Measuring Ōzeki Quality

[Seiyashi 4,097]

6 hours ago, Asashosakari said:

Let's keep in mind that "kunroku", by its very nature as a word denoting "9-6", is a relatively modern invention since it obviously came into being in the 15-bout era only. As I've had reason to mention a couple of times recently, the first decade and a half of those tournaments - ignoring the brief prior period in Futabayama's heyday which had its own issues - from 1949 to 1964 had significantly more restrictive match-making rules which resulted in many high-rankers of the day getting a much easier ride than high-rankers did after 1965, and even moreso than they do nowadays where almost no heya manages to put more than two rikishi into the joi concurrently anymore and the joi competition is very close to the ideal top 16 roundrobin (injuries aside).

I'm probably a broken record on this point - especially over on Reddit which is the unofficial headquarters of "all ozeki suck" sentiment - but I would contend that the reason a great number of ozeki don't actually hit 10+ wins all that often isn't that they're bad or that too many unqualified rikishi are getting promoted to the rank, it's that the double-digit expectation is 50+ years out of date, and arguably only survives because "double digits" is such an easily digested notion that people (the Kyokai brass quite possibly included) don't bother to question its meaning. It was appropriate to denigrate 8/9 wins as unsatisfactory when top-rankers were facing some selection of 15 rikishi out of the top 20 to 25, but it just doesn't match the historical evidence of what high-caliber rikishi are actually capable of achieving when the schedule is closer to the ideal top 16.

Unless people want to make the ozeki rank into a more exclusive thing than yokozuna, it's high time for it to be acknowledged that 9 wins is a perfectly acceptable result for an ozeki, both in an individual basho as well as in their career average at the rank.

I can definitely buy that when the yokozuna KK is considered to be 10-5, so while the kunroku term exists, it seems incongruous that ōzeki are also asked to live up to that same standard. If anything else, I'd have thought that an ōzeki KK should be 9-6! If the term is that modern then perhaps it's not entirely impossible to track down its invention, although my preliminary attempt on Google has failed quite badly. 

I'd think there's room here for combining this with the sum of opposition score analysis that @rhyen has suggested with the thought that @Yamanashi voiced regarding what happened with multiple ōzeki, to see if the Hakuhō-era ōzeki really can be given a pass for putting up bad scores because their opposition was just so much tougher.

One minor thing, though: while kunroku might date from 1949 onwards, this survey only covers ōzeki after the kadoban changes in 1969, so unfortunately this dataset can't cross-compare whether ōzeki schedules and results were likely to be stronger in the old days of torikumi-making. Not yet, at least.

1 hour ago, Gurowake said:

One thing that deserves a mention is just how much volatility there is to sumo results.  With the exception of the absolute worst and best in the sport, performances tend to spike up and down quite a lot.  There are a lot of rikishi who only make it to a rank or division for one basho and then get nowhere close after that.  It's pretty typical for rikishi who peak at ranks less than sanyaku to not have many tournaments near their high rank.  Given the rank protection for Ozeki, it's not surprising that when someone manages to string together their best 3-tournament performance of their life and have it be just good enough to get to Ozeki that they are going to likely stay around the rank for a couple years and put up mediocre results.  I'm sure if people were to run simulations after figuring the appropriate amount of variance, they could see that these kind of career paths are totally expected and not unusual at all.

This is where the strength of the nagora rating (NR) comes into play, because you can distinguish those quite easily especially when looking at the kunroku table. The ones who spike and make it into ōzeki with a run of good form will be those with a pretty good NR for their first ōzeki basho, then have relatively low standard deviations coupled with a high % of KR KK and a low minimum NR. So the most obvious candidate is Kaiketsu, but because of the brevity of their tenure Tochinoshin' and Daiju's stats can't reflect that. 

That being said, there are also a number of ōzeki who put on a burst of good sumō to reach the rank, languish for a few basho or so, then go on sporadic runs of good form as well. That's actually the story of most of the ōzeki surveyed, to be honest. Their standard deviations for their NR shows that they hang around anywhere from 5.7k to 6.8k NR about 65% of the time, a spread of results which corresponds to (if I understand nagora correctly) a 57% to 68% win rate per basho which is quite bang on 9-6 at the top end, and then you have monsters like Konishiki and Wakashimazu who were routinely averaging 80% win rates at their peak - or 12-3s in other words. But the point is definitely taken re the rank protection of ōzeki skewing things a lot in this respect.

[Seiyashi 4,097]

2 hours ago, Gurowake said:

I also think a lot of the disappointment with Ozeki that put up results that barely keep them at the rank is, right now especially, fans really want to see new Yokozuna.  Every time an Ozeki is promoted, there's some hope that his most recent string of tournaments is no fluke and he'll continue to perform at that level long enough to string together a Yokozuna run.  Since all our Ozeki are either recently promoted or still young and thus has time for improvement, there's the hope that we might get another Yokozuna, but the results we get instead show that they're nowhere close.  Particularly with only one ailing Yokozuna left, people sorta have the right to be disappointed when no one else can step up and take the rope.  Fans want someone to be excited about and the marketers want people to put on promotional material.  It's really not all that strange to have a recent Sekiwake Yusho winner featured with the rest of the Ozeki, even if in my opinion that Yusho was something of a fluke and it's unlikely to be repeated very often.  I think they also have Ura highlighted a bit more as well, even though he's done nothing that great to deserve it.  At least they aren't defaulting to Endo.

To add to this, the last "organic" ōzeki promotion to yokozuna was Kisenosato all the way back in 2017. Terunofuji is kind of weird because he's clearly not this generation's primus inter pares, and he would either have been promoted earlier or not at all, and in either case probably be retired by now if Isegahama hadn't convinced him to stay. So on top of wanting a new yokozuna, we want this era's yokozuna, and none of the ōzeki have been able to step up to the plate and bat. I wonder if Asanoyama's disgrace has something to do with it, because realistically he was the only likely yokozuna candidate amongst the four ōzeki of this batch, and his sudden fall has just sort of refracted the burden of that expectation on the other three.

Edited June 9, 2022 by Seiyashi

[Seiyashi 4,097]

11 hours ago, Asashosakari said:

Let's keep in mind that "kunroku", by its very nature as a word denoting "9-6", is a relatively modern invention since it obviously came into being in the 15-bout era only. As I've had reason to mention a couple of times recently, the first decade and a half of those tournaments - ignoring the brief prior period in Futabayama's heyday which had its own issues - from 1949 to 1964 had significantly more restrictive match-making rules which resulted in many high-rankers of the day getting a much easier ride than high-rankers did after 1965, and even moreso than they do nowadays where almost no heya manages to put more than two rikishi into the joi concurrently anymore and the joi competition is very close to the ideal top 16 roundrobin (injuries aside).

Scratch this hypothesis:

Ōzeki	ŌB	Percentages
		ŌQ KK	KR KK	MK
Yutakayama	34	47	26	26	0.601	9.01
Tochihikari	22	36	36	27	0.593	8.90
Kitabayama	30	30	50	20	0.585	8.78
Wakahaguro	13	31	31	38	0.567	8.50
Kotogahama	28	36	21	43	0.561	8.41
Matsunobori	15	0	53	47	0.495	7.43
Ōuchiyama	7	43	14	43	0.585	8.78
Mitsuneyama	8	38	25	38	0.553	8.29

These are the ōzeki promoted after basho became regularly 15 days. The numbers on the right are their win average and win average multiplied by 15 for a basho result extrapolation. So even the great Yutakayama, while scoring relatively well compared to his peers, didn't even manage to always stay at or above 10 either, although he does have the generally highest percentage of "ōzeki quality" KK.

And their averages, while a little higher relative to maybe modern ōzeki, aren't that much higher to clearly prove the hypothesis that they did get an easier ride either.

Apropos of Yutakayama, it's quite interesting that the names supposedly often tossed around as good ōzeki, or at least editorialised in their Wikipedia articles as strong ōzeki, (Yutakayama, Takanohana I, Kaiō, Chiyotaikai) don't actually have that fantastic stats in terms of high performance. Their claim to fame is mainly their longevity. It's Asanoyama, Kotokaze, Baruto, Konishiki, and Wakashimazu who generally lead those stats in terms of performance, yet three in that list of five aren't very much talked about in lists of good ōzeki.

So I guess the upshot of it is, what made those ōzeki good, if it wasn't regularly hitting 10+? And regularly hitting 10+ like Kisenosato did got him slapped with the reputation of being "not quite good enough" to make it to yokozuna. And if longevity is a subconscious expectation, it's clearly daft to examine the current batch of ōzeki with that lens. The more I delve into this ōzeki expectation, the more it smacks of impossibility - damned if you do, damned if you don't, as it were.

Edited June 9, 2022 by Seiyashi

[Yamanashi 3,860]

2 hours ago, Seiyashi said:

To add to this, the last "organic" ōzeki promotion to yokozuna was Kisenosato all the way back in 2017. Terunofuji is kind of weird because he's clearly not this generation's primus inter pares, and he would either have been promoted earlier or not at all, and in either case probably be retired by now if Isegahama hadn't convinced him to stay. So on top of wanting a new yokozuna, we want this era's yokozuna, and none of the ōzeki have been able to step up to the plate and bat. I wonder if Asanoyama's disgrace has something to do with it, because realistically he was the only likely yokozuna candidate amongst the four ōzeki of this batch, and his sudden fall has just sort of refracted the burden of that expectation on the other three.

Some good thoughts here.  I don't feel qualified to give qualitative judgments on historic Ozeki, as I haven't followed the sport for as long as many of you have.  Because of that, I don't have a feel for whether this is an anomalous period in Ozumo.  There hasn't been an Ozeki -> Jonidan -> Yokozuna path before, and it seems that Asanoyama's less-than-intai punishment is unique [whether that matters depends on how far he comes back].

[Seiyashi 4,097]

22 hours ago, rhyen said:

Yes, Sum of Opponent Score (SOS) would be a good indication of the opposition.

Philosophical question: if you're interested only in when the rikishi shows up to fight, then you'd exclude only zenkyu, right? That still leaves in cases like Asanoyama in Kyūshū 2021, where he fought three days then withdrew, leaving him with a bad but non-zero sum of opponent score? Where would be your discard threshold for this? 

[Hankegami 537]

9 hours ago, Seiyashi said:

To add to this, the last "organic" ōzeki promotion to yokozuna was Kisenosato all the way back in 2017. Terunofuji is kind of weird because he's clearly not this generation's primus inter pares, and he would either have been promoted earlier or not at all, and in either case probably be retired by now if Isegahama hadn't convinced him to stay. So on top of wanting a new yokozuna, we want this era's yokozuna, and none of the ōzeki have been able to step up to the plate and bat. I wonder if Asanoyama's disgrace has something to do with it, because realistically he was the only likely yokozuna candidate amongst the four ōzeki of this batch, and his sudden fall has just sort of refracted the burden of that expectation on the other three.

Depends what one means for a generation. Terunofuji is undeniably the best rikishi of the 1987-1991 batch, while Kisenosato got the last train for his own batch (1984-1986) which also included Hakuho, Harumafuji, and Kakuryu. Not to underestimate Kisenosato, but his story is somewhat comparable to other latecomers like Asahifuji (last Yokozuna of the 1960-1963 boys ,and he is the 1960 guy), Takanosato (the last of the 1952-1955 boys, and again the 1952 guy), and Kotozakura (1938-1944 batch from Kashiwado to Tamanoumi, he was class 1940). Granted, all of them wrestled in eras dominated by all-time champions (Hakuho for Kisenosato, Chiyonofuji for Asahifuji, again Chiyonofuji and Kitanoumi for Takanosato, and Taiho for Kotozakura) so there are reasons for them.

Granted, Terunofuji could have made it already in 2017 had not suffered from injuries. He was still 25 back then (26 in November), a good age to get the rope. The simple fact that no one jumped on the occasion and got the rope between 2017 and 2021 shows he really is the best of his batch which were in their prime during those years. And to be honest Terunofuji fully belongs to the present: a wrestler his age (30) is still very active. Saying he would have retired by now requires imagining an alternative universe in which Terunofuji clinched the rope and suffered from the same injuries only afterwards. His comeback story is full part of his persona, Yokozuna rank included.

The "awkward" part for Terunofuji is that we know he won't last. With a healthy Terunofuji in town no one would look for another guy. We would just sit and enjoy the one-man show like pretty much everyone did in 2010 with Hakuho, or 2004-2005 with Asashoryu, and so on. Someone would eventually show up anyway. The only reason we need a new guy now is that we instinctively need some backup ready for the day Terunofuji's knees will give up.

About Asanoyama, he surely has the physique for the rope but I am not 100% sure he would have been the answer. He and Terunofuji basically swapped places in 2017 and 2021, one down and the other up both ways. Asanoyama was 23 when he reached Maakuchi in 2017, yet he reached Ozeki "only" in 2020 and since then he never had a serious chance to be a Yokozuna candidate before being suspended. The idea he would have managed to get there in 2021-22 is just hypothetical. Strictly speaking, Terunofuji is the dominant rikishi of this banzuke. He just managed to get a yusho at what, 80% of his peak form? He's strong enough to play the Asashoryu were his knees ok. And Asashoryu had to wait for a guy 5 years his junior to show up (insert here a gif of Takakeisho, class 1996 and roughly 5 years younger than Terunofuji, painstakingly trying to rub his hands for the perspective).

Back to the topic (sorry everyone for the gigantic OT), my impression is that the 9-6 thing is just an idealistic fetish. In my opinion, what the fans actually look in an ozeki is the number of times he's pushing for the rope, that is his honours. This way Tochiazuma (3 yusho as an ozeki), Kaio (4 yusho and 7 JY), Chiyotaikai and Takanonami (both 2 yusho and 7 JY) but also Yutakayama (5 JY, 8 overall) are understandably regarded as strong ozeki, no matter how often they actually got them. Hype is a thing. This is also why many are not going to drop their Takakeisho pin despite his injuries, or aren't really worried about Mitakeumi's MK. They already did good, so everyone expect them to do good again sooner or later. Also, selective memory helps fans getting rid of bad basho records and remember only good ones.

Edited June 9, 2022 by Hankegami

[Reonito 1,469]

Late to this, but might be informative to break out performance vs Y and vs O separately rather than lumping them. The win rates seem surprisingly low, and my guess is that this is explained by the "dominant Yokozuna" effect.

[Reonito 1,469]

11 hours ago, Seiyashi said:

I can definitely buy that when the yokozuna KK is considered to be 10-5

Isn't this basically based on an off-hand comment by Harumafuji?

[Asashosakari 19,403]

10 hours ago, Seiyashi said:

And their averages, while a little higher relative to maybe modern ōzeki, aren't that much higher to clearly prove the hypothesis that they did get an easier ride either. (...)

So I guess the upshot of it is, what made those ōzeki good, if it wasn't regularly hitting 10+?

I didn't say the pre-1969 ozeki were any better, I said that the circumstances of the time made it more palatable to expect them to be better. There were routine public complaints about both the yokozuna and the ozeki in the post-war period, usually alleging that politicking was leading to unqualified rikishi being promoted to the ranks (or conversely, qualified candidates being refused), which first led to the creation of the YDC and then more gradually to the development of today's relatively tough ozeki promotion standards as the authorities came to grips with the expanded annual schedule of tournaments.

Also, a one-size-fits-all comparison purely of win-loss records and averages doesn't address the main issue, namely that the opponent schedules weren't uniformly easier, but rather (comparable to the Futagoyama effect in the 1990s) in very uneven fashion from one yokozuna/ozeki to the next.

Edited June 9, 2022 by Asashosakari

[Seiyashi 4,097]

1 hour ago, Asashosakari said:

I didn't say the pre-1969 ozeki were any better, I said that the circumstances of the time made it more palatable to expect them to be better. There were routine public complaints about both the yokozuna and the ozeki in the post-war period, usually alleging that politicking was leading to unqualified rikishi being promoted to the ranks (or conversely, qualified candidates being refused), which first led to the creation of the YDC and then more gradually to the development of today's relatively tough ozeki promotion standards as the authorities came to grips with the expanded annual schedule of tournaments.

Also, a one-size-fits-all comparison purely of win-loss records and averages doesn't address the main issue, namely that the opponent schedules weren't uniformly easier, but rather (comparable to the Futagoyama effect in the 1990s) in very uneven fashion from one yokozuna/ozeki to the next.

Out of reacts, but thanks for the explanation. I wonder whether the public complaints about unqualified rikishi being promoted used politicking as a fig leaf, and the root cause was just simple discontent with the performances of the ōzeki like Hankegami has suggested.

I'm working on the strength of opposition score now (just need rhyen - or someone else more conversant in how to produce and interpret ratings scores - to suggest how to deal with kyujō).

[Godango 996]

4 hours ago, Seiyashi said:

Out of reacts, but thanks for the explanation. I wonder whether the public complaints about unqualified rikishi being promoted used politicking as a fig leaf, and the root cause was just simple discontent with the performances of the ōzeki like Hankegami has suggested.

I'm working on the strength of opposition score now (just need rhyen - or someone else more conversant in how to produce and interpret ratings scores - to suggest how to deal with kyujō).

How with Strength of Opposition be grouped? The kyujo would only effect a by basho grouping, whereas if you had a time period (calendar year, tenure, whatever it may be) you could look an average/median SOO for that period?

DISCLAIMER: I may have fundamentally missed the point.

[rhyen 1,842]

6 hours ago, Seiyashi said:

Out of reacts, but thanks for the explanation. I wonder whether the public complaints about unqualified rikishi being promoted used politicking as a fig leaf, and the root cause was just simple discontent with the performances of the ōzeki like Hankegami has suggested.

I'm working on the strength of opposition score now (just need rhyen - or someone else more conversant in how to produce and interpret ratings scores - to suggest how to deal with kyujō).

Sum of opposition score works on an aggregate basis, so just total up the opponent hoshitori. (Getting the means is the same)

For rikishi going kyujo, it just means that there will definitely be a lower score, which translates into a vaguely weaker opposition (since said rikishi was competing injured, like Mitakeumi in Natsu basho).

[Gurowake 4,087]

Here's another measure: how often the Ozeki managed another run of 3 tournaments that would likely get them promoted again.  For this rundown, I've only done so far those peaked at Ozeki and whose first promotion was during my lifetime.  The criteria for a run to be included is that the sum of wins is at least 32 and the last is at least 10.  Technically KK in all three tournaments is needed, but I don't think anything was disqualified that way (yet).  If the runs overlap at all, they're included together; that means some of the intermediate tournaments were not the end of a three-basho run of promotion quality.  The first two tournaments following promotion are handled separately and don't have the wins listed.  Basho count includes Ozeki basho where the rikishi competed at least once plus any Sekiwake tournaments immediately after demotion where they secured repromotion the easy way.  Tournaments where the Ozeki withdrew after getting their 8th win are marked with "+kyujo"

Upshot: The last 40 years, which is what most people are probably going to remember, have had much stronger Ozeki in general than the last 10.  Takayasu and Takakeisho don't look too bad, nor do Asanoyama or Baruto, but Goeido, Kotoshogiku, and Kotooshu have very few (or none) despite many years at the rank.  The disappointments of the 2000s (Dejima, Miyabiyama and Musoyama) are counterbalanced by two very good Ozeki in Chiyotaikai and Kaio.  Before that there was a good run of Ozeki who succeeded quite a lot while still at the rank, with only Asashio being anything close to a disappointment with a single 4 basho run over 6 years.

Maybe this should be a separate thread, but it's directly related to this thread.

Rikishi	Basho	Promotion quality runs after promotion
Mitakeumi	2	First basho after promotion
Shodai	10	None
Asanoyama	6	First two basho after promotion
Takakeisho	17	2019.09-2020.01 (12-9-11)
		2020.07 thrugh 2020.11 (8+kyujo-12-13)
Tochinoshin	7	None
Takayasu	13	2017.11 through 2018.03 (8+kyujo-12-12)
		2018.07 through 2018.11 (9-11-12)
Goeido	33	None
Kotoshogiku	32	First basho after promotion
		2015.09-2016.01 (11-8+Kyujo-14)
Baruto	15	First basho after promotion
		2011.07-2012.03 (11-10-11-14-10)
Kotomitsuki	17	First two basho after promotion
Kotooshu	47	First basho after promotion
		2009.03-2009.11 (10-9-13-9-10)
Tochiazuma	29	First two basho after promotion
		2002.01-2002.05 (13-10-10) These were directly after promotion
		2005.01-2005.05 (11-10-12)
Kaio	64	First two basho after promotion
		2000.09-2001.03 (11-11-10-13) These were directly after promotion
		2002.01-2002.05 (9-12-11)
		2003.03-2003.07 (10-11-12)
		2003.11-2004.11 (10-10-13-10-11-13-12)
Miyabiyama	8	None
Musoyama	26	None
Dejima	12	First two basho after promotion
Chiyotaikai	61	2000.05-2000.09 (11-11-10)
		2002.05-2002.09 (11-14-10)
		2003.03-2004.07 (12-10-11-11-10-10-13-9-10)
Takanonami	38	First basho after promotion
		1994.03-1995.01 (12-9-12-12-12-11) These were directly after promotion
		1996.01-1996.11 (14-11-12-12-9-11)
		1997.07-1998.01 (9-12-14-10)
Kirishima	16	1990.09-1991-01 (13-10-14)
		1991.05-1991.11 (11-10-12-10)
Konishiki	38	Second basho after promotion (but NOT first)
		1987.09-1988.01 (12-8-13) This includes the second basho after promotion
		1989.11-1990.07 (14-10-13-12-10)
		1991.03-1992.07 (9-14-12-11-13-12-13-9-10)
Hokutenyu	44	1983.11-1984.07 (11-9-12-10-10)
		1985.07-1985.11 (13-9-12)
		1988.11-1989.05 (10-10-12-10)
Asashio	35	1984.11-1985.05 (10-9-13-11)
Wakashimazu	28	First basho after promotion
		1983.05-1985.03 (13-11-13-11-11-14-9-15-11-11-9-12)
Kotokaze	22	First two basho after promotion
		1982.09-1984.01 (9-10-14-11-11-12-11-11-11)

Edited June 10, 2022 by Gurowake

[Seiyashi 4,097]

10 hours ago, Reonito said:

Late to this, but might be informative to break out performance vs Y and vs O separately rather than lumping them. The win rates seem surprisingly low, and my guess is that this is explained by the "dominant Yokozuna" effect.

Yes and no. Dominant yokozuna are indeed responsible for quite a lopsided thrashing of some ōzeki, but their win rates against their fellow ōzeki aren't necessarily any better:

Ōzeki	Y				Ō
	W	L	Most frequent	%	W	L	%
Mitakeumi	0	1	Terunofuji (0-1)	0	2	1	67
Shōdai	0	3	Terunofuji (0-2)	0	1	8	11
Asanoyama	0	0	NA	0	2	2	50
Takakeishō*^	0	5	Terunofuji (0-3)	0	8	3	73
Tochinoshin*^	1	4	Hakuhō (0-2)	20	1	5	17
Takayasu	5	5	Kakuryū (5-0)	50	6	5	55
Gōeidō	12	38	Hakuhō (3-15)	24	25	25	50
Kotoshogiku	13	46	Hakuhō (2-24)	22	29	39	43
Baruto	2	11	Hakuhō (2-11)	15	17	24	41
Kotomitsuki	6	19	Hakuhō (3-13)	24	18	16	53
Kotoōshū	11	39	Hakuhō (6-23)	22	55	43	56
Tochiazuma*^	6	11	Asashōryū (5-10)	35	22	28	44
Kaiō	15	49	Asashōryū (8-21)	23	81	59	58
Miyabiyama	0	7	Takanohana (0-4)	0	3	7	30
Musōyama*^	3	9	Asashōryū (0-5)	25	11	27	29
Dejima	3	11	Takanohana (1-7)	21	7	9	44
Chiyotaikai	22	52	Asashōryū (6-25)	30	48	81	37
Takanonami*^	7	24	Akebono (4-22)	23	17	23	43
Kirishima	8	9	Hokutoumi (3-3)	47	5	10	33
Konishiki	24	40	Hokutoumi (10-11)	38	23	33	41
Hokutenyū	21	54	Chiyonofuji (9-27)	28	49	67	42
Asashio	18	39	Chiyonofuji (8-19)	32	28	53	35
Wakashimazu	7	22	Chiyonofuji (2-18)	24	37	38	49
Kotokaze	4	35	Chiyonofuji (1-17)	10	23	19	55
Masuiyama	2	9	Wakanohana II (0-5)	18	1	3	25
Asahikuni	6	36	Kitanoumi (1-17)	14	18	28	39
Kaiketsu*	3	6	Kitanoumi (3-6)	33	8	12	40
Daiju	2	3	Wajima (1-2)	40	2	7	22
Takanohana	22	67	Kitanoumi (6-32)	25	38	29	57
Daikirin	12	24	Kitanofuji (7-9)	33	23	27	46
Maenoyama	2	15	Kitanofuji (1-7)	12	12	11	52
Kiyokuni	10	29	Kitanofuji (5-11)	26	29	28	51
Max	24	67		50	81	81	72.7
Mean	7.7	22.6		22.3	20.3	24.1	43.3
SD	7.2	18.4		13	18.4	19.7	13.4
Median	6	17		23.7	17.5	23.5	43.2
Min	0	0		0	1	1	11.1

The current batch being thrashed by Terunofuji aside, the standouts are Takayasu, Kirishima, and Tochiazuma at least, who have roughly similar win rates against yokozuna and ōzeki alike. Special mention to Daiju and Tochinoshin, who while being terrible ōzeki have the rare distinction of a higher win rate over yokozuna than ōzeki. It does generally make them look better but it doesn't entirely rescue their win rates. 

[Seiyashi 4,097]

4 hours ago, rhyen said:

Sum of opposition score works on an aggregate basis, so just total up the opponent hoshitori. (Getting the means is the same)

Sum of opponent score is in:

Ōzeki	Max	Basho	Mean	SD	Median
Mitakeumi	124	2022.05	121	4	121
Shōdai	125	2022.05	107	24	114
Asanoyama	121	2021.01	81	46	102
Takakeishō	122	2022.03	89	38	110
Tochinoshin	122	2019.03	85	38	113
Takayasu	120	2019.05	86	43	109
Gōeidō	125	2014.11	105	26	114
Kotoshōgiku	131	2015.05	108	25	116
Baruto	126	2011.09	104	31	114
Kotomitsuki	117	2008.03	101	29	111
Kotoōshū	126	2008.01	102	26	111
Tochiazuma	128	2006.09	82	46	111
Kaiō	125	2004.03	101	31	114
Miyabiyama	114	2000.07	107	9	109
Musōyama	127	2004.05	93	37	112
Dejima	119	2000.05	102	22	109
Chiyotaikai	126	2000.09	102	33	114
Takanonami	125	2000.05	104	12	104
Kirishima	132	1992.01	105	30	116
Konishiki	128	1992.05	108	29	116
Hokutenyū	125	1989.11	108	23	114
Asashio	128	1986.01	105	29	114
Wakashimazu	128	1986.01	107	22	113
Kotokaze	129	1985.01	114	15	116
Masuiyama	119	1981.01	95	32	117
Asahikuni	128	1979.07	109	27	119
Kaiketsu*	125	1975.05	120	3	121
Daiju	126	1974.05	100	26	116
Takanohana	127	1972.11	108	24	116
Daikirin	127	1972.11	101	36	117
Maenoyama	120	1971.11	94	39	111
Kiyokuni	129	1972.05	103	29	114
Yutakayama	123	1963.11	113	13	115
Tochihikari	128	1966.01	106	15	108
Kitabayama	125	1962.05	110	23	116
Wakahaguro	126	1960.09	110	32	119
Kotogahama	123	1960.01	88	37	110
Matsunobori	122	1957.09	102	32	113
Ōuchiyama	116	1956.03	99	29	111
Mitsuneyama	130	1955.05	107	18	111
Max	132		121	46	121
Mean	125		102	27	113
SD	4		9	11	4
Median	126		103	29	114
Min	114		81	3	102
Ōzeki	Max	Basho	Mean	SD	Median

The first thing that jumps out at me is that Kaiketsu seems to have had the toughest opposition very very regularly, having both one of the higher means and the smallest standard deviation. Kirishima and Kotokaze faced generally better opposition which makes their stats look even better. On the other hand, the quartet of Asanoyama, Takakeishō, Tochinoshin and Takayasu have relatively bad scores because of a relatively high percentage of kyujō; this is also evident in Tochiazuma's record. Miyabiyama probably cements his record as worst ōzeki with not just bad stats but also the easiest opposition, so it's definitely a good thing he didn't get repromoted.

(If I can be arsed to rework the script logic a bit, this is probably the best way also to find out which was the cheapest yūshō, but that's an exercise for another day.)

Edited June 10, 2022 by Seiyashi

[Seiyashi 4,097]

4 hours ago, Gurowake said:

Maybe this should be a separate thread, but it's directly related to this thread.

On the contrary this is probably one of the best qualitative assessments of ōzeki tenures, and in combination with the other metrics really confirms and rounds out the picture of the ōzeki rank. I guess we really have the lacklustre ōzeki of the Asashōryū and early to mid Hakuhō eras to thank. 

[Gurowake 4,087]

Here's the rest of the ozeki in the time period we're considering.  Now these guys resemble the ones recently much more than the last of the pervious batch.  Even the supposedly good Ozeki Takanohana really didn't do much better than Kotooshu by the criteria I'm using, and he's the best of the bunch.  That's probably why he was seen as a good Ozeki - the rest that didn't make Yokozuna then weren't that hot.

Masuiyama	7	none
Asahikuni	22	1977.05-1978.01 (9-9-14-8-10)
Kaiketsu	9	First two basho after promotion (first time)
Daiju	5	None
Takanohana	50	1974.11-1975.03 (11-10-13)
		1976.11-1977.05 (9-12-13-10)
Daikirin	24	Second basho after promotion (but not first)
		The above started a good but technically nonqualifying run of 5 (11-10-10-11-10).  That run might have been good enough taken as a whole.
Maenoyama	9	None
Kiyokuni	27	First basho after promotion
		1970.07-1970.11 (11-9-12)
		He also managed a 13-2 but that only finished 31 in 3, then he only had 9 then 8 the next two.

[Reonito 1,469]

On 09/06/2022 at 19:05, Seiyashi said:

Yes and no. Dominant yokozuna are indeed responsible for quite a lopsided thrashing of some ōzeki, but their win rates against their fellow ōzeki aren't necessarily any better:

I'm finding it mathematically surprising that ozeki win % against ozeki isn't 50% ... but I guess that's ozeki who don't go on to become yokozuna losing to those who do?

[Gurowake 4,087]

1 hour ago, Reonito said:

I'm finding it mathematically surprising that ozeki win % against ozeki isn't 50% ... but I guess that's ozeki who don't go on to become yokozuna losing to those who do?

Obviously, yes.  The win rate of people ranked Ozeki against other people ranked Ozeki is clearly 50%.  But the win rate of people whose highest rank is Yokozuna when ranked at Ozeki while facing other Ozeki is higher than the win rate of of those whose highest rank is Ozeki when ranked at Ozeki and facing other Ozeki.

[Reonito 1,469]

Here's another measure of Ozeki quality. How often (if ever) did they come close to Yokozuna promotion? I'm doing this manually, so only covering 1990 (when promotion standards got tightened up after the Futahaguro debacle) to the present. Somewhat arbitrarily, I've defined "close" as 24+ wins over two basho, with at least 11 wins in each. The results are:

Musashimaru 10 (succeeded on #10)

Takanohana 8 (succeeded on #8)

Kisenosato 5 (succeeded on #5)

Konishiki 5

Takanonami 5

Wakanohana 5 (succeeded on #5)

Hakuho 2 (succeeded on #2)

Kaio 2

Terunofuji 2 (succeeded on #2)

Succeeded on the first instance: Akebono, Asahifuji, Asashoryu, Harumafuji, Kakuryu

Came close once: Baruto, Chiyotaikai, Tochiazuma, Takayasu, Takakeisho

Other Ozeki over this time period never really troubled the YDC and the judging department.

 

 

Edited June 13, 2022 by Reonito

[nagora 88]

Wow. Lots to read tonight.

[Ripe 71]

20 hours ago, Reonito said:

Here's another measure of Ozeki quality. How often (if ever) did they come close to Yokozuna promotion? I'm doing this manually, so only covering 1990 (when promotion standards got tightened up after the Futahaguro debacle) to the present. Somewhat arbitrarily, I've defined "close" as 24+ wins over two basho, with at least 11 wins in each. The results are:

Musashimaru 10 (succeeded on #10)

Takanohana 8 (succeeded on #8)

Kisenosato 5 (succeeded on #5)

Konishiki 5

Takanonami 5

Wakanohana 5 (succeeded on #5)

Hakuho 2 (succeeded on #2)

Kaio 2

Terunofuji 2 (succeeded on #2)

Succeeded on the first instance: Akebono, Asahifuji, Asashoryu, Harumafuji, Kakuryu

Came close once: Baruto, Chiyotaikai, Tochiazuma, Takayasu, Takakeisho

Other Ozeki over this time period never really troubled the YDC and the judging department.

 

 

I'm not sure your criteria for being "close" matches those of NSK when they considered someone for promotion... in fact I'm pretty sure it doesn't. Because between 1989 and 2014 criteria for promotion was pretty simple: back to back yusho. Meaning you were considered only after winning a yusho as a Ozeki. It was only in 2014 the criteria was somewhat relaxed leading to promotions of Kakuryu and Kisenosato.

[Reonito 1,469]

45 minutes ago, Ripe said:

I'm not sure your criteria for being "close" matches those of NSK when they considered someone for promotion... in fact I'm pretty sure it doesn't. Because between 1989 and 2014 criteria for promotion was pretty simple: back to back yusho. Meaning you were considered only after winning a yusho as a Ozeki. It was only in 2014 the criteria was somewhat relaxed leading to promotions of Kakuryu and Kisenosato.

If you look, most of these involved a yusho, a playoff, or at least a jun-yusho. I said "close", not necessarily an actual promotion case.

[Godango 996]

On 13/06/2022 at 06:34, Reonito said:

Here's another measure of Ozeki quality. How often (if ever) did they come close to Yokozuna promotion?

I'd be really interested in this as an exercise in measuring Yokozuna quality (how many more times would they have been promoted?). Would you mind if I stole your idea?

Edited June 13, 2022 by Godango

[Ripe 71]

3 hours ago, Reonito said:

If you look, most of these involved a yusho, a playoff, or at least a jun-yusho. I said "close", not necessarily an actual promotion case.

Problem is that prior to 2014, playoff losses or jun-yusho didn't matter... it was yusho and only yusho. So I'm not sure how you can classify something as "close" when there is no chance of them being promoted. When there is probably not even a discussion if they are worthy of promotion.

Edited June 13, 2022 by Ripe