I have tracked various scheduling roles (including B2B, 1DayRest, 2+Days Rest, 3rdin4, 4thin5 etc) and find no consistent correlation between any of them for either Over or Under results.
Projecting Totals has far more to do with correctly identifying the nature of the teams in play. And recent play is a good indicator as well. A team's past ten games provide a good lean to Over/Under. When combined with their opponent you can get even more of a Lean.
Note that Books adjust Totals lines accordingly once a team persists in one direction for six to eight games.