Miguel is also a tremendous thinker on combat sports who talks about sumo not infrequently if you don’t follow him already!
English
Ozeki Analytics
480 posts
Ozeki Analytics
@OzekiAnalytics
Data driven sumo writing
国技館 Katılım Ocak 2024
15 Takip Edilen56 Takipçiler
Great thread on sumo development. I often bang on about the only American in pro sumo, Hikarumusashi, and these are some tough truths for boosters like myself as he got a later start
Miguel Class@MigClass
Going to hijack this to teach you all about sumo by going through the current top ranked rikishi. Onosato (Yokozuna): started sumo in first grade. Hoshoryu (Yokozuna): scouted and recruited into sumo after middle school. Aonishiki (Ozeki): started sumo at 7 years old. 2/6
English
@SumoZeki As in machine learning algorithms on reaction times, first hit, etc
English
@SumoZeki That’s one of my dreams is getting a massive archive of matches historic and modern for fun first off but I would actually put in the work to see if there could be algorithmic classification because there’s definitely stuff we could learn from that
English
on Reddit got some heat for classifying Akebono as a “Balanced” wrestler rather than “Pusher Thruster” which I was worried about too when I saw it initially but I verified he actually had 156 Oshidashi and 154 Yorikiri
GIF
English
@SumoZeki Yeah it’s a bit unfortunate there’s no existing classification so I could reverse engineer “pusher thrusters win this way but belt grapplers that way”
Someone was commenting similarly on the post on substack how it’d be nice to classify their attempted strategy
English
@OzekiAnalytics There's no great way other than to watch all the sumo, but it might help to look at which kimarite more likely come from which style. I'd say 7 of his top 10 were usually the result of his oshizumo.
English
@SumoZeki Helpful! You’re someone I trust on this stuff so any idea how I could maybe look at it differently from a data perspective? Definitely not acting like my analysis is the end all but rather hopefully part of the picture
English
@OzekiAnalytics A lot can happen in a match and still have it end in yorikiri. You can push all the way to the edge, but if your torsos connect in the final moment, that's yorikiri. Doesn't necessarily mean you're a yotsu expert.
English
@SumoZeki Yeah it’d be another whole project as I was looking at aggregate kimarite but I was wondering if/hoping to maybe see if sumo style evolves over time
English
@OzekiAnalytics He went through a period in the middle of his career where he focused on grabbing the belt and winning by yorikiri. He was less successful during this time, but came back to his thrusting in the 2000s and won another two yūshō.
English
Average engagement by platform
Twitter: a few likes
Substack: all positive comments
Reddit: get regularly flamed 😅
Funny it works out that way and certainly not what you’d expect from general discourse
Ozeki Analytics@OzekiAnalytics
on Reddit got some heat for classifying Akebono as a “Balanced” wrestler rather than “Pusher Thruster” which I was worried about too when I saw it initially but I verified he actually had 156 Oshidashi and 154 Yorikiri
English
Hope everyone enjoys!
ozekianalytics.substack.com/p/classifying-…
English
Article is up! Using all kimarite and ~2000 wrestlers, we establish wrestler archetypes. There’s tons of info in there including most common kimarite by archetype, how the Makuuchi in January were classified and how the most recent Yokozuna were classified
Ozeki Analytics@OzekiAnalytics
New article with research, graphs and (entirely new) data is complete. Expect it to drop this Tuesday 🙏🤝
English
New article with research, graphs and (entirely new) data is complete. Expect it to drop this Tuesday 🙏🤝
English
It’s being written up but very excited
Ozeki Analytics@OzekiAnalytics
Very excited about my next article. It’s entirely new research on sumo I believe. Lots of data.
English
Ozeki Analytics retweetledi
横綱 白鵬関・鶴竜関によるシューティングセレモニー⛹️♂️🔥
鶴竜関 見事3Pシュート成功😳😳
📺NHK総合で生放送中
📱バスケットLIVEでライブ配信中
#JAPANMADNESS
#AkatsukiFive
日本語
Over 100k rows in the dataset I’m building for this one. I’m really excited for this one and I also already know the Reddit commenters will have some issues with it 😅
Ozeki Analytics@OzekiAnalytics
Very excited about my next article. It’s entirely new research on sumo I believe. Lots of data.
English
Very excited about my next article. It’s entirely new research on sumo I believe. Lots of data.
English
Hikarumusashi - the only American in pro sumo - falls to 1-4 guaranteeing a Makekoshi (losing record). It’s been a tough tournament for him.
Ozeki Analytics@OzekiAnalytics
Another 2 matches but fortunately a win. The American stands at 1-3.
English
Another 2 matches but fortunately a win. The American stands at 1-3.
Ozeki Analytics@OzekiAnalytics
Bad news for fans of the only American in pro sumo wrestling right now: Hikarumusashi started 0 wins 2 losses. He’s higher than ever before but still a bummer he hasn’t gotten a win yet. Still plenty of time to turn this tournament around.
English
Bad news for fans of the only American in pro sumo wrestling right now: Hikarumusashi started 0 wins 2 losses. He’s higher than ever before but still a bummer he hasn’t gotten a win yet. Still plenty of time to turn this tournament around.
English
@SumoZeki Can’t believe yours is the only tweet mentioning the Kainahineri (in English). That was crazy! Murray did a good job pointing out rarity!
English