rankings.py [-sport] [-d '|'] scores.txt --output rankings.txt
The "-sport" option is either -football or -basketball to properly
normalize scores for these sports. Most other sports don't require this
option.
A "-d" option specifies the file format. The default is CSV with column
titles. A value of | (must be quoted to protect from the shell)
is a common alternative to specify pipe-delimited data with no column
titles. In this case, the column order must be date, team1, score1, team2, score2.
For scores extracted from web sites, a separate parser should be used to create the required CSV file
A "--output" option specifies the output file. This option generates an output file containing the rankings.
scores.txt is a CSV file. The -d option allows different formats.
Each row in the file has results of a game:
'date', 'team1', 'score1', 'team2', 'score2'
This is used to compute a "power" ranking that compares their total share of points and games with a projected share. The projection model is iterated until it matches the actual outcomes.
This application has been tested and runs on Python 3.3. For any other version of Python, your mileage may vary.
I have existing data files for pro football, college football, and Virginia high school football on Github now. Some are in space-delimited format and others have been migrated to pipe-delimited format.