Week 16 2019 picks
I consider Week 15 in the NFL “moving week” for the contests. If you’re on the cusp of the money and you keep pace, you are well positioned for a shot with two weeks remaining. On the flip side, if you throw a meatball, your chances at winning are all but torpedoed…unless you hit 9/10 down the stretch.
Well as it turns out, my Week 15 official picks were a dud. Very disappointing. It was one of those weeks where you could argue either side of the games and a lot of chalk came through, which is usually to my detriment—not to mention a ridiculous bad beat in the WSH-PHI game on the last play of the game. Weeks like last week are a real eye-opener for me and challenge me to assess the true efficacy of the algorithm. For one, while I consider the consensus line when analyzing each game, I hadn’t realized how well (though modest) my algorithm has been performing when compared against the other models. After disavowing the algo’s suggested plays last week and getting burned, I have realized the next enhancement to the model must be to weight the individual factors systematically, rather than arbitrarily favor one or the other each week. I had subconsciously overweighted the consensus estimate relative to my own algorithm without adjusting for in-season performance. For how numerically driven the system is, it goes to show how difficult it is to avoid bias in decision-making.
Week 15 in review:
Raw algorithm picks: 4-1 (80%) <— should’ve been 5-0 had it not been for this play!
Unofficial plays: 1-4 (20%)
Official plays: 1-4 (20%)
Kahneman is right; the loss still stings a week later. But let’s move on. Maybe 5-0 is in our sights this week!
Raw algorithm picks:
CAR +7 ★★
NYG +2.5 ★★
CHI +6.5 ★
LAR +6.5 ★
GB +5.5 ★
I’ll go into the star ratings in a future blog post but in the meantime the stars simply project the confidence level in the selection. A single star shows value against the number. Two stars show value and identify teams that have agreement between all of the models. In 2019, the two-star games have covered at a 55% clip through Week 15. One-star games have covered at 46%. Now that I’ve introduced that, let’s see how our Sharp friends are weighing in on these games.
Sharp/Contrarian sides:
TB (HOU) - Moderate - The models agree that TB is the right side here but the market’s moved from the opening number of TB -1 to now +3. The public is pounding HOU at a 80% clip. My gut wants HOU: with TB’s injury concerns at the WR position, Jameis’ penchant for throwing INTs, and HOU in a dog fight for the division, I can see why the public is on HOU. Nonetheless, it smells of an overreaction. TB is on the card as a play. (This is a Saturday game)
NE (BUF) - Weak - Wow, for once the public is laying off NE, favoring BUF with a 65/35 split. Yet, we can’t forget that capital flows where it’s treated the best and NE has treated it well. NE has covered in 69% of it’s games since the late-2017 season. The algo likes NE by over 8 points but the market’s followed the BUF money (opened +7, now +6.5). This is not a strong play but lay the 6.5 with NE. (This is a Saturday game)
LAR (@SF) - Moderate - The line’s remained steady since the early hang: LAR +6.5. SF still needs to win games to wrap up the division and since LAR has been eliminated from the playoffs, 3 out of 4 public bettors think the Rams will mail it in. The algo only likes SF by 2 so LAR getting 6.5 shows a ton of value. I like the Rams. (This is a Saturday game)
CIN (@MIA) - Strong - Ugly game but tasty sharp play. CIN opened +3, but sharp money seems to be moving the line towards CIN even though they’re only getting 37% of the public backing. My algo has CIN as a 1-point favorite, which matches the current market line. With the SuperContest offering MIA +1, this seems like an arbitration opportunity of value. I’m taking CIN +1 ATS this week.
CLE (BAL) - Weak - There could be a small case to be made to play CLE here getting 10 against a division opponent. The line seems fair at BAL -9.5 so you get some slight value and the public is with BAL at a 5:1 clip. However, BAL has blown out teams this season and CLE seems to be in disarray. These two teams are on the polar ends of the emotion-process scale. CLE is a team that doesn’t seem to have strong organizational leadership and certainly doesn’t have control of the locker room. BAL, on the other hand, has a top-3 head coach in the NFL and outstanding chemistry. Against a well-coached team with the better QB and better defense, 10 points still might not be enough. Hard pass.
IND (CAR) - Weak - Conflicting signals here. My algo favors IND by only a FG but they’re getting a full TD in the SuperContest. Topping it off, 43% of bettors as of this writing are on IND and the line’s moved a critical 0.5-point towards IND and the key number of 7. The consensus models like IND here as well. I’m going to pass.
CHI (KC) - Moderate - An astonishing 8:1 tickets on KC moving the line from KC -4 to KC -6, and who can fault the public here? You get the better QB-HC combo, and a red-hot team vying for playoff home-field advantage. The consensus model line agrees with the market, while my algo thinks the opening number was close to right. The SuperContest gives CHI the hook at +6.5, so if you’re really feeling contrarian, you can take the Bears at home. I’m tempted and I’ll keep it on my card but let’s see what else is out there.
NYG (@WSH) - Moderate - The algo has the G-Men as a strong top-5 play and the market is moving this line towards a pick ‘em (currently WSH -1). The Giants are getting a respectable 55-60% of the public tickets, which is not enough to scare the discerning bettors away (usually I look for disproportionate public backing > 80% before taking the other side). With the SuperContest giving 2.5 to NYG, I like this arbitrage opportunity here.
MIN (GB) - Moderate - The money seems to be coming in on MIN even as the Packers tickets pile up. Line’s moved slightly from MIN -4.5 to -5.5; MIN getting 35% of the tickets. My algo likes GB here by about a point but it’s throwing the Pythag yellow flag (meaning GB may be overrated), so whether or not I play this game will come down to which factor I consider more heavily: the market consensus or my independent handicap? Pass.
Unofficial plays: