Week 12 Machine Learning Point Spread Picks

This week we stick to the strategy that made us winners last year. Find games with crappy teams facing tough teams with a single digit spread. The models certainly found some and we're going to make money off of it. 

I can't hide that it has sucked this year so far. But I know it will get better and hoping it does this week and the coming ones to build confidence to make big bets during the playoffs. Mistakes last week included putting money on single digit weak teams like the Eagles and the Patriots. Miami was a tough one, not sure if I would've changed it except that Tua is a small sample QB and now even smaller. Lessons learned, moving on to week 12.

I'm going to be very short and concise this week and just stick to the models. Mostly because I've had a lot of work this week and can't wait to sit back and eat turkey with the family. Next week I'll be back with more detailed analysis. Also, I might bring back high-stakes poker player for a short podcast tomorrow to debate these 3 picks and the 3 honorable mentions. 

Las Vegas -3

The Raiders look good and should've beaten the Chiefs again. Not sure how Derek Carr got so good from one year to the next but he's being real smart. Vegas has something to play for but not the Falcons. The models range anywhere from 5 to 15 point estimate for Vegas. Go Raiders!

All picks available here: https://rokfin.com/article/2307/Week-12-NFL-Point-Spread-Picks



Week 11 NFL Machine Learning Point Spread Picks

Home team favorites are covering 45% ATS this year. Models detect two opportunities in this trend. Two other picks find undervalued teams. First week where all the data used to predict is in 2020. Models are 58% ATS overall since week 5. We bet all top picks and honorable mentions this week. Let's dig in.

Last week wasn't great. Only one of the top 3 picks hit and 2 of the 3 honorable picks. Seattle's crumbling and the weather in Baltimore didn't help and we paid for it. We did bank on fading the 49ers and Indy on TNF. My Chicago-born wife told me to never bet on the Bears and she was right. On MNF their offense looked as bad as their DVOA or worse. This week, we make all seemingly good picks official picks and we got 4 solid ones.

Miami -3.5

The typical 3.5 trap to bet the other side. Don't fall for it. The models have the underrated Dolphins winning by almost a TD and other models have it by 2 TDs. This spread will move higher as Miami is getting lots of money from the sharps.

  • Captain + Injuries: Tua > Driskel
  • Trends: Away favorites by 3.5 are 57% ATS since 2010
  • DVOA: Denver has the worst rank offense in the NFL. They won't score much against a slightly above average and hot Miami defense.
  • SOS: One point for Denver. The Broncos have had a tougher schedule than Miami according to teamrankings.com, but it hasn't been that easy. They've faced the Cardinals, Rams, Seahawks, and Bills. While the Broncos have lost against the Steelers, Titans, Bucaneers, and Chiefs.
All picks are available on Rokfin.com/nflpickles.



Week 10 NFL Machine Learning Point Spread Picks

We stick to two basic philosophies this week: 1) weak teams get beat out badly by strong teams with single digit spreads and 2) spot recency bias with strong teams that didn't perform last week. Let's dig in.

Let's be honest, the top picks weren't good last week. We hit 1 of 4 with Green Bay on TNF being the only win. The Monday morning QB in me should've known better. McCaffrey was coming back to Carolina and Thomas to New Orleans changing those teams offenses significantly. Both should've been stay away games. The more you know... The next games based on confidence covered (BUF, LV, BAL) which made me this week at least come back to these honorable mentions. 

As we know and was reinforced from last week, not everything is analytical. The analytics help detect possible opportunities, but digging deeper into the factors below and know how many of them go to the side that the anlytics point is important. Here are other factors that I like seeing agreeing with the models:

  • Injuries - This is the #1 issue. Is the team the model is picking better or worse off injury-wise when compared to previous weeks?
  • DVOA - Does the matchup make sense? A good passing offense vs a weak passing defense? etc.
  • Captain Class - We talked about this a lot last year. It's based on a book. But essentially I'm trying to see if the QB for the team we're on is a significantly better player and leader
  • Trends - Are there any significantly biased ATS trends for the situation or the team? I'm not talking 3-0 ATS in last 3 games, a bigger sample. 
  • Strength of Schedule - Analytics maybe biased if teams have had a significantly different schedule. We look at teamrankings.com for this info.
This will also help me structure the analysis around the models. I don't like long paragraphs and articles. I'm a straight to the point kind of person. Maybe one day I'll have more visual tables with check marks, for now I'll stick to bullet points. Here we go!

Baltimore -7
Quick review: Strong team vs a weak ass Patriots team who should've lost against the Jets last week. 
  • Injuries - Mark Ingram will play. The list for the Ravens is much smaller for the Patriots. Advantage: Ravens
  • DVOA - Baltimore is #1 against the run which is all the Patriots can do. 
  • Captain Class - I'll take the 2019 MVP any day over Cam.
  • Trends - The Ravens are 14-5 (73%) ATS since 2018 on the road
  • SOS - Baltimore has had a tougher schedule.
Get all my picks & analysis and other channels' thorough analysis on Rokfin for only $9.99/mo.



 

Week 9 Machine Learning Point Spread Picks

66.7% ATS on locks last week. This week we keep pounding weak and banged up teams and find opportunities based on trends and recency bias. 

Last week we should've been perfect if Bruce Arians listened to Tom Brady and had gone on 4th & 1 at the 20. I still think it was the right spot and fortunately Daniel Jones missed so many long ball if not Tampa loses that game. This week a dug deep because there are so many injuries and new QBs that forced me out of many games. At the same time, if the models are pointing to the right direction and the team is banged up, we're going to take advantage. Which is why the first team I'm gonna talk about is Green Bay. Raiders and Chiefs were money. 

Green Bay -7 TNF 
The Packers and Aaron Rodgers coming off a loss are 31-15 ATS. The list of injured players in SF is too long for me to write. The 49ers are a totally different and weaker team than when they started this year and especially who they were last week. Green Bay couldn't stop the run last week and their defense has been pretty bad against the pass. This week Kittle is out and Bourne tested positive. Mostert is out so they won't be able to run the rock. Packers should be ok on defense. Their #2 offense will score +35 points against a pass D ranked #17 and sinking. Also, road teams on TNF this year have covered 5-2 ATS and road favorites since 2000 have gone 42-31 (57%). Not a high confidence game from a model perspective, but this week the market realizes how bad the 49ers have become.

Get all my picks at Rokfin.com/NFLpickles