Those estimates of 5 points and the resulting winrate were quite conservative actually, based on late game fouling often giving the winning home team an extra few points, and on teams averaging 90 points per game when the true average seems to be 85.
If it's possible for BB-Justin to do a game engine experiment, I'd like to suggest this; 24 identical players across two teams, one team has HCA. We could figure out just how potent HCA is and also maybe get a control sample on average scores in the current game engine (assuming the players have roughly equal offence and defence skills).