Sign in to follow this  
Fukurou

Acceptable Ozeki Record?

Recommended Posts

29 minutes ago, Dapeng said:

I want to know the average winnings when they successfully kept their rank at sekiwake. They lost their rank when MK. Also, they never made Ozeki. Surprisingly, both of them had average of 9 wins or 60% winning rate when they successfully kept their sekiwake rank.  

And you don't see an issue comparing this to ozeki and including their MKs? An ozeki only successfully keeps the rank with an MK because the rules say they must MK twice in a row. That isn't successfully defending their rank, it's just enjoying a perk. Comparing only good performances to good and bad doesn't add up. 

Share this post


Link to post
Share on other sites

I think Dapeng might be trying to figure out what statistically would be the average (typical) score for stationary sub-Yokozuna sanyaku.

Since K/S are theoretically less skilled than ozeki, yet must get KK to minimally maintain their rank (in most circumstances), this seems on its face to be a likely way of gathering that information without deep data-mining or mathematical modeling.

Share this post


Link to post
Share on other sites
14 hours ago, TillerTheChiller said:

Yeah, an Ozeki (and Yokozuna) should be like a trial, a representation of the best of the best, and one of the foes you have to beat to be one of the best.

I suppose that does tie in with the literal meaning of the word 'Ōzeki', which of course is 'great barrier/gateway'. i.e. Something to be passed like Gandalf.

Edited by Eikokurai

Share this post


Link to post
Share on other sites
On 16/01/2019 at 15:58, Fukurou said:

 

Just to give some data to look at, here's Goeido's record since his ozeki promotion before September 2014 basho. For these purposes I'm using "record equivalents," where an actual record of 7-3-5 is the same as 7-8.  What I found is that in the 26 basho since his promotion Goeido has finished 7-8, 8-7, or 9-6 17 times (65%). Four times he's been worse than 7-8 (15%), 5 times he's been better than 9-6 (20%). Details below. Is a consistent 8-7/9-6 good enough to remain ozeki (obviously I guess it must be) or is an ozeki really supposed to win more.

And what about the other ozeki?

Terunofuji was 104-210 before he was demoted to Sekiwake and he never had an yusho as Ozeki.

Share this post


Link to post
Share on other sites

Of course the average result for a KK Sekiwake is better than 8 wins--it literally has to be.  "Sometimes they do better than the worst result in the dataset" is the only thing you can say from that number.  

  • Like 2

Share this post


Link to post
Share on other sites
11 hours ago, itchyknee said:

I think Dapeng might be trying to figure out what statistically would be the average (typical) score for stationary sub-Yokozuna sanyaku.

Since K/S are theoretically less skilled than ozeki, yet must get KK to minimally maintain their rank (in most circumstances), this seems on its face to be a likely way of gathering that information without deep data-mining or mathematical modeling.

Sure. The current rule makes many good rikishi getting stuck at sekiwake, thus it's feasible to use sekiwake data for statistics. For example both Wakanosato and Kotonishiki had successfully maintained their sekiwake for multiple consecutive bashos. Other ranks below sekiwake are too dynamic and few stays at the same rank for more than 2 bashos. 

If some one can perform a complete statistics for all sekiwakes (who never made ozeki) like I did for the two, I believe their average winnings will be around 9 wins (60%) per basho. 

 If the Ozeki rank is made the top rank and promotion/demotion same as other ranks, I believe those who successfully defend their Ozeki rank will win at least 9 wins per basho on average. The conclusion is that the passing grade for Ozeki is winning average 60% or 9 wins per basho, instead of just 50% or 7.5 wins per basho.

Edited by Dapeng

Share this post


Link to post
Share on other sites

Ok, I'm going to say this again, this time with numbers.  

I just pulled up the last 100 KKs at M7, an arbitrarily chosen rank.  The average number of wins amongst that group is 9.17.  You will find roughly the same number all the way down to J14.  

(The Sekiwake over the same time period have slightly more KK, and a slightly higher average number of wins per KK--9.52.  This is still not a useful metric.)

Share this post


Link to post
Share on other sites

That metric has one major flaw: you won't expect most of those rikishi with 9.17 wins at M7 to get another KK next basho, let alone on average.

  • Like 1

Share this post


Link to post
Share on other sites

To maintain their rank a Sekiwake can average 8 wins per tournament (by actually hitting 8-7 in all tournaments), 9 wins per tournament, or 10 wins per tournament, plus or minus the fractions.  But if they ever average 11 wins per tournament for more than 2 tournaments, they're called Ozeki (and sometimes they're called Ozeki when averaging less than that - e.g. Kisenosato).

Something was niggling, so I took the average of all KK makuuchi wrestlers in the last four tournaments:

average wins for KK wrestler over the last four basho = 9.35

 

 

 

With the exception of visiting Juryo and withdrawals (both relatively infrequent compared to the total number of makuuchi wrestlers) every win is going to be balanced by a loss, so you're going to end up with records falling into something like a normal distribution, especially over enough basho.

And the thing about a normal distribution is that the average for winners, and the average for losers, is going to be closer to the midpoint (7.5) than either end.  Which is what we see with the average close to 9.

This doesn't deviate by much per rank, except for the tails and certain pinch-points (for obvious and presumably non-obvious reasons).  From the previous four basho:

image.png.dcb7b3711aaa145c830dfcedb38508f9.png

Anyone know the math to model this?

Edited by itchyknee
bad syntax

Share this post


Link to post
Share on other sites
3 hours ago, itchyknee said:

I took the average of all KK makuuchi wrestlers in the last four tournaments

You gotta get a bigger sample than that, man, especially if you do it one rank by one rank. In 10 ranks your sample is just two or three KK basho. 

Share this post


Link to post
Share on other sites
6 hours ago, Atenzan said:

You gotta get a bigger sample than that, man, especially if you do it one rank by one rank. In 10 ranks your sample is just two or three KK basho. 

I know mine wasn't a statistically accurate sample, I just wanted a trend.  And the trend is there for makuuchi as a whole.

I tested M13 (which did very well the last 4 tournaments) for the last 58 tournaments and it reverted to trend as expected, but saw no point in posting the trend.

Share this post


Link to post
Share on other sites

However, for those who are interested, here are the trends for the last 120 basho for a sampling of makuuchi positions.  For YOSK+M8 the right column is normalized to 240 rikishi to allow comparison with the regular M ranks: 612561687_Screenshotfrom2019-01-1910-48-40.png.9a419ad6452acd91c2279b81a131f009.png

  • Like 1

Share this post


Link to post
Share on other sites

I swear to gawd we had this discussion last Summer (July?).  I think I even made a graph about Ozeki wins vs losses.  I can't find the thread now, though ...

Share this post


Link to post
Share on other sites

Takanohana deleted the thread with a karaoke remote and blamed it on Harumafuji because of the OBSC. Also, the answer is 8 wins, but it should be 10 wins; therefore, 9 wins is acceptable, except for when it's 11 wins to qualify. 

Share this post


Link to post
Share on other sites
2 hours ago, Yamanashi said:

I swear to gawd we had this discussion last Summer (July?).  I think I even made a graph about Ozeki wins vs losses.  I can't find the thread now, though ...

I think you must mean this one-- 

http://www.sumoforum.net/forums/topic/38020-kunroku-ōzeki/

And an excellent discussion it was, too. Good analysis in there.

Edited by Shio-kago
  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
19 minutes ago, Shio-kago said:

I think you must mean this one-- 

http://www.sumoforum.net/forums/topic/38020-kunroku-ōzeki/

And an excellent discussion it was, too. Good analysis in there.

Yep, that was it!  Thanks so much.  Reading through the thread, it puts you in awe at the brilliance of the forum members in the old days (i.e. last Summer).

  • Like 1

Share this post


Link to post
Share on other sites
3 hours ago, Yamanashi said:

I swear to gawd we had this discussion last Summer (July?).  I think I even made a graph about Ozeki wins vs losses.  I can't find the thread now, though ...

We did. I started it. :) The thread was called something like Kunroku Ozeki.

  • Like 1

Share this post


Link to post
Share on other sites
6 hours ago, Eikokurai said:

We did. I started it. :) The thread was called something like Kunroku Ozeki.

Did anyone also start one for Sekiwake?  Komusubi?  M1?  M5?  M11?  :-P      I'd take this thread more seriously if the title of this thread is something like "what is better than average Ozeki win-loss record?"   After all, C- is still a passing grade.  Goeido is still passing the grade IMO, an acceptable Ozeki. 

Share this post


Link to post
Share on other sites
6 hours ago, robnplunder said:

After all, C- is still a passing grade.  Goeido is still passing the grade IMO, an acceptable Ozeki. 

Depends on the school. In my area schools C- ends at 70 and F begins at 59. 70% of 15 = 10.5 and 59% of 15 = 8.85 (fail!)

Share this post


Link to post
Share on other sites

I can't seem to import Excel files or jpegs anymore, so here's a description:

 

1) I made a table of wins per basho from 1949-2018 for the Komosubi, Sekiwake, Ozeki, and Yokozuna.

2) I then calculated a weighted mean value of wins for a 15-bout basho:

Komosubi  6.91

Sekiwake   7.70

Ozeki   8.91

Yokozuna   9.18

This result is, of course, distorted by the special rules for Ozeki and Yokozuna: O's can bow out (like Tochinoshin) and hope to overcome kadoban next basho; Yokozunae can go kyujo whenever things are looking bad, even skipping a tournament.  This is seen in the table for number of zero win basho:

K   9

S   18

O   50

Y   143

This skews the trends to a large extent; in fact,

the most likely # of wins for a Yokozuna from 1949 - 2018 was zero!

3) To fix this distortion somewhat, I redid the numbers for Ozeki and Yokozuna, throwing out the zero-win basho.  The new weighted averages are

Ozeki   9.27

Yokozuna   10.75

This procedure doesn't completely fix the distortion (because we'll never know how many wins a 3-4-8 Yokozuna would have received if he were forced to finish the basho), but I think it gets us close to the Standard Ozeki Performance we're trying to discuss.

 

Edited by Yamanashi
Frustration

Share this post


Link to post
Share on other sites

I don't have the vast knowledge and I'm not into statistics as much as many of you. For me, however, there are indeed two answers to the question. 

1. Acceptable is what retains rank. Scoring a three tournament average of 11 with joi schedule allows you a nice privilege of going 8 - 22 in two consecutive basho if you time it right. It's the kiyokai's definition of acceptable Ozeki scores, although if it was to happen, I'm 100 % sure those in power would be very displeased with this hypothetical Ozeki. 

2. I like my Ozeki to be able to push Yokozunae to the limit, win an occasional yusho or Jun-Yusho. I personally consider strong Ozeki to be the backbone of Makuuchi in general. I don't expect them to be in Double digits all the time, but a convincing Ozeki will take 10+ more often as a rule and not as an exception. 

I guess it is quite often, that Ozeki seem comparable in strength to a Yokozuna in the fans eye. If they are hit with injury and illness, they loose said strength. They can't however take repeated kyujo breaks and often come back way too early to avoid demotion.

I know it's Sumo and everyone is injured, but Ozeki is the level at which you have a) no directly result-related banzuke movement, b) constantly facing a mix of high calibre (other O and the Y) and top form competitors ("junior" sanyaku and top Maegashira, often recently promoted) and/or declining former top competitors aswell as c) having one tournament protection only from demotion and therefore pressure to complete. Tochinoshin looked insanely strong with his 14-10-13 last year. Went 5-0 in his first tournament as an Ozeki, got his loss against Tamawashi plus injury.

A Maegashira would have had a significant drop, and a probably very different set opponents after falling six to seven ranks. If he would be able to compete, he would score really good at those ranks, with lower skilled, past their prime or less experienced opponents. The Ozeki faces all top guys again, may eek out a 8-7, and might drag further injury with him. Rinse and repeat. It is quite the achievement to stay at Ozeki for a few years. Goeido can be frustrating and I honestly think Takayasu had more of an impact despite bad luck with injuries, flu etc. Still Goeido is a strong competitor who stayed at a very high rank for a few years now and won a Yusho at his rank. 

Sorry for my rambling. I guess my tl;dr: would be: It is very hard in a sport like sumo to score KK in joi and most long time Ozeki manage that, which makes them acceptable to me. Although I'd like them to do a little more. 

I hope I got something across. 

 

  • Like 2

Share this post


Link to post
Share on other sites

Not a direct response to @Thorbjarn  , rather another general asessment:

After thinking and clicking around a bit more, I realize that the generalization of the original question of this thread is a problem.

First of all, at which (arbitrarily chosen) period are we looking exactly? All-time doesn't really make sense. 15 day era? 6bpy era? Last 15 years?

The reason why the era somewhat matters, is that at certain times it is almost unconceivable for an Ozeki to regularly push the Yokozuna to the limit and keep the yusho race open, because the respective Yokozuna of the era is/are so dominant. The Ozeki will either fall short of such a virtual expectation or be bound to be a future Yokozuna himself. In the other direction, the quality of the regular joi'i matters as well. If we have a couple of Wakanosatos, Akinoshimas or Aminishikis hanging around, your Ozeki scores cannot thrive as much as with a handful of Iwakiyamas, Takamisakaris and late-career Miyabiyamas (no offense to any of them; loved them).

Secondly, you can surely distinguish types of Ozeki. There were cases when rikishi were promoted (partly) for banzuke balance (i.e. when there was no or almost no Ozeki left. Look at Kaiketsu (1st stint) and Kotokaze for instance. Personally, I wouldn't demand such Ozeki to perform at some average standard, I would rather critisize that they were given the rank in the first place. (In a certain sense, the "political Ozeki" Kotoshogiku, Goeido and [in a twisted way] Kisenosato are a sub-set of this.)

Besides such special cases, at the top end of the spectrum you will find almost-Yokozuna Ozeki, who were either victims of bad injury timing, politics and/or era-related handling of Yokozuna promotion "rules". Half a stratum below are high performing Ozeki without real chances for yokozunahood who are just fit exactly for the rank. Archtypically embodied by Takanohana II, I would submit. Below that are the non-special-case underperformers, guys who got lucky or peaked within a pretty short spell. Their average scores will be most sensitive to the quality of competition; see two paragraphs above.

Finally, Ozeki win averages in general are completely distorted by the special demotion rules for the rank, which -- you knew that -- changed over time (kosho rule in various forms). The current rules encourage group co-operation to retain rank (OBSC) and early kyujo in a basho with early losses. In the most extreme legal version an Ozeki could stay at his rank with a winning average of 4 (alternating between 8-7 and sitting out). Of course that doesn't really happen, but to a certain degree, analyzing actual numbers of average Ozeki performances have to account for the rules bias. (Sidenote: All this is additionally the sole reason, why there was only one guy ever to get to ozekihood twice without using the 10-5 backdoor; everyone else either managed to negotiate the 8-7s or was just simply done.)

 

Conclusion: Ozeki should do what they can, but if they don't, we at least have a lot to jeer and feel superior about. So they provide entertainment either way.

Edited by yorikiried by fate
intro clarification and typos
  • Like 3

Share this post


Link to post
Share on other sites

I forgot: Personal sumo socialisation is a factor when arguing about proper Ozeki performances as well. I started watching properly with Ozekis Hakuho, Kotooshu, Chiyotaikai, Kaio and Tochiazuma on board, so my personal expectations might be an inch or so higher. (I get that Koooshu is a bit of an outlier in that group; then again, even if he'd bottled it time and time again, he was a proper Oezeki in the sense of "what other rank should he realistically be put at?")

Share this post


Link to post
Share on other sites
53 minutes ago, yorikiried by fate said:

Conclusion: Ozeki should do what they can, but if they don't, we at least have a lot to jeer and feel superior about. So they provide entertainment either way.

Amen to that.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this