- Elo Rankings: The Prediction Bracket
- T8M Elo Rankings – 4/6/17
- Unpopular Opinion: US Quidditch Cup 10
- March Madness: The Championship Showdown
- Hats Off to Thee: The History of Minnesota Quidditch
- March Madness: Sweet 16 Recap
- March Madness: Round of 32 Part Two Recap
- March Madness: Play-In Round Recap
T8M Elo Rankings – 4/6/17
As the quidditch season comes to its culmination, we would like to present our final pre-US Quidditch Cup 10 Elo ratings of US Quidditch Cup-qualified teams.
These ratings use data going back to Fall 2012 and follow methodology stated here. Elo ratings, unlike The Eighth Man rankings and USQ’s standings, seek to give us quantifiable values that evaluate the relative strength of two teams. Using these, we are able to observe the relative strength of pools, teams within pots and then look at those pots and predict what a team’s likely outcome from that pot will be. One large caveat with these ratings is that first-year teams, specifically first-year teams that have played a bare minimum amount of official games like the BosNYan Bearsharks, will have under-inflated Elo ratings, due to the fact that they are given an initial Elo rating of 1300 to begin the year. This is why BosNYan, Lake Erie Elite, Carolina Heat and the Richmond Ravens (the only first-year teams to make nationals this year) all have the lowest Elo ratings of any team within their pots.
Pools of Life and Pools of Death
A simple way to quickly look at the relative difficulty of each pool is to look at the average Elo rating of the teams in that pool. Using this metric, it’s obvious that Pools One, 10 and 11 all stand as relative pools of death, having an average Elo rating of above 1700. A quick glance at team reputations would certainly make Pool 11 seem to be the “Pool of Death”, but looking at the averages, it actually falls slightly behind Pool 10. This difference lies primarily in the pools’ pot five teams. Pool 10 has Baylor University, the second highest-ranked pot five team, who will certainly be a more significant confounding factor for their pool than Pool 11’s Illinois State University, the second lowest-ranked pot five team. Even still, Baylor only has slightly less than a one-third chance of making it to day two. If you were to remove the pot five teams from the consideration, Pool 11 is clearly the pool of death, and it isn’t close. Without pot five teams, Pool 11’s average Elo rating is 1789, while the second closest, Pool One’s, is a mere 1750. While Illinois State may have a long shot of making a splash, the fight between Texas Cavalry, The Warriors, Michigan Quidditch and Tufts University in Pool 11 to see who will make day two will certainly be a thrilling one to watch.
While the high average Elo ratings of these pools of death may predict intense games to watch, it is hard to say the opposite about the pools with the lowest Elo ratings. While Pools Two, Three and Eight all have average Elo ratings of under 1650, only one of these, Pool Eight, seems to be a “Pool of Life” for its predicted winner, Lone Star Quidditch Club. Pool Two suffers from having a highly under-inflated BosNYan team that stands as a wildcard. Whether they are able to give Bowling Green a competitive game or end up playing to their current rating, Pool Two will either be hotly contested at the top or feature a competitive three way race between BosNyan, UC Berkeley and University of Virginia for who can make it out of the pool. Likewise, Pool Three, while having a relatively low average Elo rating, has four teams that lie relatively close to each other in Elo ratings. This closeness in ratings means that it will be unlikely for any one team to run away with the pool and the fight to go 3-1 or 2-2 should be hotly contested with bracket seeding on the line. Pool Three is one of two pools (along with Pool One) where four teams have greater than a 20 percent chance of going 2-2. A three-way tie leading to heartbreak for one team that may miss bracket is not an unlikely chance in Pool Three, and despite its low average Elo rating, it will certainly be worth having on your radar late in day one for the upsets that may ensue.
The Hunt for Day Two
Throughout pool play, it is likely that most spectators’s focus will be on what teams will make it to bracket play. Like at last year’s tournament, the process of narrowing down 60 teams to a 36-team bracket means that most teams that can win at least two games should have a lock on a spot in bracket play. Still, there is some chance that a team that goes 2-2 may miss bracket play and a much smaller chance that a team that only wins one game still makes it to day two (due to the potential for rare three way ties in multiple pools).
To give us an analytical perspective on this, we can simulate every game day one using each team’s current Elo rating, assign a likelihood of each team winning each game and predict a team’s odds of going 4-0, 3-1, 2-2, 1-3 or 0-4 in pool play. Full odds for each team’s chances can be found here. From these odds, we can estimate a team’s likelihood of going 2-2 or better and thus have a rough approximation for that team’s odds of making it to bracket play. It should be noted that because there are more potential situations where a team goes 2-2 and does not make bracket play than there are situations where a team goes 1-3 and does make bracket play, that the chance of a team going at least 2-2 is a slightly-inflated approximation of whether that team will make bracket play. However, because those situations are relatively rare, that inflation is minimal (usually only by a percentage point or two), and thus a team’s odds of going 2-2 or better serve as a useful stand-in to predict their chances of making it to bracket play.
Using this metric, the team with the highest chance of making it to day two is Quidditch Club Boston, with greater than a 99 percent chance of doing so. There are 16 teams that have over a 90 percent chance of going 2-2 or better and–barring any major upsets–should lock in an appearance on day two. Outside of those 16, however, there are 37 teams with greater than a 20 percent chance of going 2-2 and they will be fighting hard to seal up one of those other 20 seeds in the bracket. Even still, 90 percent is no lock and with 16 teams predicted to make day two, 90 percent of the time or greater, simple probability predicts that at least one of these “favorites” may end up seeded outside the top 36. No doubt a crushing event for that team, but in a tournament as large as this, such a cruel turn can all but be expected. As I write this, Mississippi State’s women’s basketball team has just done the unexpected and snapped UConn’s 111-game winning streak to proceed to the NCAA championship. While 90 percent (or in UConn’s case, 87 percent) odds may feel like a sure thing in sports, one in 10 times, they won’t be. When you add in the snitch as a factor, which keeps any game within three scores more or less a toss-up, what may seem like certain odds in quidditch have an easy way of being turned around in an instance, making upsets, suspense and excitement a fact of the game.
About Joshua Mansfield
A native of Cincinnati, Ohio, Joshua Mansfield began playing quidditch when he founded the Tulane University team in 2013. He currently plays for Gulf Coast Gumbeaux and serves as the Head of Statistics for Major League Quidditch. Additionally, he is the third-largest consumer of cilantro in the greater New Orleans area.
- RT @BarcelosKnows: i may have gotten scooped by @The_EighthMan but I'm posting this anyway. https://t.co/is1e0ad1Fd
about 2 days ago
- Guest writer Harry Greenhouse gives his thoughts on how community players can assist with collegiate quidditch.
about 2 days ago
- When @davidfoxxrocks wants to end a game, he ends the game!
The catch that sent @BOSNYbearsharks to the Final Four… https://t.co/mdBvsuAkYo
about 1 week ago