Subj: Scheduling/Travel Questions
Date: 5/17/01 7:53:36 AM Eastern Daylight Time
I got your email address from Amy Pickett.
Let me start by introducing myself. My name is Jason Jennette, and I am a hockey fan in Macon, GA. I'm also a very active member of the CHL mailing list on the internet.
I am a graduate student in Computer Science here in Macon and have begun working on my masters thesis. Without getting too detailed, my thesis revolves around scheduling and optimization for sports league. Because I am most familiar with the CHL, I have used the CHL for the majority of my test data.
You can see some of the preliminary work at:
I do have a few questions that tend to be of a more technical nature that I hope you can help me with. With the BC and Topeka troubles this season, I have held off of emailing earlier, as I figure you were rather busy. I know the next season is coming soon in the scheduling process, so I hope to get my foot in the door before the process gets too far along.
Without digging in too deep, my thesis revolves around the scheduling of the league schedule with an eye towards minimizing travel distances for the entire season. With the idea that more travel increases costs, hopefully an "optimized" schedule would save money.
1) Is there a maximum distance a team would be scheduled to travel overnight for a game? I currently have records from the 1996-97 season where Macon had to travel almost 1000 miles overnight for a game, but most overnight travel for that season was closer to the 400 mile range.
2) I know there is a restriction against any team playing 4 games in 4 nights. I also have noted that no team has ever played 3 home games in 3 nights. I dont know if the second is a league restriction or not. I suspect it would be a bad idea to try 3 consecutive home games, but I wasnt sure if there were any restrictions on it. I was wondering if there were any other restrictions along these lines? I know there is no restrictions on how many times 2 teams can play each other in a row, as Macon and Columbus played 4 or 5 games in a row over a span of 2 weeks 2 years ago.
3) When it is time to start creating the schedule, I know that most leagues ask the teams to submit "available dates" when their arena is open. I was wondering if this list may include the teams preferences. For example, Macon always wants a new years eve game, even if it isnt a weekend date. Could they request that specific home date at the expense of possibly a weekend date? If preferences aren't taken into account, is this something that could be useful?
Currently, my algorithm "drafts" home dates.. going through each team's list of preferred home dates and assigning them 1 game for each team, until the schedule is done. This section of the algorithm may change in the near future, but while I have the questions I figured I might as well ask.
4) Do teams get to ask for specific opponents for games? Macon has not been in the league long enough to have any traditional opponents yet, but I have heard that some of the western teams see the same team on special dates each year. I dont know if this is true or just a fluke of the schedule.
5) Is there any easily identifiable way to say how you decide when a team will take its swing to the opposite division? In Macon's first season, all of the eastern teams went west in the first half of the season, and the western division teams came east in the last half of the season. With the increase in inter-divisional games, and the unbalanced schedule between divisions, there isn't really much of a "western swing" anymore, but I know it is preferable to pick up 2 or 3 games in the other division while you are on that side of the league.
6) Lastly.. and this questions just came up in the past 4 or 5 days. I have assumed that most road teams would prefer 2 game stands in a city to cut down on travel costs. A friend from Wichita told me that their attendance actually suffers from Fri/Sat series where two weekends with 1 home game each are better overall than 1 weekend with 2 games and 1 weekend off. I will continue to evaluate past attendance figures on this topic and see if I can find anything. If you already have information on this, that could save me a lot of work.
(You can actually see some of my preliminary work on this topic at the link above.)
For right now those are the major questions I am facing. very quickly, I will give the sequence that my current algorithm uses for creating a schedule. If you see anything that needs changing or would like to give me a quick review of how the CHl schedule gets created, that would be wonderful:
* Get the list of preferred home dates from each team, in their order of preference.
* "Draft" the home games. Using the lists from the previous step, nail down all of the home dates for the entire season.
* Begin assigning the road teams for each of the games, trying to keep their travel distances to a minimum.
Any assignments that will not work need to get kicked out as an error. How to deal with these errors is still up for debate. My advisor wants me to look into picking the home and road teams at the same time, instead of 1 pass for the home dates and then 1 pass for the road teams. This would cut down on impossible matchups (due to distance) but would drastically increase the work needed to handle other issues, in particular it may back a team into a corner for the last game of the season, forcing them to criss cross the country 4 or 5 times in 2 weeks.
I know this is a lot to digest. I have tried to make everything as clear as possible so that I won't need to have any follow up questions. If there are any unclear parts of my questions, please feel free to let me know.
my private email is JasonJ1052@aol.com
my work email is Jennette_JL@mercer.edu
home phone is 478-741-6408
work phone is 478-301-288
After my program is up and running, and if you think it could be useful, perhaps I could attempt to apply my program to the CHL schedule for next year. With the possible WPHL merger, and all of the uncertainties for the future (interleague play, mergers, new teams, etc) I think it may be a bit much to try and wedge in, but hopefully my program will be flexible enough to handle anything you can throw at it.
That is the hope.. we'll see how it goes.
I thank you for your time,