DBF Week 4 Update
DBF Week 4 Update
Recap of week 3, overpalooza over-correction, how to win when you aren't "winning", and a deep dive in to Betting the Under in large spreads
“If you aunt had balls she’d be your uncle…” I heard this expression a LOT in college - usually while someone bemoaned a backdoor cover or last second field goal in a blowout to carry the OVER. It happened on Saturday this weekend when Ohio State beat Notre Dame by 3 (I had them +3.5) to ruin a perfectly good 5 leg parlay. Did I hedge? Yes… did I feel good about not landing the big one? Also, yes… As it would turn out, this quote was a favorite of the guy who ran the sportsbook in our house… It figures!
So what does this have to do with the DBF process? It made an appearance on Sunday when I nearly blew a hammy victory lapping the Road Dog Saints ML call I made last week. Sure enough, Derek Carr gets injured, Jameis Winston comes in, and the Saints blow a 17-0 lead. If Derek Carr had JUST finished the game… “If your aunt had balls, the saints might have won the game”.
Week 3 Recap
The model selected 14 lines as favorable and hit 6 with 8 losses (42.86% Win Rate). It was always going to be a learning process for the model and there are now corrections in place for recurring themes from Weeks 1-3.
One, the model doesn’t want to believe that bad teams are… just bad. The model couldn’t get the bears to +12 against the chiefs so they were a “slight” favorite to cover. The Chiefs are good, the Bears are defeated.
Two - the model liked the Jaguars, Vikings, and Bucs given their early season offensive/defensive performances (note, not how they actually faired in the Wins and Losses Column). I’ve since updated the model to increase relevancy of recent offensive and defensive performances against 2022 (heaviest weight) and 2021 (lesser weight) baselines. The lines still liked the Vikings and Jags, just not as much.
The other 4 “whiffs” were the Ravens -8 (still working on an injury/weather factor), the Titans +3.5 (Browns Defense is elite), the Commanders +5.5 (Bills offense/Defense returned to 2022 glory), and the LA Rams +2.5 (3 point road loss). As you see in this week’s projections, Vegas is making similar adjustments in their ATS lines which is why there are fewer early-week misalignments. Both Vegas and my independent line are relatively in lock step (with some fun exceptions).
The model predicted 8 of 12 OVER/UNDERS correctly, with the 4 misses being OVER Picks. The UNDER took 11 of 16 games this week - impressive when you consider Vegas called for 706.5 points and teams scored a collective 706. 11 of the 16 games failed to make it to 45 points this week (avg. points per game if 706 points are scored = 44.125 points). Given that 706 is “above average” in points for a week in 2022 and 2023, I, personally, would have expected a closer split between the OVER and UNDER.
This is another sign pointing to the downward trend in scoring from 2021 —> 2022 —> 2023… I’ll continue to monitoring scoring trends and Vegas expected points to determine if the points totals need the downward adjustment in 2023.
Overall, the model returned $113 for every $100 wagered (assuming +100 odds) - you’d have been up an extra $3.50-$5.50 if you followed the ML process from last week with a Steelers Win (+135 - +155) and a Saints Loss (+100 on Friday). We were consistently minimizing losses and maximizing gains as of Monday Morning. For ATS it was $46 won and $39 lost (still +$7 despite being at .500). For totals, we won $66 and lost $20. That had us up $53 staring down what felt like pretty even MNF matches. Then Tampa didn’t cover (+5.5) nor did they score enough points (UNDER 45), and the punchless Rams lost by 3 (+2.5) to the Bengals in a boring clunker (UNDER 44)… That’s how you go from $53 to $13 in a hurry.
Week 3 Observations
Five things I think I think, including the focus area deep dive
The Dolphins are on a historic pace offensively - the 2013 Bronco’s who currently hold the NFL scoring record, scored 606 points (643 points in a 17 game season). The Dolphins are on pace to score 737 points… They’re more than 5 ppg above the Bronco’s pace.
Teams that completely snuck up on us in terms of offensive efficiency - Browns, Colts, Packers (New QB), and Cardinals - all were bottom-half efficiency teams in 2022.
Speaking of the Cardinals, they’re 3-0 ATS - and they’re getting 14 points on the road against the 49ers. Hard to not take the points. Meanwhile the Bears and Broncos are 0-6 ATS… We know that HAS to change (assuming the line holds at +3.5, a push this week would be comical).
The Underdog MLs (deep dive from last week) had opportunities with two +300 teams winning on the road (Colts +300 , Texans +300) and one winning at home (Cardinals +495). The model didn’t love those teams to win… but it would have been a nice little boost to have gotten even one.
Three of the games to hit the OVER had large spreads (Texans +7.5 v. Jags , Chiefs v. Bears +12.5, Cowboys v. Cardinals +11.5) - and one could argue that the Broncos +6 v Chiefs should have been larger. This pattern runs counter to the old line “Bet the under in large spread games” - which I’d always heard explained as either:
The game is a defensive struggle and the final score is closer than many expect.
One team gets ahead and stops scoring - so you end up with a 35-10 final that barely sneaks UNDER.
Is betting the under in large spread games actually a reality and could there be a viable way to take advantage of this?
The answer is no… not at all.
As we found in Week 2’s analysis of OVER/UNDER trends - 54.5% of games hit the UNDER. In order to even be a reality, we need to see more than 54.5% of large spread games - games with spreads of greater than +7 and +10 - hit the UNDER. I selected two spreads to determine if there was any trend between spread size and % of games hitting the UNDER.
In 2022, there were 76 large (greater than +7) and 32 XL spreads (greater than +10). In 2023, there have been 10 large and 3 XL spreads. In 2022, 52.6% of games with a greater than +7 spread hit the UNDER and 47.4% the the OVER.
If you include only spreads +10 or greater, the odds don’t change drastically - and are still below the 54.5% threshold - with only 53.1% taking the UNDER.
If you tried to take the UNDER in larger spread games in 2023, you’d see even worse returns - with 63% hitting the OVER.
NFL Week 4 Picks
Adjustments are in to the model to favor recent performance more than historical performance and efficiency data. We had a great pull hitting the OVER and LIONS on TNF, so we’re running in to Sunday up $20.
Share this post, leave a comment, or reach out to me on Twitter @dbfinsights