Since there's so much talk of the sim being "broken" as a result of the success of teams built around, primarily, ATH and DEF, I thought I'd look into the real impact of a guy who's the exact opposite:
Floyd Felix. I recruited him because at the time I was desperate for PER (as you can see, that situation has since corrected itself). But he is one of the absolute worst human-recruited defenders you will ever see, with 21 ATH and 9 DEF.
In order to "hide" his defensive shortcoming, I play him at all 3 wing positions. I did some analysis of his minutes and it looks like he plays between 40 and 45% of his minutes at SG, with the rest reasonably evenly split between PG and SF. I don't want to give a ton of details on his usage patterns since that's essentially laying out a roadmap to exploit my own player, but I do think that a good coach who was willing to put in a lot of work gameplanning for 1 matchup (understanding both Felix's usage patterns and his own team's substitution patterns) could probably play a guy for ~10-12 minutes and have half of them come matched up with Felix with a little bit of luck. With a good enough scorer, and assuming he manages to avoid much time matched up with Friedlander, I could see where jacking up the distro on such a player could yield close to +2 points per game vs. playing with normal distros against average backup defenders. Probably not quite 2, but close.
Compared to other guards on my team, or an average competitive human D3 team, I would think that Felix should generate a slightly above average number of turnovers but a below average number of rebounds. So far neither of those effects have manifested in his stat line, but the sample size is not huge for these events. Regardless, I think the easiest thing to do is say that his impact on possession count is roughly a push with an average player, and that's a good enough estimate to run with.
So obviously, the final factor here is what happens on the offensive end. Based on my numbers, I have Felix worth .268 points per possession better than my team average of 1.092. If you take him out of the team average, it drops to 1.053. That makes him .308 better per possession than the rest of my team. So far, he's averaging 9.312 possessions/game; this means he is scoring 2.86 points per game more than the rest of my team would be in taking over those possessions from him.
Putting it all together, this suggests that even a good coach, at high effort, and with a player who could score a lot against a weak defender without making him even more efficient offensively, and who at the same time is going to create a net positive effect for his team by being placed in a limited (10-12 minute) role - and I would consider this a rare optimal set of circumstances for a Methodist opponent - could probably, at best, be giving me a 1 point per game advantage by playing Felix 15+ MPG. Frankly, I find this result rather shocking. I'd been trying to migrate minutes away from Felix, but I'm probably going to reverse that decision after seeing this. Granted, a player like Felix is still only useful if you have at least 3 more balanced guards and an SF to rotate to keep him mobile throughout the lineup. As a starter he would get lit up by an average D3 guard. But he's not even a top-tier scorer, he's at the bottom of the bottom tier defensively, and in the usage pattern I've established for him, which is not particularly complicated or unusual, he's a net asset for my team.