Chart set design is a popular topic in the Paydirt message boards. My data collection, statistical analysis, and quality assurance yield accurate and playable charts that are as good or better than the original Paydirt sets.
Step 1. Data Collection
The only way to produce accurate charts is to use play-by-play data from the regular season. I use the official NFL play-by-play data that STAT’s Inc. produces, and is available at NFL.com. The Gamebook contains all the play-by-play information to create team charts. The parsing software I created reads the play-by-play file, presents the information as it will be categorized for analysis, and then persists this data to the database so it can be analyzed. I fully review each play for accuracy prior to persisting it to the database. Additionally, I watch every game using the NFL’s Game Rewind on-demand service.
Step 2. Analysis
Once all play-by-play data for the regular season is collected, then statistical analysis is performed to obtain all the needed data elements used during the design process. This analysis is performed using database and analytic software to ensure statistical computations are accurate. Also, during the analysis phase all data elements are reconciled back to NFL.com to ensure data integrity and accuracy.
Step 3. Design
After the statistical analysis is completed and all data elements are available for use, the manual process begins of mapping play outcomes to the team chart and assigning the appropriate probabilities. Most of this mapping process occurs manually, but probability and option set calculators are used to facilitate this process and ensure accuracy.
Step 4. QA Testing
I usually simulate a minimum of 100 seasons to ensure statistical accuracy, and also enlist some play testers to test head to head play. This ensures you have a statistically consistent, and a fun, playable set for head to head competition.
Step 5. Revisions
If a revision is needed or a typographic error is identified, a correction is made and new files will be distributed.
