Sign in to follow this  
Odoriou

Ozumo World Champion 2005 is...

Recommended Posts

I made up a points system which takes only this year's achievements of makuuchi rikishi into consideration. I awarded 10 championship points for a makuuchi yusho, 6 for a jun-yusho, 4 points for a 3rd best basho result etc. (10-6-4-3-2-1). One bonus point was awarded for a sansho.

So this is my attempt to make an objective list of this year's most successful rikishi:

01.	Asashoryu	60
02.	Kotooshu	 23+4
03.	Tochiazuma   17
04.	Kotomitsuki  12+1
05.	Hakuho	   12+1
06.	Chiyotaikai  12
07.	Tamanoshima  11+1
08.	Kaio		 10
09.	Kyokutenho	9
10.	Roho		  9
11.	Kyokushuzan   8+1
12.	Kokkai		8+1
13.	Futeno		7+2
14.	Miyabiyama	7+1
15.	Ama		   6+1
16.	Tokitenku	 6+1
17.	Tochinohana   6+1
18.	Jumonji	   6
19.	Asasekiryu	6
20.	Aminishiki	6
21.	Takamisakari  6
22.	Kaiho		 5+1
23.	Kisenosato	5+1
24.	Iwakiyama	 5
25.	Wakanosato	5
26.	Kasugao	   5
27.	Hakurozan	 5
28.	Takekaze	  5
29.	Dejima		4
30.	Tosanoumi	 4
31.	Buyuzan	   3
32.	Kakizoe	   3
33.	Toyonoshima   3
34.	Kasuganishiki 3
35.	Toyozakura	3
36.	Hokutoriki	2
37.	Kotoshogiku   2
38.	Shunketsu	 2
39.	Tochinonada   2
40.	Takanowaka	2
41.	Shimotori	 2
42.	Tamaasuka	 2
43.	Hayateumi	 2
44.	Kotonowaka	1
45.	Katayama	  1
46.	Tokitsuumi	1
47.	Tamakasuga	1

OzumoWorldChampion2005.gif

It's certainly no surprise that Asashoryu could secure the title "Ozumo World Champion 2005" while Kotooshu was runner-up. But maybe there are some surprises elsewhere on this table. For instance, I didn't expect Kotonowaka as low as on position 44 - he scored only one championship point in haru basho.

This way to measure the success of rikishi probably doesn't make much sense, but it's non-basho time again and I was bored. (In jonokuchi...)

Edited by Odoriou

Share this post


Link to post
Share on other sites
OzumoWorldChampion2005.gif

Yeah yeah yeah. I've been dreaming of seeing charts like this for sumo. Specifically I'd like to see the banzuke done up like this for the past year, 5 years etc... but I can imagine it would go to spaghetti quite quickly.

What I've been wondering is exactly how wide the gap is betwen Asa and his Ozeki. I'm a math major so I'll say it this way: If the average skill difference between two rikishi side-by-side on the banzuke is represented by X, by how many X is Asashoryu better than the number 2 Rikishi ( I guess you could pretend it is Kotooshu right now ) ?

I was trying to figure out how you'd calculate this, maybe using his 84-6 record as a guide stick. If 45-45 over the year as Yokozuna meens about a Sekiwake level ( -7.5 X ) and 90-0 puts him 15 'X' above everyone else then his 84-6 translates to ... 12 X? Have I lost everyone by this point?

So he's roughly 12 spots better than the Ozeki? Is the difference between Tochiazuma and Dejima roughly equal to the difference between Asashoryu and Tochiazuma?

Share this post


Link to post
Share on other sites

Wow, a math major! I'm in awe. Too bad that the rest is lacking a lot in theoretical strength.

Meanwhile you may have a look at some strength ratings calculated from some guy and draw your own conclusions to answer your initial question. I don't think my math is strong enough to help you there. (In jonokuchi...)

Share this post


Link to post
Share on other sites
Wow, a math major! I'm in awe. Too bad that the rest is lacking a lot in theoretical strength.

Well, I know enough to say that I'm not being scientific, that's for sure.

Alright so the 84-6 roughly translates to 12 spots above the Ozeki by my earlier calculations but is obviously a pretty hacky way to do it.

Taking a look at your strength charts I came up with the following number: 8.43 . I averaged the 'power' difference between the 2nd and 6th ranked guys each tournament and came up with 23. So my 'X' from the first post is now 23 in this scheme. Since Asa was on average 194 'power' above the 2nd ranked guy it works out to 194/23 = 8.43 which means he's a bit more than seven ranks better than his current Yokozuna rank.

Relativistic Banzuke as might have been seen before Kyushu ( with stats based on Kyushu ) :

YE: Asashoryu YW: -

YE: - YW: -

YE: - YW: -

YE: - YW: -

OE: Tochiazuma OW: Chiyotaikai

OE: Kaio

SE: Kotooshu SW: Kotomitsuki

That's about where I pictured him in my head so I'm glad the numbers worked out about the same.

Edited by Tetsuzukiyama

Share this post


Link to post
Share on other sites

Well, 23 is a pretty high number. If you go down only slightly more, say, to the tenth instead of sixth, this number should go down a lot as the difference between sixth and tenth usually is a lot smaller than between second and sixth. So I'm still not very satisfied with your method as it is anything but robust (long ago when I was young I used to be assistant professor at university so I have seen a lot there...). Come on, delve a bit deeper into the numbers (you can click on each basho by the way to get a lot more info for each basho) and come up with an approach which can satisfy my critical eyes. (Sign of approval)

Share this post


Link to post
Share on other sites
Come on, delve a bit deeper into the numbers (you can click on each basho by the way to get a lot more info for each basho) and come up with an approach which can satisfy my critical eyes.

Hmmph. I kind of liked how high he was on the scale... this can only... wait a second? Lower difference? So his Relative rank should be higher? Maybe I really will have to do the work, now.

I wonder if the 84-6 calculation of 12 is closer to the mark after all. If the difference do consistantly shrink the lower you go on the banzuke then we've got a curve and some equations to work out if this is to be done in a robust way. Actually I'm getting the feeling that Statistics would be more useful here as we're dealing with the extreme end of the Sumo bell curve...

I probably won't go that far but I'll spent a bit more time on this tonight. You've certainly thrown the gauntlet down now...

Share this post


Link to post
Share on other sites

For some reason I didn't expect Hakuho to get only 12+1 points. The way you read about it, you could expect him and Tochiazuma to change places - sort of - in a way.

Edited by Phelix

Share this post


Link to post
Share on other sites
Come on, delve a bit deeper into the numbers

It took me a few days but I went a bit deeper into the numbers on your page. I do like the formula you use to represent their relative abilities and it made me curious to know how often your Sumo predictions go against the numbers or if you ignore the values completely when picking.

I'm pretty lowly on the Japan-Guide sumo game and I always figured that the top guys would be using the power of databases to analyze the winning percentages. If you had a robot player that always predicted matches based on the biggest power-differences, I wonder how he'd do.

Anyway, my numbers: So I had come up with a magic number of 23 to represent the average difference between Rikishi but as Doitsu pointed out that was only sampling the very top of the Banzuke where the numbers spread out a bit. Here are the other averages over 2005:

Average Difference:

From 2 to 16 - 15.43

From 2 to 4 - 23.1 *used last time

From 2 to 5 - 21.7

From 6 to 10 - 16

From 11 to 15 - 7

Using 15.43 as our new X puts Asashoryu 12.6 ranks above the number 2 rikishi. Unfortunately the number keeps getting bigger and so each 'bonus' rank should be harder and harder for him to earn, and for that we need the quadratic equation.

Quadratic Equation for the difference between ranks in '05 : 26.84 - 1.17x - 0.0162x^2

( x is banzuke position )

So here are the plotted differences for each Rank and the running subtraction from Asa's 194 point perch:

Rank 1: 25.6 ( 194 - 25.6 = 168.4 )

Rank 0: 26.84 ( 168.4 - 26.84 = 141.56 )

Rank -1: 28 ( 141.56 - 28 = 113.56 )

Rank -2: 29.1 ( 113.56 - 29.1 = 84.46 )

Rank -3: 30.2 ( 84.46 - 30.2 = 54.26 )

Rank -4: 31.26 ( 54.26 - 31.26 = 23 )

Rank -5: 32.3

So by this calculation Asa should be about 5 ranks higher than he is now, or:

YE: Asashoryu YW: -

YE: - YW: -

YE: - YW: -

OE: Tochiazuma OW: Chiyotaikai

OE: Kaio

SE: Kotooshu SW: Kotomitsuki

The last time when I only used 23 as the value of X and he was two ranks higher. If I were to improve this analysis I would have plotted the quadratic equation of the entire Makuuchi but that would only change things by a rank or two, I postulate downwards. Do you still demand satisfaction?

Interesting that there's often a break-away pack of four guys each tournament. There is almost always a sharp drop-off between the 4th and 5th scores after each basho.

And thanks for forcing my eyes across your stats page, it is huge.

Share this post


Link to post
Share on other sites
I'm pretty lowly on the Japan-Guide sumo game and I always figured that the top guys would be using the power of databases to analyze the winning percentages. If you had a robot player that always predicted matches based on the biggest power-differences, I wonder how he'd do.

Don't get fooled. I'm probably the only one. Ok, Igiski too I think, going by his picks. I use my power-differences, yes. But I'm adjusting them with injury updates, age developments, actual form and personal observation - plumbing it in manually for a change.

Share this post


Link to post
Share on other sites
Do you still demand satisfaction?

No, your result looks good enough - and I'm essentially result-oriented. (Sign of approval)

Still, you could have undergone some tests at first to see which is the best curve to approximate the ranks from 1 to 42, and after that calculate the parameters. The linear calculations don't lead anywhere.

And nice to hear anyone likes these pages. (I am not worthy...)

Share this post


Link to post
Share on other sites
Still, you could have undergone some tests at first to see which is the best curve to approximate the ranks from 1 to 42, and after that calculate the parameters. The linear calculations don't lead anywhere.

I agree, I was hoping to see the curve smooth out more through 2-16 but it was obvious that the the entire Makuuchi would be needed to validate the results. At least I can still do quadratic curve-fitting.

One strange thing that popped up ( in the linear bumbling ) was that it would be something like 110 ranks before his difference of 194 would fit the curve... I mean the power differences start near 0 and rises to 23.1 at the top Ozeki and a difference of 194 is only normal for a -110 ranked Yokozuna... We know he's good, I know, I know.

Well I feel better now that I have a firm grip of the 'rank gap' he's holding right now.

I might do a Robo-Sumo test next basho and see how your naked power ratings do in the Sumo Game. Can't be worse then my 'guys I like' strategy.

Share this post


Link to post
Share on other sites
Yeah yeah yeah. I've been dreaming of seeing charts like this for sumo. Specifically I'd like to see the banzuke done up like this for the past year, 5 years etc... but I can imagine it would go to spaghetti quite quickly.

That's quite nice, as long as you only look at a few rikishi at once:

Banzuke-sel.gif

If you want to see some more of them, you really get a plate full of spaghetti:

Banzuke-all.gif

Hmmm... that 2nd chart looks appetizing. (Eating...)

Share this post


Link to post
Share on other sites

I'm pretty lowly on the Japan-Guide sumo game and I always figured that the top guys would be using the power of databases to analyze the winning percentages. If you had a robot player that always predicted matches based on the biggest power-differences, I wonder how he'd do.

Don't get fooled. I'm probably the only one. Ok, Igiski too I think, going by his picks. I use my power-differences, yes. But I'm adjusting them with injury updates, age developments, actual form and personal observation - plumbing it in manually for a change.

You think that's why my picks are so similar to yours? :)

Yeah, I do about 90% of my picks straight from my ratings. Though it's more helpful in some games than others.

Igiski

Share this post


Link to post
Share on other sites

I don't get it. Takatoriki won the yushu in Haru, 2000

http://www.banzuke.com/~movies/haru2000/da...i_interview.wmv

yet he is nowhere to be seen in the top six ratings for that tournament

http://130.149.112.57/ratings/index.htm

was his level of opposition That low?

At any rate, for me that was one of the great "moments" in sport.......Easily!!!

Edited by the italian halibut

Share this post


Link to post
Share on other sites
I don't get it. Takatoriki won the yushu in Haru, 2000

http://www.banzuke.com/~movies/haru2000/da...i_interview.wmv

yet he is nowhere to be seen in the top six ratings for that tournament

http://130.149.112.57/ratings/index.htm

was his level of opposition That low?

They're not the ratings for that tournament alone, but for a rikishi's entire career up to that tournament. What you're looking for is a rikishi's change in rating for a given basho - Takatoriki indeed comes out on top there for Haru 2000, with +133 points. No other Makuuchi rikishi had more than +47.

Edited by Asashosakari

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