Handicapping The Haru Banzuke – Part 2


The Meat Grinder & Cannon Fodder

In the first of our series speculating on the Haru banzuke, we took at look at the San’yaku ranks, which face fierce competition for the competitive ranks, and significant injury and problems in the “permanent” ranks.

Today we look at the rikishi who have hard work to earn their pay, the upper half of Maegashira ranks.  As with the first group, this is all purely speculative, and based on some formula concocted by myself in an attempt to guess where the Nippo Sumo Kyokai will rank the men for the March tournament in Osaka.

Plugging everyone’s win/loss record, the difficulty of their foes, a scoring factor for their rank (harder to move up the higher you were in the banzuke) and a few other magic elements, we get this prospective ranking:

East Rank West
Ikioi Maegashira 1 Takekaze
Takarafuji Maegashira 2 Sokokurai
Tochinoshin Maegashira 3 Shohozan
Arawashi Maegashira 4 Takanoiwa
Yoshikaze Maegashira 5 Endo
Ichinojo Maegashira 6 Chiyonokuni
Hokutofuji Maegashira 7 Aoiyama

Ikioi is an amazingly popular rikishi with the public, and his posting to Maegashira 1e for the Osaka basho will only ramp public interest higher, above and beyond the current Kisenosato mania sweeping Japan.  Ikioi’s sumo has been improving steadily, and the NSK probably assume it’s time to give him a test for a San’yaku slot in the near future. Interestingly enough, the sumotori with the highest “mathematical” rank is Sokokurai! Sokokurai had 11 wins, his “rank velocity” (win vs loss * rank factor * schedule difficulty) was an astounding 9.9, higher than anyone and well ahead of second highest Ichinojo. But I think the Maegashira 1 ranks are prized positions, and the Nippon Sumo Kyokai will likely put Ikioi’s impact on the popularity of sumo foremost. Takekaze moves up from Maegashira 5 to a Maegashira 1 spot at Haru, and we will see if the veteran can fend off the up-and-coming crowd.

Takarafuji benefits from the chaos and blood bath at the upper end of Makuuchi in January, landing solidly at Maegashira 2e, and a chance to rack up kinboshi against a wounded Yokozuna crew.  Joining him is Sokokurai at Maegashira 2w, the rikishi who computed out to the highest “rank velocity” of anyone coming out of Hatsu. If he continues his strong streak from Hatsu, he will present a really good opponent to many top rikishi.

Tochinoshin drops 3 slots from Komusubi to Maegashira 3, he had a terrible record before his injuries forced him to withdraw. To be honest, it will be interesting to see if he is even healed up enough to compete, but as always Tachiai wishes the big Georgain the best of fortune. Joining him at Maegashira 3w is Yokozuna Kisenosato‘s dew-sweeper, Shohozan.

Takanoiwa,  who has been really blowing the doors off of his competition, raises from Maegashira 10 to Maegashira 4. I put him on the west side, which draws a slightly easier schedule.  However, if there is a lack of fierce Ozeki class competition, we may once again see score inflation among the up-and-coming rikishi, and I would look for Takanoiwa to excel. Joining him is Arawashi falling from Maegashira 2 in January.

Yoshikaze (a favorite of mine) seems to have been ranked in a very comfortable spot, as the computation put him at the same rank, but moved him from West to East. Joining him at Maegashira 5 is fan favorite Endo, whose make-koshi in January pushed him down from M2.

Another of the levitating next-gen rikishi, Ichinojo, leaps to Maegashira 6 from his prior spot at Maegashira 13.  Frankly, I am not sure if he is ready for this intensity of competition, but we will see in March how he fares. His computed “rank velocity” was an impressive 7.7, which was more than Takanoiwa.  Joining him is Chiyonokuni, who turned in a solid performance in January at Maegashira 8.

Rounding out the upper portion of the Maegashira ranks, we the rather impressive Hokutofuji at Maegashira 7e, with man-mountain Aoiyama broadly occupying the Maegashira 7w position.  This is one of the cases where even though Aoiyama was able to turn in a winning record (8-7), there was a huge cohort with strong winning records, with victories over higher ranked rikishi, and they ended up passing him by.

Keep in mind – this exercise is for discussion purposes for the most part. So feel free to leave comments, and alternate opinions. I am hoping to tune my formulas over time, and this first attempt should not be taken too seriously.

Tune in Friday for part 3!

24 thoughts on “Handicapping The Haru Banzuke – Part 2

  1. My simple formula based on previous rank and won-loss record produces the following predicton (I’ve left out Tochinoshin because of the withdrawal, but my most optimistic position for him would be M6, and possibly much lower).

    M1 Takekaze Sokokurai
    M2 Ikioi Takanoiwa
    M3 Shohozan Yoshikaze
    M4 Takarafuji Arawashi
    M5 Endo Hokutofuji
    M6 Chiyonokuni Aoyiama
    M7 Ichinojo Chiyoshoma

    The ranks for Takanoiwa and Shohozan are equal, but Shohozan got make koshi at M2. Also equivalent ranks and could be switched: Yoshikaze and Takarafuji; Arawashi, Endo, Hokutofuji and Chiyonokuni; Aoyiama and Ichinojo.

    • Nice list. I am going to be really curious to see how this one turns out. I am going to post the lower half of Makuuchi on Friday, so please – if you don’t mind – feel free to post your picks for those ranks, too.

      I would love to see Yoshikaze at M3, but I get the feeling they are not going to give him that slot.

  2. Happy to post the rest of my picks as a comment on Friday’s post. The main big uncertainties for me are what they’ll do with Tochinoshin, and how high they’ll place Daieisho and Ura from Juryo. By rank alone I’d have those two at M8 and M10, but history suggests they’ll be a lot lower.

    • If one judges by all the pink key chains on sale…yes. When I was in Nagoya, Satoyama (aka “Sexy Sugar Mountain) seemed to have the most vocal female cheering section.

    • Well, yes. But here’s the odd thing – it’s not so much the girls, but the little old ladies can’t get enough Endo or Ikioi. It’s a crazy ass thing, but it happens!

      • Rikishi are almost naked, i could imagine women having fun watching them. Specially competing in battle of strength.

        Yeah ikioi is good looking, but endo is more handsome.

  3. I hope Yoshikaze learns how to keep his face out of the mayhem, my wife and I have been calling him “The Cyclops” last 2 basho because he seems to get smashed around his right eye for some reason.

    • He’s gotten more than one battering there around the nose and eyes. I agree, it seems to be a constant target. Maybe he has a reputation as a bleeder?

    • My wife adores Yoshikaze, but gets so worried about his face. I think Andy is right, he has a reputation as a bleeder, so you see many opponents go for his face. If you take a look at photos of him 5 years ago, he had a normal looking face, but now he seems to be perpetually swollen.

      I can only hope that once he retires, things go back to normal.

  4. Oh, how I do love new arbitrary metrics. (Nerd, here.) I’m going to assume you’ve identified and scaled these variables in an attempt replicate the esoteric math of previous banzuke. How well does your current model “predict” those previous outcomes?

    • Oh yeah – backcasting! So yes, I did a good amount of back-casting with the formula, and they got me to 80% in most cases, which I figured was good enough of spitballing a banzuke on a sumo blog. it’s clear that the basic banzuke is math driven, and then undergoes some measure of hand adjustment based on NSK driven factors. Please not I am not criticizing this at all.

      You don’t seem them moving anyone up or down more than two slots, except in the case of Osunaarashi’s brutal demotion in 2016, which defies all math. But I suspect that my formula may be too simple to get much closer than my 80%. Here’s how I compute it.

      Firstly, you get a win vs loss record. so you can get things like +5 or -3 depending on if you had more wins or more losses. Then you weight the wins or losses based on relative strength of weakness of opponents. For example. if you a Maegashira 5 with a 7-8 record, but you beat 2 Yokozuna, that is better than 7-8 and losing to 2 Maegashira 10. Then you take into that moving from M15 to M12 is less difficult than M3 ro M1, or M2 to Komusubi. You roll all of that together, and I compute what I call “rank velocity” which you can think of as a first order derivate of their record during the prior basho.

      Each rank below Ozeki has a numerical value. Rikishi current rank value x rank velocity and you get “computed rank”. Take the computed ranks and short them in ascending order, and put them into the banzuke. Thats where the math ends.

      Then (if you are me) you make sure that you consider velocity, the clout of each heya, and public support for some rikishi. For example on this one – Sokokurai ended up with the highest computed rank. But I put Ikioi at Maegashira 1e. Ikioi computed out at M2w, but his popularity is such that he will likely get a prominent role on the banzuke.

      Just for discussion, below is the resulting data for velocity and computed rank, sorted by computed rank

      Computed Rikishi
      -0.5 Mitakeumi
      -0.31 Takayasu
      0.25 Tamawashi
      1.25 Shodai
      2.1 Sokokurai
      3.25 Takekaze
      3.98 Tochinoshin
      4.4 Ikioi
      4.5 Takarafuji
      4.5 Shohozan
      5.5 Arawashi
      5.7 Takanoiwa
      6.25 Yoshikaze
      6.6 Endo
      7.3 Ichinojo
      7.6 Hokutofuji
      7.6 Chiyonokuni
      8.2 Aoiyama
      8.75 Chiyoshoma
      9.2 Okinoumi
      10.1 Kaisei
      10.25 Kotoyuki
      11.4 Tochiozan
      12 Kagayaki
      13.7 Ishiura
      14.6 Myogiryu
      15 Takakeisho
      15 Daishomaru
      15.5 Sadanoumi
      18 Nishikigi
      18.5 Chiyoo
      19.3 Chiyootori
      19.3 Chiyotairyu
      20.5 Gagamaru
      28.5 Osunaarashi

      Sorted by rank velocity

      Velocity Rikishi
      9.9 Sokokurai
      7.7 Ichinojo
      6.3 Takanoiwa
      3.75 Takekaze
      3.5 Mitakeumi
      2.4 Hokutofuji
      2.4 Chiyonokuni
      2.31 Takayasu
      1.5 Sadanoumi
      1 Kagayaki
      0.9 Kaisei
      0.8 Aoiyama
      0.75 Tamawashi
      0.75 Yoshikaze
      0.6 Ikioi
      -0.25 Shodai
      -0.5 Shohozan
      -0.6 Endo
      -0.75 Chiyoshoma
      -1 Takakeisho
      -1 Daishomaru
      -1.5 Takarafuji
      -1.5 Arawashi
      -1.5 Chiyoo
      -1.98 Tochinoshin
      -2.25 Kotoyuki
      -2.7 Ishiura
      -3.3 Chiyootori
      -3.3 Chiyotairyu
      -4.2 Okinoumi
      -5 Nishikigi
      -5.4 Tochiozan
      -5.5 Gagamaru
      -5.6 Myogiryu
      -10.5 Osunaarashi

      • Nice. I need to make some time to play with the historical data to devise my own approach, but yours is an interesting one. I suspect the actual process may be a bit more algorithmic than computational, but if the model fits…?

  5. Hi, my 5 cents worth…

    Takekaze 1 Sokokurai
    Ikioi 2 Takanoiwa
    Shohozan 3 Takarafuji
    Yoshikaze 4 Hokutofuji
    Chiyonokuni 5 Endo

    probably a good thing for Yoshi and Endo NOT to be ranked 1-3 as they might avoid the sanyaku fights (at least in the beginning of the basho).

    Sokokurai has never been up so high and I wonder if he will get out alive (means 4 wins at least…)
    for Takekaze I predict his last hurray and a 5-10 or worse… I mean, the guy is sixtyseven years old or so.

    Ikioi could be up for a surprise, Osaka is his homebasho and he’s getting stronger physically constantly.

    If Takarafuji is back to top form, he’s my best pick for kk among these 10 guys, maybe a kinboshi candidate.

    I predict only 3 kks among those ten guys, gotta be a tough basho for the Jo-Jin :)

    • You and LK did very well projecting your banzuke. Bruce crushed me, even with Tochinoshin out of place. Y’all downright schooled me.

  6. Much of the upper Maegashira blood-letting comes from the Ozeki and Yokozuna during the first week of any basho. Both Andy and I have genuine concerns over the health and viability of the Ozeki for March, and at least 2 Yokozuna in a questionable state.

    During a normal basho, I would say you are dead on right about 3 KKs for the whole group you listed. But March may be a complete nut-house in terms of what takes place.

    My biggest wish is that Goeido and Harumafuji show up healthy and able to compete at 80% of their capability as a minimum. That would be Goeido 2.0, rather than the kadoban compatible 1.0 or 1.1 version. If we can get that, all kinds of carnage in the upper Maegashira. But I fear that it might be the other way around – the upper Maegashira dominate these weekend upper ranks.

  7. IMHO Tochinoshin will collapse much more than you think.
    A 3-12 makekoshi = 8 ranks down in the banzuke. Tochinoshin has 0-6-9 !
    Remember Aminishiki who lost 10 ranks after 2016 Nagoya basho with a score of 1-2-12…

    • I was in fact probably too light on his demotion. I do wonder if he can come back healed and ready, or will he limp through another basho.


This site uses Akismet to reduce spam. Learn how your comment data is processed.