For Sasaki, I think we're on solid ground making him eligible in 2009.

Fascinating analysis by Jim Albright at baseball-fever.com, regarding overall placement of Japanese all-time great players.

http://www.baseball-fever.com/showthread.php?32451-Albright-s-musings/page21

http://www.baseball-fever.com/showthread.php?32451-Albright-s-musings/page20

He names 12 players as worthy of his personal Hall of Fame, which is slightly smaller in size to the Hall of Merit:

Akira Bessho
Yutaka Enatsu
Isao Harimoto
Kazuhisa Inao
Masaichi Kaneda
Masaaki Koyama
Shigeo Nagashima
Katsuya Nomura
Hiromitsu Ochiai
Sadaharu Oh
Victor Starffin

Looking mainly at the comparables list Albright produces, Yutaka Enatsu is the only member who looks to fall short of HOM worthiness, while Tetsuya Yoneda's comparables make him look HOM worthy.

From looking at albright's comparables, I would say that Yutaka Fukumoto is on the cusp, while Hiromitsu Kadota and Tetsuharu Kawakami are a cut below the Hall of Merit line.

The Baseball Fever Hall of Fame has elected:

Akira Bessho
Isao Harimoto
Kazuhisa Inao
Masaichi Kaneda
Shigeo Nagashima
Katsuya Nomura
Sadaharu Oh
Victor Starffin
Kazuto Tsuruoka - only member shy of albright's list.

Has anyone found comprehensive analysis of Japanese players besides the fine work by Jim Albright?

Tango was working on a study in 2002 - any new info on this?

Yutaka Enatsu is the only member who looks to fall short of HOM worthiness

Enatsu is a difficult case - he's Japan's version of Dennis Eckersley in that he spent half of his career as a starter, and half as a closer, making his career stats less impressive than if he stayed in one role. He was absolutely dominant in both - as a starter in 1968, he was 25-12, 1.83, and in 329 innings allowed only 200 hits while striking out 401. He won one Sawamura Award, and two MVPs (both as a closer, one in each league - only starters are eligible for the Sawamura, so they were in different seasons than his Sawamura).

He had troubles with both baseball authority and the law (he hit his manager once, which in Japan is unthinkable, and served prison time for drug possession after his retirement). My impression, though, is that if you asked a knowledgable Japanese baseball fan who the greatest lefthanded pitcher in NPB history was, he'd more than likely nominate Enatsu.

Enatsu's NPB stats

18. KJOK Posted: December 03, 2007 at 06:42 PM (#2632482)
Not sure why you reconstructed this thread John, but since you did, I'll do a bit of self-promotion on my series of articles on Japanese MLE's here:

The Japanese Are Coming

KJOK, the link on this is dead, can you repost?

And does anyone have access to the information that Tango posted back in 2002, or can anyone share any new links?

Do we have any MLEs for Japanese play for the 90s and early 2000s? Curious for when Ichiro is eligible and also would like to give Matsui a complete look.

Carl-

I wish I had it bookmarked, but in one of the past ballot discussion threads from about 10 years ago (in real-life), Dr Chaleeko generated MLEs with Win Shares (before the WAR boom), but off the top of my head, I can’t remember what year it was. Maybe he still has those figures somewhere.

I think there is a misconception that it's easy peasy to hit home runs. Mainly this is because of Sadaharu Oh hitting 868 Home Runs, and American League basseball players from time to time having an incredible season. Truth is only two people have careers with over 600 home runs. Oh was an anomaly for Japan baseball.

I don't think its a misconception. There is a long, long list of AAAA or AAA type players who go over to Japan and hit 30+ homers per year. For 2017 that includes Brandon Laird, Zealous Wheeler, Alex Guerrero, Wladimir Balentien, and Jose Lopez.

A good rule of thumb is a player moving from Japan to MLB will lose 40-50% of his homers. Hideki Matsui hit 39 homers per 650 PA in Japan, in MLB he hit 22 (57%). Ichiro went from 19 to 7 (38%).