NCAA Tournament Home Court Advantage

It's a common assumption that neutral court games should be treated differently from games played at one team's home court, but is that really true?  The SI article that looked at home court advantage concluded that it was primarily due to the referees treating the home team differently.  That jibes with something I found -- that large home dogs don't get a HCA.

Presumably the refs don't give the benefit of the doubt when they know the home team is overmatched.

I did some other experiments (prior to starting the blog, so they aren't documented here) where I trained a predictor on regular season games using just a strength measure for each team, so that the prediction equation looked like this:

    MOV =  (C1 * Strength of Home Team) + (C2 * Strength of Away Team) + C3

C2 was negative, and C3 (along with any C1/C2 ratio) was the "home court advantage".

I then tested the accuracy of this predictor on NCAA tournament games, first treating the higher seed as the home team, then the lower seed as the home team, and then washing out HCA altogether by dropping C3 and forcing C1 & C2 to be equal. 

What I found was that the best prediction was made treating the higher seed as the home team.  This makes some intuitive sense -- the refs are giving the benefit of the doubt to the team that they "know" is the better team.  So I'm a little dubious that there's really no "HCA" in tournament games, although I don't know that anyone else has looked at it.