|
Now that I'm to something of a checkpoint—it isn't a perfect tool, but it does generate schedules that I wouldn't be terribly unhappy to be forced to play (my Elsdorf Argonauts are playing with one next season, in fact)—I'm interested in your input on what my next move should be with this tool. There are two paths I can take.
Option one: I have in mind some changes that would make the schedule generator better at handling arbitrary, asymmetric league structures, like tracking which teams were excluded from divisional play the last time the scheduler attempted to schedule a divisional game, or doing scheduling in a team-first rather than series-first manner, to do better at guaranteeing even matchups between teams where possible. The scheduler would still have a large random element, but would be better at generating sensible schedules for a wide variety of league structures. I think that some additional schedule realism would also shake out of this effort—games every day, with different break days for different teams.
Option two: provide deeper user configuration for schedule structure. You'd have the option to specify a list of series for each type of play, which the scheduler uses instead of automatically generating them, or a list of series per opponent per type, which gives you the ability to design a schedule almost to the game. Note that it would be hard to make that setup work with asymmetric divisional structure inside of subleagues, and may pose obstacles to asymmetry in any form.
I feel like I'm eventually going to get to both of these—in particular, specifying series per opponent may require doing part of the first option anyway. That said, I'm at a point where I'm more or less personally satisfied with the outcome of this project, and as such, I'm willing to let the time I put into it be more community-directed. Let me know what you'd like to see.
|