Ultiworld Ultiworld DG

USAU Approximated Club Rankings: Post Pro-Elite Challenge

[caption id=“attachment_37013” align=“aligncenter” width=“800”]

Nicky Spiva (Truck Stop #36) catches. Photo: Alex Fraser – UltiPhotos.com[/caption]

Ultiworld’s Cody Mills has put together a statistical tool to approximate the USA Ultimate rankings by using publicly available score data and applying the algorithm that USA Ultimate uses to run their numbers. While the results below are unofficial and could include some scores which are liable to be altered or removed from the official rankings, they are a good representation of the what the current rankings would look like.

Throughout the remainder of the club season, we will be running this USAU ranking approximation and, later, possibly applying different algorithms and/or ranking methods to test alternatives to the existing algorithm.

For more information about the USA Ultimate algorithm, ranking system, and some of its perceived flaws, see this FAQ from last year.

Men’s Division

The rankings below include only teams with at least five sanctioned games.

[table id=635 /]

Implied Bid Allocation

Great Lakes: 2 bids
Mid-Atlantic: 2 bids
Northeast: 3 bids
Northwest: 2 bids (+1)
North Central: 2 bids
South Central: 2 bids (+1)
Southeast: 2 bids (-1)
Southwest: 1 bid (-1)

Discussion

Very solid season opening campaigns from Truck Stop and Doublewide land them both squarely in the top 10, with Truck’s high margin wins over top opponents vaulting them straight to the top of the rankings. The bid picture, though, is still very much in flux. A dreadful weekend from Temper means they won’t earn a bid for the Mid-Atlantic, and Patrol has slipped out of the top 16. With Floodwall likely to drop out later this season, the onus will be on Patrol to claim a second bid for the region.

Guerrilla and Condors also have work to do to earn a bid beyond the one that Revolver will get for the Southwest. Guerrilla had their chances this weekend but didn’t come away with enough big wins. The task may fall to the new combo Condors team.

Rhino’s strong weekend only moved them up to the #16 spot, so they will need to play tough at the Pro Flight Finale to ensure they hold a second bid for the Northwest.

Mixed Division

The rankings below include only teams with at least five sanctioned games.

[table id=636 /]

Implied Bid Allocation

Great Lakes: 1 bid
Mid-Atlantic: 3 bids
Northeast: 2 bids (-1)
Northwest: 2 bids
North Central: 3 bids
South Central: 3 bids (+2)
Southeast: 1 bid
Southwest: 1 bid (-1)

Discussion

The two predominant powerhouse regions – the Northeast and Southwest – are staring in the face of a combined three bids. While the regions are down a bit from their usual strength, it is still a scary sight for teams in those regions (particularly the Southwest) who know that many opponents are capable of winning on any given Sunday. There is time left, and teams like the Polar Bears and Blackbird have a chance to play back into the top 16, but the pressure is real.

The South Central can’t quite start celebrating yet, but a three bid outcome seems plausible. shame. – a relatively new team from Fort Collins, CO – is undefeated save for a one point loss to Love Tractor from June, a close loss that is floating them to #13 in these approximated rankings. They also have their 10 games, so they could opt out of future tournaments and possibly land a bid at the end of the season.

The Mid-Atlantic continues to sit pretty with three bids as well, after Alloy held their own this weekend at the Pro-Elite Challenge.

Women's Division

The rankings below include only teams with at least five sanctioned games.

[table id=637 /]

Implied Bid Allocation

Great Lakes: 1 bid
Mid-Atlantic: 1 bid (-1)
Northeast: 3 bids
Northwest: 3 bids
North Central: 1 bid
South Central: 2 bids (+1)
Southeast: 2 bids
Southwest: 3 bids

Discussion

Showdown’s solid weekend in Denver moved them into the top 16 and knocked out Green Means Go, sending a bid from the Mid-Atlantic to the South Central. Otherwise, steady as she goes with plenty of potential drama at the cutoff. Currently, there are six teams within 100 points of one another from #15 to #20. Even second-order effects could cause changes in the final two bid slots.

One worry has to come from the Northwest. Despite Traffic’s solid debut at #7, the region doesn’t pick up a bid because Schwa got beat down and went winless at the Pro-Elite Challenge. Their regional opponents will surely help Portland plays better at the Pro Flight Finale very close to home.

USA Ultimate is expected to release the initial official club rankings later today. Score alterations and ineligible teams can cause some differences between these approximated rankings and the official USAU ones. We will be watching closely!

Originally published at: http://ultiworld.com/2016/08/03/usau-approximated-club-rankings-post-pro-elite-challenge/

Can you include the region that each team is a part of in another column?

@cmills Is there any easy way to calculate uncertainty with this particular algorithm? It would be really interesting to see not just the mean rating, but also some measure of the distribution.

The algorithm is deterministic, the uncertainty in the process comes from the uncertainty of the input set. I suppose you could measure sensitivities of the teams’ ranks to small perturbations in the results, but I’m not sure how well-defined that would be.

Well, a very simple measure would simply be to report the standard deviation of the ranking points for each team (since it is a set of games that are averaged together once those averages have converged). A histogram would be even better, but there might not be enough games per team for that to really be meaningful.

Ah, okay I see what you mean now. I guess it’s a weighted average, so you’d want to include the weights in the histogram/account for them in the std dev. Now I want to see this too; having the set of game values for each of these teams would be nice. I suppose now that the rankings are up, it wouldn’t be too hard to compute them manually.