Fantasy Basketball Performance Grid

I downloaded all the data from the first 7 weeks of the fantasy basketball season, for the purpose of analysis. The goal was to identify statistical categories that I should target for improvement.

The data I chose to select was each team's final number in each category for each week. I disregarded whether the team was superior to its head-to-head opponent in that category. I just took the raw number, so that I could rank ALL the teams in a given category.

So, one line of it looks like this:

Once all the data was entered for each team and week, I then sorted the teams within each week, and assigned a rank for each category. For instance, in week 1, Andrew's Ballas ranked 12 in FG% with .397, Cameroon Robusto ranked 3 in FG% with .471, and so on.

With this process complete, I computed my average weekly rank in each category. On one particular statistic, my average rank was 9.3, out of 12 teams. I can only expect to win that category a couple of times over the whole season. In fact, it would probably be a waste to try to improve that category, as I'm likely to lose it virtually every week. In another category, my average rank was 1.4! That number is so high, I probably have too many specialists in that particular statistic. I want to win as many categories as possible, not win categories by as much as possible.

The categories I'm most interested in are the middle ones. I have four stats for which my average rank ranges between 4 and 8. If I can improve those by giving up some of my strength in my strongest couple of categories, or by giving up a player who excels in one of my weakest categories, I'm likely to improve my weekly results.

This approach doesn't take into account the fact that I have different players counting toward my total each week, as I bench the ones that are likely to be less productive. It also doesn't take into account the trades and free agent moves I've made. Nevertheless, I haven't been transforming my team drastically on a week-to-week basis, and since these numbers represent the actual performance of my team as a whole, they should be effective in guiding my strategy.

I intend to use these results to drive mutually beneficial trades between my team and others in the league. Teams: Do not fear my science. I seek fair trades that improve us both.

UPDATE: The player behind LoopersHoopers points out that my system doesn't evaluate head-to-head performance, and after all, our league is a head-to-head league. An alternate approach might be to take my team's performance and evaluate it against every other team in the league, for each week, and see what categories I tend to lose, and by how much, and what week-end tallies I tend to get. Identifying categories where I tend to lose by a small amount would be another fine targeting method.

To win the league, you must qualify for the playoffs, by finishing 6th or higher, and win your head-to-head matchups on those final weeks, so this approach might best prepare a team for the playoffs. My team is currently ranked sixth, so I believe general shoring up of statistics is the approach that best serves me currently.