USTA League Nationals will be underway in less than two weeks, and I'll begin posting my previews and summaries of my simulations soon. But something I'm experimenting with is showing the unflighted round-robin random schedules in interesting ways and I wanted to start sharing that right away.
I started experimenting with this as a result of looking at how the unflighted round-robin format was used at Florida Sectionals this weekend, and it is basically graphically showing the schedule in the form of a directed graph with the teams as nodes and the matches shown as edges.
To start, here is the graph for one of the first weekend events, the 18 & Over 3.5 men.
I've used abbreviations for the sections that are hopefully self explanatory.
This shows what appears to be a reasonably random schedule where all teams are somewhat connected and no obvious islands or isolated groups of teams.
There are a few "mini-pods" though. These are:
- Caribbean plays Southern Cal who plays Northern Cal who plays Caribbean
- Midwest plays New England who plays Middle States who plays Midwest
Then there is also a larger pod with Missouri Valley playing Eastern, Florida, and Pacific Northwest, Florida also playing Eastern and Pacific Northwest.
Then there is one more larger pod with Hawaii playing Intermountain and Northern, Texas also playing Northern and Intermountain.
When I do the simulation we'll see what the risk of 5 or more undefeated teams are.
What do you think of seeing schedules this way?
I like the visualizations. But I might be biased as a computer science/math guy and fellow software developer haha. It definitely makes it easy to tell if it was truly randomized like is shown in this map or if it was less than random line the Florida mixed.
ReplyDelete-Stephen F
Thanks for the comment. Always good to hear from another nerd :)
DeleteIt would be interesting to see you do the Florida mixed schedule as a graph where you randomly place the 10 teams nodes in a circle like this to test how effective or ineffective it would be at spotting scheduling issues.
Delete-Stephen
I did it! https://computerratings.blogspot.com/2022/09/what-if-florida-had-used-random.html
DeleteI am a bit confused why it's a directed graph? What is the meaning of the arrow? Seems like the direction doesn't matter?
ReplyDeleteYou are right, for Nationals it doesn't matter, but I did a directed graph to show the visitor/home team when it does matter, e.g. if this was used for a regular season flight.
DeleteGood point. I believe the local leagues are allowed to do a PRR (partial round robin) if they have an EXCESSIVE number of teams. They fail to define the size of EXCESSIVE though...
DeleteThanks for taking the time to do this. I certainly shared your thoughts with the USTA Florida tournament desk and they assured me this will never happen again. (Ie no PRR with 10 teams)
Their logic was this is how nationals does it, but apparently they can't differentiate between the numbers 2 and 4.