Sumomasen 01 - A Brief(ish) History of Yokozuna

Analysing how sumo's top dogs have fared over the years and some stats on their success.

Hakkeyoi and welcome to Sumomasen!

Each month I plan to post about some sumo analytics I’ve been researching - have been following sumo pretty avidly for almost 2 years now and there are so many topics I’d like to get into: rikishi from each decade, foreign born rikishi, how current oyakata’s careers compare, analysis of individual stables etc.

For my first post I thought a great place to start would be taking a look at some of the creme de la creme of the sumo world - the Yokozuna. There have only ever been 75 rikishi to officially claim this rank in the history of the sport. In fact, the origins of the Yokozuna rank are somewhat debated, with its first appearance on the banzuke around 1890. Before this it was more a bragging rights licence for Ozeki with influential patrons, allowing them to perform a dohyo-iri. Nonetheless, the Yokozuna is undoubtedly the most formidable opponent a rikishi could face in the ring, and one that steeped in Shinto tradition.

For this research I used sumodb - which has data on all 74 Yokozuna (Onosato not being listed yet at the time of writing as he’s not had his first bout at the rank). However, I’ve only looked at data for Yokozuna-destined rikishi who debuted in the modern, 6-basho era in an attempt to level the playing field. This takes us back to the 51st Yokozuna Tamanoumi, and while this does mean we miss out on some legends such as Taiho, Wakanohana I and Futabayama, it still encompasses the most recent third of all Yokozuna.

So, what makes a Yokozuna?

First, I decided to look at the number of basho a Yokozuna goes through as well as their winning percentage on their way to earning the rope.

I wanted to highlight three things in particular on these two graphs - the ‘average’ Yokozuna (dashed blue line), and the data for Hoshoryu (red line) and newly promoted Onosato (yellow line). Notable here is that win% drops 20% at a pretty consistent rate by until Komusubi, before future Yokozuna seem to find their feet and rise to the top.

Interestingly, out of the last 25 Yokozuna, only 1 has ever had a higher-than-average number of basho as Jonokuchi - that man being the great Hakuho. While that may be somewhat of an outlier (with the payoff being that he also has the most basho, wins, jun-yusho and yusho as Yokozuna of all time), something that provides hope for a thrilling future is Hoshoryu and Onosato both rising to the top faster than the average of 56 basho (Onosato’s 13 total basho to Hoshoryu’s 42 is certainly something in itself despite Onosato starting in Makushita).

Next, I looked at how different factors affected the following stats: wins as Yokozuna, win% as Yokozuna, career win% and yusho as Yokozuna.

Surprisingly the correlation between height and wins, win% and yusho was very weak¹. While there were a couple of outliers (namely Hakuho skewing the wins and yusho data, and Kisenosato’s 50% win rate as a Yokozuna an obvious anomaly), and a relatively small sample size of 25, I had originally expected a stronger correlation here. The heights² of the Yokozuna in question ranged from 177 cm (Tamanoumi) to 204 cm (Akebono), with a mean height of 186.4 cm (The same height as Takanohana, Harumafuji, Kakuryu and Hoshoryu).

For the yusho graphs here I have omitted Hoshoryu, Futahaguro and Wakanohana with their 0 yusho as Yokozuna, otherwise the r-squared value for these drops to 0.

Weight proved just as inconclusive in deciding a Yokozuna’s wins, yusho or win rates. Of these indicators, a higher weight seemed to suggest a slight disadvantage when it comes to yusho, wins and win% as Yokozuna. This is in contrast to height which, if anything, suggested a taller Yokozuna had a slight advantage. Yokozuna weights were more widely distributed with a range of 109 kg separating the lightest (Chiyonofuji at 115 kg), and the heaviest (Akebono at 200 kg), with an average of 150.5 kg (equivalent to Kakuryu, Kitanoumi & Kotozakura).

Earning a raise and good first impressions

What I found interesting was the 7 total scores of 8-7 or 9-6 in the basho before a successful Yokozuna run. Only two of these, from Kakuryu and Hoshoryu, were not followed up with consecutive yusho. Takanohana and Harumafuji were both able to claim the maximum 30 wins through two zensho yusho to earn their promotion. Takanohana also boasts the highest number of wins with 41 over his final three basho as Ozeki, narrowly beating Wakanohana II and Onokuni by 1 win.

At 2 yusho and 1 jun-yusho, Takanosato, Kitanofuji, Mienoumi and Takanohana had the most dominant starts as Yokozuna - Takanosato claiming top spot with 41 wins, 1 ahead of the others. Terunofuji also had an impressive start with two back-to-back yusho, including one zensho, and was one of only three rikishi (along with Takanohana and Musashimaru) to manage 4 yusho over this 6-basho stretch.

Going into promotion, this batch of Yokozuna averaged 35.8 wins in their last three Ozeki basho, 25.8 over the two basho immediately before promotion, and impressively averaging 14 wins in their final Ozeki basho.

Post-promotion however, the three basho average is slightly lower at 32.2 wins, with new Yokozuna managing on average 21.6 wins over their first two basho, and a reputable but not dominant 10 wins in their first outing.

With the average age of a debutant Yokozuna being 26, it shows the importance of rising to the top quickly. Here we can see that age has the strongest correlation of the factors looked at, with the length of a Yokozuna’s tenure steadily decreasing with a higher age at debut. Yokozuna who were promoted age 30 or above averaged just 10.3 basho, while those under 30 at promotion averaged 35.7 basho.

The youngest debutant Yokozuna was Kitanoumi at 21, while Tamanoumi and Kotozakura were the oldest at 32.

Ozeki Analytics wrote a very interesting article on Makuuchi debut ages and how a younger debutant most often leads to a more prosperous career.

Here we can see age also has a reasonable correlation with the number of yusho won as Yokozuna. Hakuho obviously has the most Yokozuna yusho (42), followed by Chiyonofuji with 29 and Asashoryu with 23. However, Asashoryu has the highest yusho rate, with him coming 1st in 54.8% of his 42 basho as Yokozuna.

Only 6³ Yokozuna in the modern era have ever managed a 1st or 2nd place finish in >50% of their tenure. They are: Tamanoumi (90% of 10 basho), Hakuho (70.2% of 84 basho), Asashoryu (69.0% of 42 basho), Chiyonofuji (64.4% of 59 basho), Kitanoumi (58.7% of 63 basho) and Takanohana (55.1% of 49 basho).

The final factors I looked at were the number of basho a Yokozuna spent in the lower ranks (Jonokuchi up to Ozeki), and their win% pre- and post-promotion. Here it seems that while a weak correlation, the greater the length of time spent in the lower ranks, the lower the win% as Yokozuna.

Based on these rudimentary correlations, with Onosato debuting at 24 it suggests that he’ll have a ~43 basho tenure with about 15 yusho and 7 jun-yusho as Yokozuna. Although I feel he’s definitely capable of surpassing that.

A good workman never blames his tools

Looking at the breakdown of Yokozuna by birthplace, we can see each of the three nationalities has a different set of unique tools at their disposal:

• Japan has the highest number of Yokozuna (with there having been only 8 foreign Yokozuna out of the last 25⁴) and therefore the highest number of basho and wins,

• The USA (or rather American Samoa and Hawaii) having the tallest and heaviest Yokozuna, as well as the longest tenured on average and most successful in terms of win%,

• Mongolian Yokozuna are able to boast the highest average wins per Yokozuna.

Speaking of tools, of the 6,382 victories, 58 unique kimarite were used, with the usual yorikiri (37.8%), oshidashi (11.5%), uwatenage (10.7%), hatakikomi (5.0%) and sukuinage (3.0%) being the most common. Hakuho and Chiyonofuji were the most technically diverse with 37 and 35 unique kimarite used respectively during their Yokozuna years, while Kotozakura (13), Futahaguro (11) and Kisenosato (8) were the least diverse.

There were a number of techniques used only once, those being: hansoku, kekaeshi, ketaguri, kotehineri, kubihineri, okuritsuridashi, sokubiotoshi, tsukite and tsumatori.⁵

Of the 1,795 Yokozuna losses (excluding those of Wakanohana I, Asashio, Kashiwado, Taiho, Tochinoumi and Sadanoyama, who were active after 1958, but had debuted earlier, and including playoffs), Maegashira ranked opponents accounted for 34.9%. Of the 65.1% defeats to Sanyaku opponents, Ozeki posed the biggest threat, with Yokozuna-on-Yokozuna crime made up almost a quarter of that, at 277 bouts, or 15.4% of the overall losses.

This batch of Yokozuna were defeated only twice by opponents ranked below M10. in the July basho of 1991, M13e Kotofuji toppled Y1e Asahifuij on the penultimate day of the tournament and went on to claim his first and only yusho (at 14-1), and his second fighting spirit prize. The second occasion came when M12w Kotonishiki defeated Takanohana in the final basho of 1998. He also went on to win the basho with 14-1, earning himself two special prizes.⁶

Hoshoryu vs. Onosato

And finally, a little look at what’s to come. Head-to-head, Hoshoryu’s had the better of Onosato in the admittedly few times they’ve met with the 74th Yokozuna winning 6 out of their 8 bouts. Interestingly enough however, Onosato has gone on to have a better record at the end of four of these basho, with a tied record twice, and Hoshoryu taking the other two.

To back this up I took a look at how both fared against the rikishi on the banzuke for this previous 2025 Natsu basho, and unsurprisingly Onosato has a much better record against the rest of the field, albeit with a little over half as many bouts as Hoshoryu. We’re entering exciting times in Sumo!

1. An r-squared value of 1 indicates a direct correlation between the dependent variable (y-axis) and the independent variable (x-axis). Usually anything below r-squared = 0.400 is considered very weak.

2. Height data was taken from sumodb and is estimated to be the height of the rikishi at the time of their promotion.

3. Technically Hoshoryu would be on this list with his 1 jun-yusho out of two basho as Yokozuna, however as his tenure is still ongoing, I’ve left him off.

4. There is a cap of 1 foreign born rikishi per stable, so this does limit the number of potential foreign Yokozuna compared to local, Japanese-born ones. Here, Akebono is taken as the first foreign-born Yokozuna as Sakhalin, birthplace of Taiho, was technically a Japanese controlled territory at the time of his birth. < Thanks to Oyster5436 on Reddit for raising this!

5. Hansoku - exact video not available on sumodb, here’s another example

Kekaeshi - exact video not available on sumodb, here’s another example

Ketaguri - exact video not available on sumodb, here’s another example

Kotehineri - Kisenosato vs Shohozan (Haru 2017)

Kubihineri - Asashoryu vs Kotooshu (Aki 2005)

Okuritsuridashi - Asashoryu vs Ama (Kyushu 2005)

Sokubiotoshi - exact video not available on sumodb, here’s another example

Tsukite - Harumafuji vs Myogiryu (Natsu 2016)

Tsumatori - exact video not available on sumodb, here’s another example

6. The lowest rank a Yokozuna has lost to, according to sumodb, is M14. On the 4th day of the January basho in 1941, 34th Yokozuna Minanogawa (Y1e) lost via sukuinage to Komatsuyama (M14w). The same tournament of the following year they rematched, with Komatsuyama now M6e, and Minanogawa getting revenge via kimedashi (a rare technique in itself, having only been used 0.38% by Juryo and Makuuchi rikishi since 1991).