Fantasy Baseball Power Rankings

Overview

This is a small Python 3 script to compute power rankings in ESPN head-to-head fantasy baseball leagues. There are often unfair results in H2H, so these power rankings are an attempt to make owners who got robbed feel a little better about their teams.

Methodology

It does this by computing, for each team, an aggregate winning percentage (AWP)(tm). AWP is simply a team's record against every other team during a given time period, either a week or an entire season. This way managers can see that the only reason their team did so badly this week was because they were playing the best team in the league, while the team below them vaulted over them into playoff position by virtue of playing the league doormat. You'll get them next week.

Output

The script produces an HTML document containing team standings according to the AWP records, and a Relative Power Matrix(tm), showing how every team would have fared against every other team. The standings also list each team's cumulative opponent AWP to date. Sometimes I manually edit the standings to add a witty comment about how each team performed, much like professional power rankings do (but funnier). The output is further controlled by a stylesheet, which you are free and encouraged to modify.

Usage

Configuration

To generate the power rankings, the script needs several pieces of information from a configuration file. The configuration parameters are:

leagueId - Your ESPN League ID number, usually found in the URL of any of your league's pages
seasonId - Another URL paramemter, for the season; it's just the year
username - The user you will log in to ESPN with
password - The above user's password
leagueName - Name of the league to show in the output
startYear - For cumulative power rankings, the year to start in
startMonth - Start month for cumulative rankings
startDate - Start day of month for cumulative rankings
lowerBetter - Categories where a lower score is better. Separate abbreviations by commas.
rankingsUrl - Where the rankings are hosted. Included in post to league message board.

The script will use the supplied credentials to log in to your league, identified by the league ID and season ID, using cURL (see below). The login cookie will be saved in the specified cookie file.

startYear, startMonth, and startDate should correspond to the first day of competition for the season to get cumulative power rankings.

Running the Script

$ python rankings_cli.py -h
Usage: rankings_cli.py [-c <file> | --config-file=<file>] [-w <week> | --week=<week> | -s | --season] [-m | --post-message] [-h | --help]

     -h, --help                   Print this usage message and quit
     -c <file>, --config=<file>   Configuration file
     -w <week>, --week=<week>     Weekly power rankings for <week>
     -s, --season                 Season power rankings
     -m, --post-message           Post a message

If -c or --config is provided, then <file> is the configuration file to use. By default, the script looks for the configuration options in a file called pr.conf

If -w or --week is provided, then the rankings for <week> are calculated. Otherwise, rankings for the most recent week are calculated.

If -s or --season is provided, then cumulative rankings for the entire season to date are calculated.

If -m or --post-message is specified, then a message will be posted to your league's message board by your user on behalf of PowerBot(tm), a sassy robot who wouldn't mind seeing the extinction of humans. It will pull quotes from the fortune command if available, and it will tell your league mates where to find the rankings, via the rankingsUrl configuration property. Feel free to modify it as needed. Also, it doesn't like being referred to as "it".

Dependencies

This part is so very important, and yet so far down the page. This program has a few critical dependencies, which are specified in the requirements.txt file. Dependencies can be installed using pip: pip install -r requirements.txt

Most critically, it only works with Python 3. However, the Python 2 version can still be found in the python2 branch.

Second most critically, it depends on Requests to perform the HTTP requests.

Third most critically, it depends on the venerable Beautiful Soup library. Additionally, it uses lxml as Beautiful Soup's underlying HTML parser. The Beautiful Soup documentation contains information on installing lxml. The Python 3 code only works with Beautiful Soup 4, while the Python 2 branch has only been tested to work with Beautiful Soup 3.

Last most critically, the command line script uses Jinja2 to produce its HTML output.

michelmansour / fb-power-rankings

readme