Rating the top guys

[Guest Vincentoryu]

RIJICHO ADDITION: The table below in this post is to be found monospaced and slightly modified later in this thread.

---

I tried to find a way to rate the rikishi in Makuuchi.

I used an approach in which the last 6 basho are rated.

The last basho's results are multiplied by 6.

The second last basho's results are multiplied by 5.

...

The fifth last basho's results are multiplied by 1.

So I get an weighted average win score for every rikishi.

Only rikishi who have had at least the rating of Maegashira 5 or higher over the last twelve months were taken in consideration.

This gives me this ranking:

• Unofficial Ozumo Ranking (Year round)
  • Shikona Current Rank Hatsu Basho 2004 Kyushu Basho 2003 Aki Basho 2003 Nagoya Basho 2003 Natsu Basho 2003 Haru Basho 2003 Total Wins Total Losses Basho Average Points Weighted Basho Av.
    
  • 1 Asashoryu HY 15 0 12 3 13 2 5 5 13 2 10 5 68 17 12,0 253 12,7
    
  • 2 Chiyotaikai HO 10 5 10 5 11 4 11 4 10 5 12 3 64 26 10,7 219 10,4
    
  • 3 Kaio NO 10 5 10 5 7 8 12 3 11 4 10 5 60 30 10,0 206 9,8
    
  • 4 Tochiazuma HO2 9 6 13 2 10 5 7 8 8 7 0 0 47 28 9,4 196 9,8
    
  • 5 Wakanosato HS 9 6 7 8 11 4 10 5 9 6 9 6 55 35 9,2 190 9,0
    
  • 6 Musoyama NO2 5 4 9 6 1 5 10 5 8 7 1 6 34 33 7,6 126 8,0
    
  • 7 Kyokutenho NM2 8 7 4 11 10 5 6 9 10 5 9 6 47 43 7,8 155 7,4
    
  • 8 Tochinonada HM3 6 9 8 7 8 7 7 8 8 7 9 6 46 44 7,7 154 7,3
    
  • 9 Tosanoumi NM4 4 11 10 5 7 8 10 5 4 11 8 7 43 47 7,2 148 7,0
    
  • 10 Takamisakari HM8 4 11 5 10 9 6 9 6 6 9 8 7 41 49 6,8 132 6,3
    

The last digit is the one I use to rank the guys.

You can see that there is a huge gap between Asashoryu (averaging 12.7 wins a basho!) and the Ozeki trio of Chiyotaikai, Kaio and Tochiazuma of at least 2.3 wins a basho!

It also shows that Wakanosato is performing at a typical permanent Sekiwake level of 9 wins a basho.

Musoyama is clearly in danger of losing his rank as is stressed by his 8.0 weighted win average over the last 12 months.

Kyokutenho, Tochinonada and Tosanoumi are the typical high maegashira - low sanyaku rikishi with there winnumbers.

Takamisakari will be out next time the ranking will be published as he is demoted to Maegashira 8E.

Please excuse me for the way these numbers are presented as I don't really know how I can improve them.

Kind regards

Vincentoryu

Edited March 6, 2004 by Rijicho

[Yoavoshimaru 0]

Umm, ok... Interesting I guess... Why did you do this?

[Dejimasama 0]

ery interesting. Thanks for sharing this. Its always neat to see how others come up with their own rankings.

[Guest Vincentoryu]

I made this ranking in an attempt to see which rikishi are performing the best over the last couple of basho (a one-year span). So it's sort of a rolling Year ranking of rikishi.

Also I tried to quantify the difference in performance of the top dogs. So that I am sort of able to say that rikishi a is averaging x wins more than rikishi b.

Also I think it is a bit clearer when you use the typical sumo interval of 0 to 15 wins and not a winningpercentage as you normally see in sumostatisctics.

[Yoavoshimaru 0]

Cool, thanks for the stats and explanation. Do you keep a running spreadsheet with this data for use in sumo games?

[Kotoseiya Yuichi 3]

# Shikona       CR  01/04 11/03 09/03 07/03 05/03 03/03 TW TL AWB  PW  PWBA
1 Asashoryu     HY  15-0  12-3  13-2   5-5  13-2  10-5  68 17 12,0 253 12,7
2 Chiyotaikai   HO  10-5  10-5  11-4  11-4  10-5  12-3  64 26 10,7 219 10,4
3 Kaio          NO  10-5  10-5   7-8  12-3  11-4  10-5  60 30 10,0 206 9,8
4 Tochiazuma    HO2  9-6  13-2  10-5   7-8   8-7   0-0  47 28 9,4  196 9,8
5 Wakanosato    HS   9-6   7-8  11-4  10-5   9-6   9-6  55 35 9,2  190 9,0
6 Musoyama      NO2  5-4   9-6   1-5  10-5   8-7   1-6  34 33 7,6  126 8,0
7 Kyokutenho    NM2  8-7   4-11 10-5   6-9  10-5   9-6  47 43 7,8  155 7,4
8 Tochinonada   HM3  6-9   8-7   8-7   7-8   8-7   9-6  46 44 7,7  154 7,3
9 Tosanoumi     NM4  4-11 10-5   7-8  10-5   4-11  8-7  43 47 7,2  148 7,0
10 Takamisakari HM8  4-11  5-10  9-6   9-6   6-9   8-7  41 49 6,8  132 6,3
CR - Current rank, TW - Total wins, TL - Total losses,
AWB - Average wins in basho, PW - Points weighted,
PWBA - Basho average with weighted points

[aderechelsea 125]

(Applauding...)

Koto-san you gave vicentoryu's fine job a glamour that it was missing and it makes a lot more sense now............

Good job from both of you...........

These stats are clearly indicating the dominance of Asa in the last year and are mirroring the truth up to a point.

[MongolPower 0]

PW - Points weighted,

PWBA - Basho average with weighted points

Can you explain the purpose of these?

Edit: How are these calculated?

For Asa--15*6+12*5+13*4+5*3+13*2+10*1=253

253/(6+5+4+3+2+1)=12.05

253/(6+5+4+3+2)=12.65

Edited March 6, 2004 by MongolPower

[Takanobaka 0]

Thanks for the comparison!

To a certain extent, I'm not sure that the limit of staying at all times above M5 is fair (for example, this would have cut out Asashoryu from consideration at the beginning of his astonishing Yokozuna run, along with anyone else who has dominated while on the rise like Kakizoe or Kokkai), but once again I love seeing all of this stuff. On the other hand, if you didn't such a limit, the result could end up (theoretically) with a guy mid-maegashira being better than the Yokozuna!

Some day when I have the patience I might try calculating a "dominance factor" which determines, from top to bottom, who has really been most dominant against who they've been facing...

[Naganoyama 6,035]

Hi Vincentoryu. Thanks for this rating system which produces interesting and believable results. (Applauding...)

Perhaps you could extend it down the banzuke by adding a 'banzuke position factor'. So a rikishi higher up the banzuke would have his results slightly amplified to account for his stronger opposition and a rikishi lower down the banzuke would have his results slightly discounted.

Perhaps the 'bpf' for a rikishi at M6 (about half way down the banzuke) should be 1.0 and for each position upwards/downwards you should add/subtract a small amount (say 0.1)

Actually this could be simplified a bit since rikishi above this level face much the same opposition.

What do you think?

[Guest Vincentoryu]

Can you explain the purpose of these?

Edit: How are these calculated?

For Asa--15*6+12*5+13*4+5*3+13*2+10*1=253

253/(6+5+4+3+2+1)=12.05

253/(6+5+4+3+2)=12.65

I only use the bouts I which the rikishi did get a win or a loss so kyujo bouts, other than the first which results automatically in a fusen loss, are not taken into consideration.

Perhaps you could extend it down the banzuke by adding a 'banzuke position factor'. So a rikishi higher up the banzuke would have his results slightly amplified to account for his stronger opposition and a rikishi lower down the banzuke would have his results slightly discounted.

Perhaps the 'bpf' for a rikishi at M6 (about half way down the banzuke) should be 1.0 and for each position upwards/downwards you should add/subtract a small amount (say 0.1)

Actually this could be simplified a bit since rikishi above this level face much the same opposition.

What do you think?

That would be a very good idea but that requires a lot of data collection.

I think I first have to do a historical analysis of the strenght of the schedule for each banzuke position so that I can use this as the basis when determing the 'banzuke position factor.'

I think I will try this somewhere in the future!

Edited March 8, 2004 by Vincentoryu