APBRmetrics

I have often wanted to compared ws/48 estimates to RAPM estimates. I calculated what I think is the rough equivalency:

ws/48 RAPM
.200 +5
.180 +4
.160 +3
.140 +2
.120 +1
.100 0
.080 -1
.060 -2
.040 -3
.020 -4
.000 -5

As has been discussed in other threads offensive winshares in heavily influenced by usage and defensive winshares is very rough and not very precise. So this rough equivalency is indeed rough and best when players are not unusually high or low on usage or significantly better or worse on defense than their teammates.