Top 50 rikishi since 1958

[Zentoryu 154]

A recent topic about the the Japanese book that is coming out listing the top 20 of the modern era (or since the 6 basho a year format began in 1958) got me to finish this little project I've been working on for some time now.

I used the following point system to rate every rikishi who competed in a least one tournament in the top division over the last 46 or so years:

1 point for every win over losses in KK basho.

15-0 = 15 points

14-1 = 13 points

13-2 = 11 points

12-3 = 9 points

11-4 = 7 points

10-5 = 5 points

9-6 = 3 points

8-7 = 1 point

No points were deducted for losing records or abscences.

Bonus points were added for the following:

Yusho = 30 points

Zensho Yusho = 50 points

Jun-Yusho = 10 points

Special Prizes = 5 points

After many hours of compiling the totals, here are the results:

(Please note that this was done almost entirely by hand, so the likelyhood that a mistake crept in here or there is possible).

First up, the Overall top 50 ranked by total points:

Rikishi 

Edited February 22, 2004 by Zentoryu

[Fujisan 533]

Seems clear the top 10 of the first list correspond almost exactly with Yushos won.

After that it gets a bit more sketchy.

[Mmikasazuma 12]

Thank you Zentoryu for showing this (Weeping...)

I think the reason for seeing active rikishis undeservedly close to the top is simply because you chosed the same measurement for their achievements. At the end of their career they will deserve their places - maybe a lot lower due to accidents, or to long career never again in Sanyaku. (Applauding...)

Others like Asashoryu may collect a lot more points. :-D

And let's hope Kotomitsuki will get his many points too. :-) :-D

[Onnagumo 4]

And let's hope Kotomitsuki will get his many points too. :-D (Weeping...)

(Sign of approval) :-D

[Yoavoshimaru 0]

Very interesting. Good work, Zentoryu-zeki (Weeping...)

[Sasanishiki 57]

Resepct for all the hard work. (In a state of confusion...)

It is interesting to consider what lies behind some of these statistics (or not as the case may be). While using averages per basho etc might help to try and span the gap between different generations when trying to compare great rikishi, this is always going to be a relatively futile exercise. There are always so many intangibles.

Much of the argument that I am going to bring up can be related to other sports such as boxing (of which I have little interest). People are now trying to evaluate where Lennox Lewis stands amongst the all-time greats.

First and foremost, any successful or dominant wrestler does so within his generation, or perhaps a couple if they have longevity. They therefore have to repsond to their opponents. They can only be compared to their opponents because that is the only opposition they have. If a certain period of sumo (or any sport) is considered weak we cannot truly evaluate the strongest athletes from that period against others because they are a product of their environment. They may have become even better if the competition was stronger, or they may not have been as successful if challenged more often.

We need to take into account (in sumo) the relative strength of each heya because of the support various ozeki and yokozuna get by not having to fight their brothers and also in getting their brothers to beat up on other opponents.

I could make other points but I have already typed plenty. Well done Zentoryu-zeki, and it certainly gives us a good guide. However, before we start raving about so-and-so and compare which is better etc., let's be aware of the limitations of such comparisons (Sigh...)

[Takanobaka 0]

Thanks for the analysis, I'm sure it was a lot of hard work and it is interesting (Sigh...)

However, I'm a bit concerned about how you assigned the points. As a student of voter assignment theory, I feel like things could turn out extremely differently if you alter the amount of points that are assigned for each of these a bit, and there's noreal way of knowing what system is the most accurate. If you had it all in an excel spreadsheet and were able to do calculations quickly, I'd suggest you try changing the numbers a few different ways and then use the average rank for each person to determine a (potentially) more accurate placement system, but that's definately a bit much to ask if you're doing it all by hand (In a state of confusion...)

Anyway, my guess is that no matter how you assigned the points, the top four guys would be condorcet winners in relation to everyone else (i.e. no matter how you stack it up, they beat everyone else). After that, things get a bit muddled. However, I'm sure the averages there will fuel some of the Asashoryu fans, especially since (I would imagine) if he wins 2 of the next 3 basho and performs well in the third he would likely jump up to #4 of all time for success per basho (calculation here for math nuts: 19.105 x 19 = 363 + 30 + 30 + 11 + 11 + 9 / 22 = 20.63, presuming 2 yusho, no jun-yusho, 2 13-2 records, and a 12-3 record) or #2 if he can score a couple more zensho (363 + 100 + 30 + 9 / 22 = 22.81).

Anyway, nice topic!

[Zentoryu 154]

First and foremost, any successful or dominant wrestler does so within his generation, or perhaps a couple if they have longevity. They therefore have to repsond to their opponents. They can only be compared to their opponents because that is the only opposition they have. If a certain period of sumo (or any sport) is considered weak we cannot truly evaluate the strongest athletes from that period against others because they are a product of their environment. They may have become even better if the competition was stronger, or they may not have been as successful if challenged more often.

We need to take into account (in sumo) the relative strength of each heya because of the support various ozeki and yokozuna get by not having to fight their brothers and also in getting their brothers to beat up on other opponents.

Well done Zentoryu-zeki, and it certainly gives us a good guide. However, before we start raving about so-and-so and compare which is better etc., let's be aware of the limitations of such comparisons (In a state of confusion...)

I entirely agree with you. I would have loved to have been able to assign each rikishi points based on the quality of the opposition he faced over the course of a given basho. Alas, the opponents of most of the involved Makuuchi rikishi from 1958 to about the mid 1970's is just not known (at least not in english). So I was left with only assigning points based on a simple Yusho/Jun-Yusho/Special Prize/Wins over losses system, which I readily admit is somewhat flawed.

Hopefully someday we can do the that type of analysis when the info becomes available.

BTW, thanks for the kind words.

Edited February 26, 2004 by Zentoryu

[Zentoryu 154]

However, I'm a bit concerned about how you assigned the points. As a student of voter assignment theory, I feel like things could turn out extremely differently if you alter the amount of points that are assigned for each of these a bit, and there's noreal way of knowing what system is the most accurate. If you had it all in an excel spreadsheet and were able to do calculations quickly, I'd suggest you try changing the numbers a few different ways and then use the average rank for each person to determine a (potentially) more accurate placement system, but that's definately a bit much to ask if you're doing it all by hand (In a state of confusion...)

I actually have all of this in spreadsheets. When I said I did it "by hand", I actually meant that it was all manual entry (which I'm sure you can imagine still takes a LONG time). (Sigh...)

After taking some time off from this, I'll probably go back and fiddle with the numbers a bit. We'll see how it changes things, if at all.

Thanks to everyone for the kind words. It makes the hard work worth it.

[Kotoseiya Yuichi 3]

It makes the hard work worth it.

I know the feeling (when it comes to the forum's glossary; not updated for a long time as it is). I'm not too experienced with statistics but whenever I see one, it's interesting to try to figure something out off it. (In a state of confusion...)