Not arguing but need to point something out.My apologies if Johnson's 20.1% TO rate had already been posted. Nevertheless, the number shows that Johnson was one of the most TO prone on the team last season, and my reply was within the context of your post.
Shifting gears to your new target, Johnson's numbers don't indicate that he was getting 4 assists every 10 possessions. IU had 16 assists per 70 possessions last season. Rob's assist percentage was 20.5%, or 3.3 of those 16 team assists. His usage rate was 16.9% per 70 team possessions, or 11.9 individual possessions. 3.3/11.9 = 27.6% of his possessions ending with an assist. It's not a trivial task to then figure out how to split the remaining (10-(2.01+2.76)) 5.2 possessions into shots and passes.
For a meaningful measurement of individual player impact on offense, I think the appropriate metric is offensive rating (ORtg%). Here are actual offensive ratings from last season for IU regulars, ranked best to worst.
Zeisloft - 129.7
Bryant - 129.0
Ferrell - 125.6
Morgan - 120.5
Blackmon - 120.2
Bielfeldt - 117.8
Hartman - 115.9
Anunoby - 115.8
Johnson - 115.0
Williams - 109.7
Now you might observe Williams' low ORtg% and refute that this metric is a valid measure because anecdotally Williams was important to the offense last year. I would temper that notion with the fact that Williams' usage rating last season was a whopping 26.2% -- highest on the team, excluding JBJr. The sheer volume of Williams' touches gave him more chances to impact the offense.
fwiw - stats show what they were intended to show. That stat was created to show efficiency, not productivity, nor impact. It does well in that manner. But, to show how much impact a player had - it needs multiplied by the usage factor (much like you're doing) or simply use a PER. (though it has defensive stats also) ... Some players don't have the gifts to use extensive possessions and a spot up shooter like NZ is a great example.