The detractors in this thread fail to understand that all metrics have limitations. The ones they recommend instead are simpler, but far more problematic.
One problem with the efficiency metric is that it assumes a similarity in positions. Comparing offensive efficiency scores of point guards to point guards makes sense. Comparing the scores of point guards to centers and forwards might not make as much sense.
I share your conclusion. Kyle should have been on the court, but he should have been shooting less and distributing more.