9-11; -$2073
Scored the second max bet of the season, but losing a near-max on the Angels cancelled out most of it.
Regarding the methodology: The starting point was a formula to figure out last year's expected runs based on a couple of offensive indicators. A few regression analyses were done to separate out luck from expectation (that is, to figure out what was reasonably repeatable). Then adjustments were made based on personnel changes and sabermetrics expectations. The season-long picks were then made based on Pythagorean expectations based on the run projections.
On a game-by-game basis, the expected runs allowed are re-calibrated based on starting pitchers, and season expected records are re-calculated (i.e. "The Lions should win 74 games, but if Jones pitched every game, they'd win 86."). Then the teams win expectations are computed head to head via established formulas (e.g. If a .600 team faces a .400 team, it can be expected to win a little over 69% of the time) and a home field (historical) adjustment is made.
Finally, the expected win percentage is compared against the breakeven win percentage of the money line odds, and bet sizing is according to the amount of the perceived advantage (for instance, the "max plays" occur when my computed percentage exceeds the offered percentage by more than 15%. So if a team is +150, the breakeven is 40%. If I calculate that the team's chance to win the game is more than 55%, it qualifies for a max bet. So the max bets aren't the most likely winners, necessarily; they're the biggest perceived overlays).