Post by George Gervin on Sept 6, 2019 15:54:42 GMT
We all know the offseason can feel like it drags on, and one thing I wanted to try to quantify in that downtime is a question we vote on each year as part of the GM Survey: what team has the brightest future? What team, using data and projected Wins Above Replacement (WAR) over the next five seasons, is set up to ascend to the top of D5 in the future seasons to come? Before the results are presented in a five part series, a few ground rules to establish on this exercise:
- The source for the WAR projections is FiveThirtyEight’s CARMELO engine, which uses similarity scores and tracking to approximate WAR for players across future seasons. Some player’s scores have broad deviations, as the engine can only project with a certain degree of confidence on how a player will perform. In short: don’t complain if you see a player from your team ranked criminally lower than expected, or if you hate WAR as a measure amongst the many analytics acronyms (VORP, BPM, etc.), because this is a MATH based exercise, not an eyeball test.
- To limit taking into account players that could aggressively decline – such as those in their late 20s or early 30s – based on analytical projections, the cut off I set for eligible players in this analysis was age 25 and younger. So long as the player was not 26 or older at the time of this analysis, they were fair game for the projections.
- Rookies that have yet to play were taken into consideration as part of the analysis, as some teams would have had no eligible players without taking that into account and, in the interest of a result for each team, I incorporated rookies.
- Rather than aggregate and sum all players under 25 by each team, I limited it to the top three players by WAR on each team to ensure some measure of control on the analysis, plus three core players seem to be the standard for team building, both in D5 and the real life NBA.
Stay tuned over the next few days, as I plan to hopefully release one part per day!
P.S. As a big music fan, these parts will be titled after some well-known – though definitely older – hits.