When analyzing projections from fantasy sites, you'll often find that the projections are higher than betting lines, which will make you bet shit tons of overs. This is a common mistake. The projections are mean based, which tend to skew higher because the range of outcomes are not evenly distributed. (i.e., a running back on a good day could go for over 100 yards, but he would never go under 0... hopefully). If you find a mean projection lower than the Vegas line, the under is likely a good bet. To bet an over, the mean projection should be significantly higher. All bets below will follow that logic.
San Francisco @ Philadelphia
Deebo Samuel u21.5 rushing yards -114 (FanDuel): Samuel is projected around 16.8 yards. He hasn't seen more then 4 carries in a game since week 3. Even if he gets 4 attempts (realistic projections are at 3.0), he'd need to average 5.375 yards per attempt to get there. He's also gone under this number in 9/15 games this season.
DeVonta Smith u66.5 receiving yards -109 (Caeasars): Smith is projected slightly below this. He'll have his work cut out for him against a 49ers defense that ranks 5th in pass defense DVOA.
AJ Brown u4.5 receptions +135 (DraftKings): I've seen projections (PFF) which have Brown at exactly 4.5. I think those projections are a bit too low, but I can't turn down +135 here.
Cincinnati @ Kansas City
Joe Burrow u17.5 rushing yards +100 (DraftKings): Burrow has gone under this number in 11/18 games this season and is projected around 15 yards. I'll happily take this at even money.
Joe Mixon u60.5 rushing yards -115 (PointBet): Mixon is projected in the low 50s. Even though Kansas City's defense has a reputation for being awful against the run, they've been far better this season, ranking 15th in run defense DVOA.
Kadarius Toney u35.5 receiving yards -114 (FanDuel): This number is too high, Toney is projected in the high 20s.