What Makes A Sekiwake
Using Numbers and Precedent to Determine the Number of Sekiwake Slots
Last week’s Banzuke prediction piece I joked at one point that I have a lot of research to do before I’m trying to determine how many Sekiwake there ought to be in each tournament. Well, the comments on that post were lively and as a result, I’m looking to answer various questions about Sekiwake so our Banzuke prediction can be improved.
Let’s start with some basic stats on the Sekiwake (is the plural Sekiwakes? I guess in Japanese it’d be Sekiwake-tachi).
First off we can see that 4 out of 5 tournaments there will only be 2 Sekiwake. That’s a good start. Furthermore, we can also observe that having 4 Sekiwake is quite rare. 2024 we’ve only had 2 Sekiwake per tournament, but 2023 was quite the year as there were 3 or more Sekiwake in every basho starting from September 2022, and even 4 Sekiwake in January 2023 and May 2023.
So with just this data, we have the data to say that 2 Sekiwake will be the usual baseline. But what if there could be more than 2 Sekiwake? There are situations where an Ozeki is demoted but there were 2 Sekiwake who both had Kachikoshi in the preceding tournament, for instance.
So next what I looked at was what happened preceding each tournament where we went from 2 to 3 Sekiwake. That’s 19 tournaments. A (not so fun I suppose) fact I learned during this research was there was actually no Yokozuna from July 1992 through January 1993. Hokutoumi (yes, the current Japan Sumo Chairman) retired and was last on the Banzuke in May 1992, and Akebono was promoted and first served as Yokozuna in March 1993. Very interesting and I suppose that slightly raises my odds for Terunofuji retiring with no new Yokozuna in place, albeit not too much as I think 2024 and the early 90’s not long after the Futahaguro scandal are quite different.
Looking through all of those instances, and the 6 times that there was a fourth slot opened (so going from 2 or 3 to 4 Sekiwake), I was able to establish some basic rules. Rules is a bit of a strong word as there’s a degree of inconsistency, but by the time I had gone through these 25 instances of a new slot being opened up I had a pretty good idea of who would be at Sekiwake and who wouldn’t. These rules are ordered in terms of their priority, so 1 will be most important, and then if 2 applies, it’ll be used next.
Current Sekiwake who went 8-7 (or better) receive absolute precedence and retain their slots
Demoted Ozeki receive the lowest Sekiwake slot, potentially opening one up
This is where it gets dicier - I looked at the tournaments where Ozeki had a Makekoshi and were demoted, and then the next tournament where they’re at Sekiwake and it’s a bit inconsistent on if they’ll open up a new Sekiwake slot or not. It’ll depend on the subsequent rules and how the Sekiwake did in the previous tournament.
A 10-5 (or better) record is required to open up a new Sekiwake slot
Newly opened slots will be ranked lower than existing slots excluding demoted Ozeki
Generally, higher ranked Komosubi are favored for opening up new slots
However it’s a lot harder to predict how they’ll handle multiple Komosubi with 10-5+ records
They also will potentially open a slot for a high performing upper ranked Maegashira (M1 or M2)
They are generally averse to opening up 2 new slots in the same tournament
It is not unprecedented however occurring in Hatsu to Haru 1996 (January to March)
Addendum: they seem to prefer to have 2 Sekiwake, but I couldn’t discern them being particularly aggressive in trying to winnow down the number of Sekiwake if there are 3 or more. Usually attrition will naturally restore them to 2 Sekiwake sooner or later.
So we now have a bit of an idea how this works so I’ll give an example where we can see this in practice using Nagoya 2020 (July) and Aki 2020 (September) - this is a pretty straightforward one.
Rule 1: Shodai and Mitakeumi both Kachikoshi so they retain their positions. Rule 3: Daieisho hit the magical 10+ win threshold so he opened a new slot. Note that Okinoumi’s 9 wins do not meet that threshold and so we shouldn’t expect him to open a slot himself. And then finally Rule 4: the new slot is the lowest opened one. Now check out the Banzuke and how it aligns with our rules.
I kind of stacked the deck in my favor for this one. It was actually an easy application of our rules. Just for fun - or perhaps to inflict the psychological damage I sustained looking at all these Bashos over the past few nights and trying to make consistent rules on others - let’s do a harder one. Here we have Aki 2001.
Now take a few beats and try to think who will end up where. The easy ones are Tochiazuma at S1E and Miyabiyama at the the last Sekiwake slot, but how many slots?
If you had Kotomitsuki opening up a slot in Sekiwake from M2E then collect your winnings. Also, fun Asashoryu cameo but his 10-5 at M1W wasn’t enough to open a slot, apparently. Basically, they can get a little arbitrary but the general contours of the rules will apply.
One more because we get to see a young Hakuho on the come up and then some fun facts! So we begin with Kyushu 2004.
Wakanosato and Miyabiyama both obviously retain their slots, and a new slot has to be opened up for Tochiazuma (he really was just constantly shuttling back and forth through the Sanyaku in the late 90’s through mid 2000’s) being demoted. Perhaps Kotomitsuki could open a slot or even Hakuho with a Jun Yusho from Maegashira 1. However, I have some bad news for you:
Neither Kotomitsuki nor Hakuho merited an additional slot in the Banzuke makers’ minds. I do kind of understand not wanting to open up 2 slots per Rule 6, but it does feel like both Kotomitsuki and Hakuho got a bit of a raw deal given 2 slots opening up in one tournament isn’t unprecedented. Furthermore, one of the slots opened up was because of an Ozeki demotion. Anyways, if you’re keeping track this is the earliest Hakuho snub by the sumo establishment I’m aware of haha.
The modern (i.e. not in the 1800’s or even before) record for tournaments at Sekiwake is Kotomitsuki (despite that snub above he did reach Ozeki so he can hang his hat on that) with 22. Mitakeumi is the active wrestler with the most appearances at Sekiwake with 19. He seems like he doesn’t have another Sanyaku run in him despite still being relatively young (or at least not near retirement, imo) so he won’t be able to catch that record that I wouldn’t describe as dubious, but also wouldn’t describe as honorable (we really only describe distinctions as dubious eh?).
I know I started this piece with a bit of a jokingly dismissive attitude, but this was actually a pretty fun topic to research. A lot of future Yokozuna and Ozeki popped up as I was going through it and I got to get acquainted with some of them for the first time, like Kotomitsuki. If you’re curious for the algorithmic Banzuke implications, I don’t have it coded up yet but I’ll paste some pseudo code and explanations below:
existing_sekiwake.sort_by_wins()
or order existing Sekiwake by wins last tournament
existing_sekiwake.append(demoted_ozeki)
add to the back of the list of Sekiwake the demoted Ozeki
If existing_sekiwake < 4 (then the below will be done)
if banzuke[komosubi:maegashira-2] wins >= 12: insert wrestler in front of demoted Ozeki if they exist
okay this one I had to explain the pseudo code in the code haha but the idea is if in the Banzuke from Komosubi to Maegashira 2 has 12 or more wins, then they’ll go in front of the demoted Ozeki, if there’s an existing Ozeki, if not then just to the back
if banzuke[komosubi:maegashira-1] wins >= 11: insert wrestler in front of demoted Ozeki if they exist
same as above but with 11 wins and restricted to Komosubi to Maegashira 1
if banzuke[komosubi] wins >= 10: insert wrestler in front of demoted Ozeki if they exist
same as above but with 10 wins and restricted to Komosubi
It’s probably not perfect, but I think it’ll do a good job and obviate a lot of the concerns over my/the algorithm’s handling of Sekiwake in the Banzuke generation process. Also, if you’ll allow me to gloat a bit: the cowards at wikipedia1 wrote, “Typically 11–4 or better as a komusubi is sufficient for promotion to sekiwake”
It’s not a guarantee, but looking at the data more times than not 10-5 is sufficient to land at Sekiwake. They should have the courage of conviction to write 10-5 or better. Here at Ozeki Analytics we’re pushing the analytical frontier of Sumo wrestling and making our cliches more accurate. Anyways, hopefully you all enjoyed this piece. I have some more research I’m already knee deep in so you can expect us to continue our journey through sumo using data to try and understand how it works better. Thanks for reading!
Editor’s note: not capitalized intentionally. Readers can draw their own conclusions on if that is meant to be disrespectful or not ;)
Interesting stuff as always!