NFL Week 3 Newsletter and Picks

Week 2 was the Orb’s all-time worst performance, and we are not here to shy away from it. If we didn’t show and break down our losing weeks, especially our historically bad ones, how could you trust our winning weeks? The numbers are comically bad. One of the ultimate goals of using predictive modeling is to improve the odds of you simply flipping a coin to predict a binary outcome. A coin flip would have crushed our models this week. Another goal of modeling even on losing weeks is to have a safety net of accuracy so that bad weeks don’t turn into season-altering ones. Unfortunately, we failed at both these goals. Our season goals don’t change, with our main ones to finish the season above 52.4% against the spread and be profitable against moneyline picks, but we would be lying if we said we weren’t off to a bad start on them. What is just as bad as the performance is the timing. Ideally, you want to bank some winning weeks early in the season so that when a weekend goes against you, you have some good performances in the bank, and can stay on the right side of the profitability line for the year. Even more important than unit performance is reader confidence in the models and the project as a whole. Andy and I are thrilled that we’ve gained so many new readers since the end of last year’s NFL season so for many of you, this is the first you are seeing of our project. To have this performance to start out the year is getting off on the wrong foot and we hope to show you over the course of the season what our models are capable of.  

From a football perspective, it was a crazy, unpredictable weekend! Underdogs dominated as they covered 11/16 games and won 8/16 outright, upsetting the favored teams at +money. This included the Cowboys, Lions, 49ers, Ravens, and Eagles all losing outright while favored by 6.5 or more points as of the time I ran the models. Insane! I can’t remember an upset week like it off the top of my head. The real insult to injury for the Orb was the Eagles on Monday night. Already down 5.6 units after Sunday’s massacre, the models had one more pick in them as they all agreed the Eagles would take care of business at home. They were -240 odds to win, which translates to a 70.59% implied win probability from the sportsbooks. Now you might think losing a pick that theoretically has over a 70% chance of winning is bad, but don’t you worry it gets so much worse. According to Next Get Stats, with 1:56 left in the fourth quarter, the Eagles had a 99.3% win probability and somehow went on to lose the game. This resulted in the 7th most unlikely comeback in the league since 2016. Just the cherry on top of a performance we hope to never replicate. This game, and the weekend as a whole is why they play the game and why we watch. We use data to try and assign probabilities and predict outcomes but as we saw in week 2, truly anything can happen. 

One quick note regarding the elephant in the room, the Carolina Panthers. Our models predicted them to not only cover the 4.5 points but win outright. I’ve joked about how much this one team has hurt our model's performance like the -2 units they cost us this week. But I now legitimately feel bad for everyone involved in this situation. I feel bad for Bryce Young who since the game has gotten benched in what is the fastest I can remember a team moving on from a #1 overall pick. I feel bad for the team itself that had to make this move in order to try and get anything going offensively. And most of all I feel bad for the fans in Carolina who not only have to root for a dumpster fire team with a terrible owner but for the next 15 years will be reminded of how they traded the pick that became Caleb Williams only to not take CJ Stroud. Hopefully, Andy Dalton can get something going and the fans can have something to root for in Carolina but for now, this may be the last that we talk about the Panthers for a little while. But they were the least of the Orb’s problems this weekend. So in line with our main belief in transparency on honesty, here is just how badly our models performed:

Spread:

• Dolphins -2.5 ❌

• Lions -7 ❌

• Browns +3.5 ✅

• Panthers +6.5 ❌

• Titans +3.5 ❌

All spread picks: 1-4, 20%, -3.091 units

Moneyline:

• Dolphins ❌

• Lions ❌

• 49ers ❌

• Panthers ❌

• Seahawks ✅

• Commanders ✅

• Eagles ❌

All moneyline picks: 2-5, 28.6%, -3.605 units

Overall -6.696 units. Is that bad? It seems bad. It was a discouraging result, to say the least. We’ve had losing weeks before, and certainly will again but to dig ourselves into such a big hole to start the year is disappointing when a new season brings so much optimism with it. So what went wrong, apart from everything? Well, we had three games in which the models made a prediction on both the moneyline and spread outcomes. In all three games, it picked the same team to both win and cover meaning the moneyline and spread outcomes were tied together. In two of these games, it picked the favorites (Dolphins and Lions) to both win and cover. In the other, it was the Panthers to cover and upset. So when a favorite loses the game, by definition it can’t cover. And when an underdog doesn’t cover, by definition it also can’t win. Missing on all three games meant missing all six picks which cost us six units against our record. So hitting on games when we have the spread and moneyline picks directly tied together will lead to good results, and missing on them will lead to what we saw this weekend. The numbers are at least in our favor that we won’t miss every single game this happens in this season.

 Here is how each model performed individually against all the games:

Spreads:

Moneyline:

Interestingly, combining the models resulted in worse performance than any individual model on its own. This is obviously the opposite of the intention of our modeling strategy but at the very least, it is interesting to see the logic of the multi-model system still work inversely when all three models perform poorly. 

Our models have a long road ahead of them to try and win back the confidence of the reader and right the ship. But even in the face of our worst week and worst start ever, we like to stay positive here at the Orb! It is time to turn a new leaf and look forward, not backward. We are officially burying the week 2 game ball Rex Ryan style and are onto week 3. One bad week or ten bad weeks, we promise to keep working hard on the models and putting time and effort into our work. I looked into the Orb and this is what it told me will happen on Sunday:

Spread:

• Giants +6.5

• Texans -2.5

• Saints -2.5

• Broncos +6.5

• Ravens -1.5

• Monday Night Double-Header Picks (Premium)

Moneyline:

• Jets (TNF) ✅

• Giants

• Colts

• Texans

• Saints

• Buccaneers

• Seahawks

• Ravens

• 49ers

• Monday Night Double-Header Picks (Premium)

The Orb missed its first big upset call of the season last week (by only 24 measly points!) but is making another bold prediction for Sunday. All three models believe that the New York Football Giants have a chance to go into Cleveland and upset the Browns. They had a big win in Jacksonville, but the Browns are coming into this game with the 31st-ranked passing offense and just the 20th-ranked rushing attack. This equates to the 4th worst expected points added by an offense in the league only beating out the Bears, Broncos, and Panthers. The Giants have been far from perfect on either side of the ball but in a battle between two teams that score less than the league average and allow more points than average, the Orb is taking the points and the upset. It is scary having 2 units of our season performance on a team that is being handed so many points in back-to-back weeks, but we just tell you what the models predict. I can’t imagine this game makes it onto Redzone too many times but we will be keeping a close eye on this terrifying pick in what has the making of an ugly game.

---

---

The Orb is also going for its first Baker of the year! For new readers, a Baker is when the Orb correctly predicts both the underdog covering and the favorite winning. Meaning the spread and moneyline picks are won by different teams. This is our favorite outcome when predicted right because the two picks aren’t directly tied together so both our spread and moneyline models are independently right. We named the Baker after the man himself in Tampa Bay last season when he managed to heroically lose by 2 to the Texans, covering the +2.5 spread pick and also not ruining the Texans moneyline pick. How appropriate that our first Baker attempt of the year involves Mr. Mayfield, the somehow current #1 fantasy quarterback, as the Orb is taking the Broncos to cover and keep the game close but ultimately lose to the Buccaneers. 

Let’s hope the Orb can start to earn back some reader trust and win some units against the sportsbook starting this weekend. Thank you for continuing to read our work and supporting this data project in good weeks and in bad! Who are your favorite picks on Sunday? 

- Team Orb Analytics

.

.

.

.

.

.

DISCLAIMER:

The information provided on this website is for informational purposes only. It is not intended to be gambling or financial advice, and should not be relied upon as such. We are not responsible for any actions or decisions taken by readers based on the information provided on this website.

The picks and predictions provided on this website are based on our own research and analysis, and are intended to be used for entertainment and informational purposes only. We do not guarantee the accuracy or completeness of the information provided, and we are not responsible for any losses or damages incurred as a result of using this information for gambling or other purposes.

By accessing and using this website, you acknowledge and agree to the terms of this disclaimer, and you assume all risks and liabilities associated with your use of the information provided on this website.