Sign in to follow this  
Doitsuyama

Rikishi Strength Analysis Revisited

Recommended Posts

I certainly use my statistics as a help for making game picks, not solely by any means, and not even exclusively as anyone is invited to do the same (the Heisei ratings are freely available after all).

But even if I would use this stuff solely and exclusively - I don't care if you agree or not. I still feel it's a valid approach.

(Laughing...) (Nodding yes...) Die einzig richtige Einstellung :-)

Share this post


Link to post
Share on other sites
Well, back to topic, as a new here it's really amazing to see how many work one (Laughing...) person have been through! My area of knowledge is much related to math and numbers, so they tell me a story. I think this analysis giving the best possibility to compare rikishi of present time to those who are out of dohyo. I hope Doitsu will be back soon to answer my question. I'm really wondering where you got the idea to use the Elo kind of ranking system? Is there any argumentation why it's better than some other, or is it just so easy algoritm to use? Can you give some light ... (Nodding yes...)

Thanks.

"Easy algorithm or not" certainly never was a deciding factor as the project was big enough the way it is. If you can come up with a good suggestion for a better ranking system, I'm eagerly hearing. Sometimes simple formulas are the best.

Share this post


Link to post
Share on other sites
These numerical ratings are very similar to those in the chess world, and your work is similar in some ways to that of Jeff Sonas:

http://db.chessmetrics.com/

In any case it's a great idea. I've been a sumo fan since around 1992, diligently watching every basho as long as I could access it. I've noticed that the Kyokai doesn't have a strict or well-defined manner in which the ranking of a maegashira rikishi changes from 1 basho to the next. The only thing that a sumo fan seems to know is that, among maegashira, a rikishi's ranking rises (if there's room for him to rise of course) if he scores kachi-koshi and falls with a make-koshi result, and better scores result in a greater rise in the ranking whereas a very bad result results in a large drop in the ranking. Sekiwake and komusubi will invariably drop out of sanyaku if they win less than 7 bouts in a basho, and will at least keep their ranking if they score kachi-koshi.

1 thing that a numerical rating for rikishi could do is to allow a fairly reliable prediction of how a rikishi will do in a given basho based on his rating and the ratings of his opponents. Of course intangible factors are involved too such as injuries and other health issues. Also, psychological factors seem to play a role, as the plethora of "elevator rikishi" demonstrates. Many of the maegashira rikishi have good results when ranked near the bottom of Makunouchi but have disastrous results when ranked near the top. I suspect that this phenomena is due not just to the strength of the rikishi they face but their mental attitude as well. It's a complicated issue.

Anyway, great research and study! It is appreciated.

You found out a lot on your own, I think. All rather valid conclusions. As for the elevator phenomenon, basically a good sekiwake rating is needed to expect just 7.5 wins in the joi-jin while the same rikishi at M5w would be expected to get 10 wins and even 12 wins at the bottom of makuuchi. The typical elevator rikishi is expected less than 6 wins in the joi-jin but 10 wins at the bottom of makuuchi rating-wise so these swings really aren't a big surprise. Still, I promise to look a bit closer at Hokutoriki and Kyokushuzan and come back.

(Nodding yes...)

Share this post


Link to post
Share on other sites

Im a little dissapointed in you bro, seems to me like you plug in this formula and use it to make your sumo game picks eh? (or anyone who uses any type of formula etc.) As they should come from you, your own head and knowledge of watching individual bouts etc. not with the help of statistics, I dont agree.

I certainly use my statistics as a help for making game picks, not solely by any means, and not even exclusively as anyone is invited to do the same (the Heisei ratings are freely available after all).

But even if I would use this stuff solely and exclusively - I don't care if you agree or not. I still feel it's a valid approach.

Sure its a valid approach I suppose, but I guess I just wouldn't hold anyone who uses formulas to make their picks in as high of esteem as someone who doesn't. Also are you saying you disagree with me in my opinion that, "It requires more skill to make picks based solely off the top of your head as opposed to plugging in a formula and weighing in the odds?"

Edited by Ryukaze

Share this post


Link to post
Share on other sites
Also are you saying you disagree with me in my opinion that, "It requires more skill to make picks based solely off the top of your head as opposed to plugging in a formula and weighing in the odds?"

I just don't want to argue with you on this topic, that's all... For me it is fun to calculate ratings, yusho odds etc. and I guess you'd prefer to die than do that. I rather like to discuss this with people who like to do such things as well.

Share this post


Link to post
Share on other sites
"It requires more skill to make picks based solely off the top of your head as opposed to plugging in a formula and weighing in the odds?"
I think You're confusing "skill" and "luck". Regardless of how to "make picks", in the end they're either wrong or correct. No method on earth would change that. Any talk about "skill" is just a delusion.

Share this post


Link to post
Share on other sites

Well, back to topic, as a new here it's really amazing to see how many work one :-) person have been through! My area of knowledge is much related to math and numbers, so they tell me a story. I think this analysis giving the best possibility to compare rikishi of present time to those who are out of dohyo. I hope Doitsu will be back soon to answer my question. I'm really wondering where you got the idea to use the Elo kind of ranking system? Is there any argumentation why it's better than some other, or is it just so easy algoritm to use? Can you give some light ... (Nodding yes...)

Thanks.

"Easy algorithm or not" certainly never was a deciding factor as the project was big enough the way it is. If you can come up with a good suggestion for a better ranking system, I'm eagerly hearing. Sometimes simple formulas are the best.

Yes, simple is fine, usually working quite well. My point actually was, that you have done a perfect job and the ratings are giving good basis for rikishi comparison. (Laughing...) I'm familiar with Elo and the origins of Elo from chess. I played chess myself in my younger years. I don't want to argue to select some other system. There are, of course, others, like tennis players ranking system, but I am not aware if there is better one. Therefore the question is how you yourself found the Elo rankings and how you did come to the idea to use it for Ozumo rikishi comparison. Ozumo and chess are not so similar (Help me...) Did you compare Elo with some other systems or just picked it up?

"easy algorithm" I thought is that it is simple for codification (just figures in formula) for any commonly used db software. Yes, the work of gathering the data to use in the formula is much more serious task, so thus my respect ...

Share this post


Link to post
Share on other sites
"It requires more skill to make picks based solely off the top of your head as opposed to plugging in a formula and weighing in the odds?"
I think You're confusing "skill" and "luck". Regardless of how to "make picks", in the end they're either wrong or correct. No method on earth would change that. Any talk about "skill" is just a delusion.

A statement with alot of philospohical potential there... Is the guy who picks the winner of a bout due to his own "knowledge," having watched both rikishi compete exibited skill in using this to determine his correct pick? Or is he just lucky that the guy he picked happened to win regardless of the "odds?"

As for my playa Doitsuyama, yessss bro I would most definately rather die than make a sumo game pick based solely on a statistic any dayyay! You still didn't answer my question tho.......

Edited by Ryukaze

Share this post


Link to post
Share on other sites

Fantastic job! Going to need more than the 3 hours I spent with it, tip of the iceberg. Fantastic analysis, thorough and engaging. Thank you so much! :D Made a newbie Sumo fan a point of reference.

-Gogo (In a state of confusion...)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this