At least, I hope it’s the last change. Point guards were very underrepresented by the winshare formula as previously presented. I figure a lot of what point guards do — pass — goes unrecorded. More than half of the passes made by anyone that lead to field goal attempts are not credited, because assists are only given for made shots. As such, I attempted to correct for this, by dividing actual assists by the field goal percent of the rest of the team (tfga-fga)/(tfgm-tfm), giving an estimate of total attempted assists. One possible problem with this measure is that better passers will make passes that lead to higher percentage shots, and this formula doesn’t entirely capture that. Given that the shooting percentage of the rest of the team is used, there is at least some compensation for the individual player and team situation. When assists are replaced in the valuable contributions formula (discussed here and here), point guards get a much fairer shake, even though the adjustment applies equally to every assist made by every player. This gives us the following (hopefully final) list of most valuable players for the 2007 season:
This list puts the two most-discussed MVP candidates at the very top, and then follows them with all the other players one could have reasonably considered for the award. Notice that Dallas won so much that a player no one would consider a star wound up in the top five, outranking McGrady and Bryant, and any number of other good players. In terms of wins, however, Terry just contributed more. His high ranking is somewhat counterintuitive, but he did contribute about 16% of the valuable stats to a 67 win team, which ought to count for something. His high ranking, I would argue, is completely legitimate; this is the value of objectivity.*
* I realize that it’s hardly objective to subjectively decide that point guards are underrepresented by a formula, and then change the formula to compensate, but given that, the formula does objectively compare across all players.