Promotion/Demotion Discussion

[Yamanashi]

26 minutes ago, Bunbukuchagama said:

He looks so happy...

His sumo is so one-dimensional, I fear he might drop back to Makushita immediately. 

You could be right; I haven't watched a ton of his matches.  OTOH, there are a good number of guys with 1D Sumo who have made a fairly stable career in Juryo

[Bunbukuchagama]

9 minutes ago, Yamanashi said:

You could be right; I haven't watched a ton of his matches.  OTOH, there are a good number of guys with 1D Sumo who have made a fairly stable career in Juryo

I am getting strong Tochimaru vibes from him. 

[Reonito]

29 minutes ago, Bunbukuchagama said:

I am getting strong Tochimaru vibes from him. 

His tsuppari is less comical

[Bunbukuchagama]

1 minute ago, Reonito said:

His tsuppari is less comical

But is it more effective? 

[Katooshu]

I think so, yes. He gets his whole body moving behind his thrusts, whereas Tochimaru is pretty much all arms and has more trouble budging opponents in high makushita than Nabatame does. Nabatame also reached juryo far faster than Tochimaru (4 years vs 11), so that's a good sign for him in that comparison. 

He could be sent back to makushita quickly, but overall I think he'll spend much longer as a sekitori than Tochimaru has.

Edited May 30, 2024 by Katooshu

[Reonito]

1 hour ago, Bunbukuchagama said:

But is it more effective? 

looks like about 27% of juryo debuts lead to an immediate demotion back to makushita...

[Bunbukuchagama]

1 hour ago, Katooshu said:

I think so, yes. He gets his whole body moving behind his thrusts, whereas Tochimaru is pretty much all arms and has more trouble budging opponents in high makushita than Nabatame does. Nabatame also reached juryo more than twice as fast as Tochimaru (4 years vs 11), so that's a good sign for him in that comparison. 

He could be sent back to makushita quickly, but overall I think he'll spend much longer as a sekitori than Tochimaru has.

So, where do you place him on Takakeisho-to-Tochimaru scale? Bushozan?

[Katooshu]

That sounds about right. 

[Reonito]

He's still young enough and new enough to ozumo to potentially improve ... there's space between "ozeki" and "M-J yoyo" (disclaimer: this is a generic observation, as I haven't paid a lot of attention to his sumo)

[Bunbukuchagama]

17 minutes ago, Reonito said:

He's still young enough and new enough to ozumo to potentially improve ... there's space between "ozeki" and "M-J yoyo" (disclaimer: this is a generic observation, as I haven't paid a lot of attention to his sumo)

I just don't see any versatility in his sumo. A narrow specialist needs to be really good at what he does in order to succeed. 

[Tigerboy1966]

Nabatame's kimarite stats are certainly not indicative of versatility. 104 bouts and one win by yorikiri. And I bet that if we check the replay on that one there would have been a shove involved at some point.

[Churaumi]

Hopefully being a sekitori will give him better access to higher-end training partners. There was one of those Futagoyama videos a few months ago where the cameraman asked Roga why he wasn't around the stable much and he said he practices with other sekitori in the morning and schmoozes with supporters in the evenings. 

[Asashosakari]

Juryo and makushita-joi guess:

Tokihayate (M15e 6-9)     J1   Onokatsu (J12e 13-2)
Takerufuji (M6e 0-0-15)   J2   Daiamami (J1e 6-9)
Kitanowaka (J5w 8-5-2)    J3   Mitoryu (M13e 2-9-4)
Shimanoumi (J11e 9-6)     J4   Asakoryu (J4w 7-8)
Tohakuryu (J4e 6-9)       J5   Daishoho (J2w 5-10)
Tsurugisho (M17e 3-12)    J6   Tamashoho (J9w 8-7)
Shiden (J10e 8-7)         J7   Tomokaze (M16e 2-13)
Shirokuma (J6e 6-9)       J8   Myogiryu (J1w 3-12)
Hidenoumi (J8e 6-9)       J9   Chiyosakae (J9e 6-9)
Shishi (J7e 5-10)         J10  Shimazuumi (J7w 5-8-2)
Tochitaikai (J14e 8-7)    J11  Aoiyama (J11w 7-8)
Hakuyozan (J10w 6-9)      J12  Tsushimanada (J12w 7-8)
Hakuoho (J8w 5-6-4)       J13  Kayo (Ms1w 5-2)
Fujiseiun (Ms11e 7-0 Y)   J14  Nabatame (Ms2w 5-2)

They could put Fujiseiun pretty much anywhere from J11w to J14w and none of those spots would really surprise me.

Kiryuko (Ms3e 5-2)        Ms1  Kazekeno (J14w 7-8)
Daiseizan (Ms3w 4-3)      Ms2  Kitaharima (Ms4e 4-3)
Kotokuzan (Ms4w 4-3)      Ms3  Oshoumi (J13e 5-10)
Otsuji (Ms7e 5-2)         Ms4  Yoshii (Ms7w 5-2)
Wakaikari (Ms6w 4-3)      Ms5  Satorufuji (Ms14e 6-1)
Akua (Ms1e 3-4)           Ms6  Chiyomaru (J13w 4-11)
Wakanosho (Ms20e 6-1)     Ms7  Asahakuryu (Ms10w 4-3)
Daishomaru (Ms5e 3-4)     Ms8  Kototebakari (Ms2e 2-5)
Dewanoryu (Ms18w 5-2)     Ms9  Kaisho (Ms14w 4-3)
Tenshoho (Ms6e 3-3-1)     Ms10 Mineyaiba (Ms15e 4-3)
Obara (Ms15w 4-3)         Ms11 Kotokenryu (Ms16e 4-3)
Nihonyanagi (Ms32w 6-1)   Ms12 Fukai (Ms23e 5-2)
Kitadaichi (Ms8w 3-4)     Ms13 Fujitoshi (Ms24e 5-2)
Tokunomusashi (Ms5w 2-5)  Ms14 Kamito (Ms36e 6-1)
Asanowaka (Ms20w 4-3)     Ms15 Kotoyusho (Ms26e 5-2)

I may have been too nice to Oshoumi, but "between the top 5 KKs and the non-top 5 KKs" felt like a good place in that tight mess. The section decompressed rapidly after Ms6, so the subsequent movements are actually fairly generous. Hard luck for Akua and Chiyomaru if this comes to pass, but as discussed earlier there have to be two losers in the 10-slot sweepstakes here.

Next ones not in:

Ms40e 6-1
Ms32e 5-2
Ms23w 4-3
Ms12e 3-4
Ms9e 2-5

Maaaybe the 6-1 (Aonishiki) goes into the joi instead of either Tokunomusashi, Asanowaka or Kotoyusho, the rest shouldn't stand a chance.

Edited June 2, 2024 by Asashosakari

[itchyknee]

On 26/05/2024 at 18:45, Reonito said:

No one's ever had consecutive yusho without being ranked ozeki at least for the second basho, period. Futabayama and Terunofuji are the only ones to win consecutive yusho at S/O.

To be fair the DB doesn't record yushos before Natsu 1909. According to Wikipedia, Raiden won as a sekiwake and then came back to win again as a komosubi. Crazy times in ranking, so it makes sense to not record them in terms of metrics.

[Kachikoshi]

On 02/06/2024 at 00:38, Asashosakari said:(Ms3e 5-2)        Ms1  Kazekeno (J14w 7-8)
Daiseizan (Ms3w 4-3)      Ms2  Kitaharima (Ms4e 4-3)
Kotokuzan (Ms4w 4-3)      Ms3  Oshoumi (J13e 5-10)
Otsuji (Ms7e 5-2)         Ms4  Yoshii (Ms7w 5-2)
Wakaikari (Ms6w 4-3)      Ms5  Satorufuji (Ms14e 6-1)
Akua (Ms1e 3-4)           Ms6  Chiyomaru (J13w 4-11)
Wakanosho (Ms20e 6-1)     Ms7  Asahakuryu (Ms10w 4-3)
Daishomaru (Ms5e 3-4)     Ms8  Kototebakari (Ms2e 2-5)
Dewanoryu (Ms18w 5-2)     Ms9  Kaisho (Ms14w 4-3)
Tenshoho (Ms6e 3-3-1)     Ms10 Mineyaiba (Ms15e 4-3)
Obara (Ms15w 4-3)         Ms11 Kotokenryu (Ms16e 4-3)
Nihonyanagi (Ms32w 6-1)   Ms12 Fukai (Ms23e 5-2)
Kitadaichi (Ms8w 3-4)     Ms13 Fujitoshi (Ms24e 5-2)
Tokunomusashi (Ms5w 2-5)  Ms14 Kamito (Ms36e 6-1)
Asanowaka (Ms20w 4-3)     Ms15 Kotoyusho (Ms26e 5-2)

This is a bit off topic, but is there a “formula” for the Makushita joi, like how in Makuuchi/Juryo you move up/down by wins minus losses? The Makushita joi rankings have always seemed way more arbitrary to me.

[Reonito]

2 hours ago, itchyknee said:

To be fair the DB doesn't record yushos before Natsu 1909. According to Wikipedia, Raiden won as a sekiwake and then came back to win again as a komosubi. Crazy times in ranking, so it makes sense to not record them in terms of metrics.

Yeah 1909 is generally seen as the first recognized yusho, and the JSA only adopted the system in 1926, so tournament wins before than were a lot less formal, as of course were the rankings.

[Gurowake]

3 hours ago, Kachikoshi said:

This is a bit off topic, but is there a “formula” for the Makushita joi, like how in Makuuchi/Juryo you move up/down by wins minus losses? The Makushita joi rankings have always seemed way more arbitrary to me.

The only thing you can really do is look at what they've done historically and try to generalize from that.  There's clearly no formula of any sort for lower division movements because there are some choices of relatively ranking that would be impossible to be obtained if there were a formula that was consistently applied and fit with the observations.  They'll obviously follow the general required principles of banzuke movement, but the relative placement of different groups that could go in a wide number of possibilities is seemingly arbitrary and probably decided by one person who has to finish it in an hour and only cares if it's consistent.

[Bunbukuchagama]

2 hours ago, Gurowake said:

probably decided by one person who has to finish it in an hour and only cares if it's consistent.

I am still convinced a dartboard and generous amounts of alcohol are involved.

[Yamanashi]

2 hours ago, Gurowake said:

The only thing you can really do is look at what they've done historically and try to generalize from that.  There's clearly no formula of any sort for lower division movements because there are some choices of relatively ranking that would be impossible to be obtained if there were a formula that was consistently applied and fit with the observations.  They'll obviously follow the general required principles of banzuke movement, but the relative placement of different groups that could go in a wide number of possibilities is seemingly arbitrary and probably decided by one person who has to finish it in an hour and only cares if it's consistent.

My guess is that because of the paywall (IYKWIM) at the top of Makushita, the joi may be the only place where there's any consistency.  If a rikishi get a 2-5 record at, say, Sd30, he could end up anywhere.  When I did a survey of men who received 4-3 marks at Ms20 (N = 207), the results were pretty similar: the ± 1 SD spanned 3.5 to 6.5 ranks.  The same study at Sd30 (N = 192) gave a span of 8.5 to 18.5 ranks.

On a side note, three cases gave a rank drop with a 4-3 record, two from Sd30 and one from Ms20.  Not to go all conspiracy theory here, but all three were in ... 1967.

[Reonito]

41 minutes ago, Yamanashi said:

On a side note, three cases gave a rank drop with a 4-3 record, two from Sd30 and one from Ms20.  Not to go all conspiracy theory here, but all three were in ... 1967.

Looks like Makuuchi shrank from 40 to 34 that year, and Juryo from 36 to 26?

[Asashosakari]

1 hour ago, Yamanashi said:

My guess is that because of the paywall (IYKWIM) at the top of Makushita, the joi may be the only place where there's any consistency.  If a rikishi get a 2-5 record at, say, Sd30, he could end up anywhere.  When I did a survey of men who received 4-3 marks at Ms20 (N = 207), the results were pretty similar: the ± 1 SD spanned 3.5 to 6.5 ranks.  The same study at Sd30 (N = 192) gave a span of 8.5 to 18.5 ranks.

Trying to average things out across all 60+ years that they've had seven-bout schedules for the lower makushita is pretty useless because they've clearly not always used the same overall approach. (6-1's in particular were upgraded in value multiple times, rightfully so I will add.) Doing it across just the last, say, 20 years will show things to be quite a bit more consistent.

In general, the current approach appears to be to try to have relatively fixed promotion sizes and just slot the makekoshi records into the resulting gaps. Where those gaps are depends mainly on how many rikishi have retired or gone 0-0-7, and consequently the available banzuke luck gets parcelled out primarily to the MK guys today. It becomes more noticeable the lower you go, because the space created by retirements naturally accumulates and the number of kyujo tends to be larger further down. See e.g. Jd50 4-3 vs Jd50 3-4.

That also means it's actually the opposite of what you stated; makushita-joi is the area with the least consistency, because it's the only one where a) it's not always feasible to create room for the kachikoshi by just demoting the makekoshi more, and conversely b) sometimes you get gaps that simply can't be filled with underdemoted MKs because there just aren't any yet. Sure, the standard deviations might be small anyway, but that's the wrong measure; they'd need to be normalized against the mean promotion/demotion sizes to be meaningfully comparable with what goes on in other sections of the banzuke.

That being said, sometimes a single banzuke will come out that has some KK rank movements that go against everything established, and then the next basho it goes right back to "normal". I'm fairly sure they use some sort of table of expected promotion sizes these days, but it's clear that there isn't a huge amount of quality control that goes into the whole process. They also don't appear to go back and fix errors properly, in the sense that the post-fix state should look as though the error wasn't made in the first place. So you get things like Daiyusho going Sd34w -> Jd5e -> Jd76w on successive 0-0-7's, when it should have separately been Sd34w -> Jd4w and Jd5e -> Jd75w. That is almost certainly happening because they initially forgot to assign rikishi to spots above him, and when the omissions were discovered, they just shoved all the already-assigned rikishi down, including Daiyusho, rather than rework things around the nominally fixed 0-0-7 demotion spots properly.
 

Quote

On a side note, three cases gave a rank drop with a 4-3 record, two from Sd30 and one from Ms20.  Not to go all conspiracy theory here, but all three were in ... 1967.

I mean, yeah, that's what happens when 8 sekitori ranks and 38 makushita ranks are eliminated. All the lower division rank movements are proper if you virtually re-add those ranks, e.g. Ms20e 4-3 -> Ms22w would have been Ms20e -> Ms14w and Sd20e 4-3 -> Sd51e would have been Sd20e -> Sd5e without the banzuke restructuring, so it's clear that the new banzuke was (special exceptions in low juryo aside) created normally and then just shifted down as needed. Another piece of evidence why blindly averaging all of sumo history frequently doesn't make much sense.

(Side note: The latest unannounced framework change is a different handling of zero-win records in jonidan, seemingly in effect since the post-Aki 2023 banzuke session.)

Edited June 4, 2024 by Asashosakari

[Yamanashi]

40 minutes ago, Reonito said:

Looks like Makuuchi shrank from 40 to 34 that year, and Juryo from 36 to 26?

Yes, there's been some discussion about strange goings on in 1967-68.  Makushita dropped from ~100 to 60 and Jonidan grew from 92 to ~130 between March and May 1967; Sandanme stayed about the same.

[Asashosakari]

14 hours ago, Asashosakari said:

In general, the current approach appears to be to try to have relatively fixed promotion sizes and just slot the makekoshi records into the resulting gaps. Where those gaps are depends mainly on how many rikishi have retired or gone 0-0-7, and consequently the available banzuke luck gets parcelled out primarily to the MK guys today. It becomes more noticeable the lower you go, because the space created by retirements naturally accumulates and the number of kyujo tends to be larger further down. See e.g. Jd50 4-3 vs Jd50 3-4.

I'm going to expand a bit on this by going into detail about why the demotions can become so much more volatile, although as I'm starting to write this I'm not entirely sure if I'm not just stating the bleedingly obvious, so if I am, just roll your eyes and ignore. 

If you have two basho with banzuke of the same size and both tournaments have the same number of new rikishi coming in from maezumo, but one basho sees 10 rikishi retire afterwards and the other has 20 retirements, naturally the follow-up rankings will differ in length by 5 ranks. The first-glance impression of what that means in practice may be, "everybody's rank movements will be displaced by at most five ranks", with almost no differences in high makushita (because there are no or very few retirements above that in either case), gradually expanding to five ranks of difference in jonokuchi. But in actual fact, the different treatment of KK and MK rikishi has a sort of magnifying effect.

Take the example of a rikishi, let's call him Makenohana, who just went 3-4 at Sd85. In an average basho he can perhaps expect to be demoted to Jd15 after that, going down 20 ranks. That average basho will have something like this: 5 retirements from makuuchi to upper sandanme (i.e. rikishi who would have stayed in front of Makenohana if they weren't leaving), and another 5 rikishi who were kyujo somewhere in mid to low sandanme (i.e. rikishi who will now fall below Jd15, but would have been above it with any real tournament result, 0-7 excepted of course).

Now, let's instead assume that none of those retirements and kyujo happen. There are now 10 additional rikishi who need to find space ahead of Makenohana, so instead of Jd15 he needs to be demoted to Jd20. But...KK records are getting fixed promotions, so about half of that Jd16-20 zone is already occupied by KK rikishi, regardless of the change from 10 intai/kyujo to zero. These are rikishi who were going to be ranked behind Makenohana after an average basho, but now they need to be ahead of him, too, so it's no longer just 10 rikishi pushing Makenohana down, but around 15 or 16...and he's down to Jd23. And of course these latest three ranks that Makenohana has lost also have KK rikishi present already, so he's losing another handful of spots to accommodate those guys, and so on. When all is said and done, he probably ends up around Jd25 or 26. The added presence of 10 active rikishi has caused the size of Makenohana's demotion to increase by 20+ spots.

The same goes the other way, too...more intai/kyujo than average, and Makenohana's demotion gets lessened disproportionately because KKs who would have jumped ahead of him no longer do.

That's all highly idealized, of course. When things get really extreme, they'll adjust the KK movements to some degree, too, especially when there's a lot of open space to fill; less so in the opposite scenario with a severe lack of space, because they really do appear to be interested in rewarding KKs with comparable minimum promotions on every banzuke.

[itchyknee]

12 hours ago, Asashosakari said:

Take the example of a rikishi, let's call him Makenohana, who just went 3-4 at Sd85.

The closet most literal matches I can find are:

Nakanohana 1965.01 Sd82w 3-4 -> 1965.03 Jd2e 6-1 / Which was a drop of 33 spots (factoring in a removal of two Ms spots), with 14 retirements above his position.

A Takanohana 1964.03 Sd89e 3-4 -> 1964.05 Jd4w 2-5 / Which was a drop of 18 spots (factoring in a removal of one Ms spot), with 5 retirements above his position.

With special mention to Ikenohana for having "kenohana" in his shikona and dropping 62 spots with a 0-7 and 15 retirements above his position.

1969.07 Sd80w 0-7 -> 1969.09 Jd11w 4-3

(closest thing I could find to a dunce cap for this totally pointless post)

[Kachikoshi]

23 hours ago, Asashosakari said:

I'm going to expand a bit on this by going into detail about why the demotions can become so much more volatile, although as I'm starting to write this I'm not entirely sure if I'm not just stating the bleedingly obvious, so if I am, just roll your eyes and ignore. 

This was not "bleedingly obvious" and was, in fact, rather observant and informative. Thanks!