With the last of the schools wrapping up their seasons on Saturday, wrestlers are left wondering, "What if the Far East schools in Japan - from Edgren in the north to Kubasaki in the south - had wrestled a pretty-much-normal season, complete with a season ending tournament, like their US and Japanese high school counterparts? Who would win the Far East?" That, of course, is unknowable because the Far East never happened. But what if it had? Well, let's find out! If we compile the results printed in the Stars and Stripes, along with some extra information provided by coaches, we can do as ESPN did for the canceled 2020 March Madness tournament: simulate it. The military does simulations frequently to test out different strategies, to measure the effect of new technologies, and even to see which fighter jet would win in one-on-one encounters (i.e. dogfights). The Koku-Jieitai (Japan Air Self-Defense Force) has a schoolhouse dedicated to professional military education, complete with experts who can recreate just about any situation or scenario. A few f those experts from Meguro Base (Tokyo) agreed lend their talents to do up a simulated Far East, with the results published here. Each of the wrestlers was converted into a fighter jet, with their stats (W/L, scores, pins, techs, times, 4-pt takedowns, etc.) representing various aircraft capabilities. For some wrestlers there was less "intelligence" available than for others. That can be treated as a new fighter jet that has seen some action (i.e. win/loss), but its full capabilities are unknown (pin? close, one-point win on a controversial call? great on takedowns, but can't pin? etcetera, etcetera). Running 10,000 scenarios is a common practice that can reveal the most likely outcomes. However, doing that with this virtual Far East would likely produce something fairly obvious like, "Kadena's Kurt Chapman takes 1st at 158," so we did essentially the opposite. There were 16 scenarios for each individual match, and one of those was selected at random. For example, if Chapman had a 95% chance of winning a certain matchup, 15 of the 16 scenarios would have that outcome (12 pins, 2 tech falls, and 1 decision, to be exact). But, every match starts 0-0, and upsets do happen. In order to do the whole tournament in one run (instead of one round of matchups, then a new set of opponents and a new round of matches, repeated until all places are determined), all wrestlers were pitted against all others. The brackets were then completed based on those results. That methodology produced many unusable results, but some were pretty fun, like Chapman (7-0 at 158) vs his teammate Mizuki Sato-Marsh (3-0 at 101). Unfortunately, this methodology also made it uninteresting to do a dual meet tournament; there is only one head-to-head result selected from the 16 possibilities, so it'd be the same result in a dual meet as in the individual tournament. Because the season was so stilted and shortened (or lengthened), some imperfect information was intentionally added. This included data from modified dual meets (e.g. "takedowns only" CAJ vs St. Mary's dual meet on 2 Dec), non-DODEA/KPASS sanctioned competition (e.g. Azabu Open), and a small amount of results from two years ago. On the other hand, the Korean schools were excluded, because there were not enough inter-school data points with the main body of data in Japan. The Rumble on the Rock, Beast of the East, and Far East results from years past provide a little, but that data is old (i.e. less reflective of current capabilities). This year's Azabu Ward Open provided a wealth of data that the supercomputers could use, so the Okinawa school and the international school data pools are sufficiently linked to the main pool of DODEA-Pac East schools. Lineups were either provided by the coaches, the wrestlers themselves, or determined by the last competition for each team, as published in Stars and Stripes. Any holes in a team's lineup at that point were filled to the extent possible at our discretion. That may have created some unlikely or awkward lineups, but artificially inserting a person with some wrestling experience during the season is more interesting than a bye. Well, that's as much explanation as there will be, so just sit back and enjoy (or argue about) the results! The first rounds will come out on Monday, 21 Feb.
0 Comments
Leave a Reply. |
Mustang MikeSend a note to [email protected] with ideas for a story or corrections/additions to these write-ups. Archives
February 2022
Categories |