My model has done well on MLB totals for 2 straight years now. This year going 168-149, 53% (documented on here as 2 unit bets) And last year going 237-196, 54.73% (didn't make a season thread last year so unproven to all except me). Up about 80 units combined in two seasons (if betting to win 2 units on each play). So with 750 plays across 2 seasons, all those sifted out of 2,000 potential model total plays. I'd have to assume that this is a large enough sample size to practically rule out being fluky. But I'm no expert and could be wrong...