Basketball Prospectus: Unfiltered Everything Else is Fluff.

November 19, 2010

Weekly fantasy projections

Filed under: Uncategorized — Bradford Doolittle @ 10:26 pm

(Note: We are not offering the fantasy downloads sheet this season, however this page is still being used as a reference point for those interested in how our weekly schedule- and position-adjusted projections are formulated.) – BD

As we expand our offerings at Basketball Prospectus, we’ve targeted more and better content for fantasy basketball enthusiasts. Part of that will be more fantasy-themed articles, which we’ll be launching soon. For now, we unveil the first Weekly Fantasy Projection spreadsheet, which ready to download right now.

The projections built on three pillars:

1. Preseason SCHOENE forecasts

2. Season-to-date advanced statistics

3. Upcoming opponents

Each player’s statistical baseline is established by blending his SCHOENE projection with his season numbers. The further we get into the season, the more important the actual results get and SCHOENE begins to be filtered out. Old Russ, though, he doesn’t go completely away until the end of the season.

Once we have the statistical baseline, it’s then time to match them with upcoming opponents. Each spreadsheet covers a seven-day window. For each game, game pace is projected to determine the likely number of possessions in the contest. Then the baseline projections are adjusted by how well each opponent has defended each player’s positional group. Wing players are grouped together as are interior players. Point guards are in their own group.

Right now playing time is based on season-to-date averages. Going forward, we’ll be setting up a playing time grid that we’ll use to reflect breaking events, such as injuries and suspensions. We’ll adjust projected rotation minutes up and down based on who we anticipate being available.

An important calculation measures how well each opponent defends the position that is player plays. For example, say the Mavericks have a 0.94 score against Derrick Rose, which means that they’ve defended point guards six percent better than the league average. The score is based on the GR ratings you’ve been seeing in PBP Roundup and is provided just to give you an idea of the difficulty each opponent presents each player. They are not however actually used in the adjustments made to a player’s baseline projection. Each category receives unique adjustments.

Top big men from the last week are ranked by individual winning percentage, the rate stat component of WARP. Seasonal rankings are based on WARP. Our exclusive schedule-adjusted projections use a combination of SCHOENE forecasts and season-to-date data. We’ll be predicting the weekly numbers for NBA big men based on a Monday-to-Sunday schedule suitable for weekly head-to-head fantasy leagues. By this method, SCHOENE is combined with actual results to product a new baseline forecast for each player. These are then adjusted for a player’s upcoming opponents and how well those competitors have defended his position. The projections compiled in a module of NBAPET, my system for projecting, evaluating and tracking the league.

Any questions? Suggestions? Drop me a line here. You can also find me on Twitter at @bbdoolittle.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress