Saturday Feelings - Week 3

Here we are back on Saturday with a full lineup of predictions for the weekend. The nervous-excitement feelings I felt around the Week 2 predictions are still lingering but I was very pleased with the results. I was nervous that many of my picks would be incorrect, but actually, the two most surprising outcomes were very close to being correct!

Week 2 in Review:

I had a couple comments saying “wow Chargers over Chiefs? That’s crazy!” And I thought so too. But the Chargers led by Justin Herbert actually took the Chiefs to overtime and fell just short of upsetting the reigning world champions. The other surprising pick was the prediction for the Raiders to upset the Saints on Monday Night Football - and it happened! I was ecstatic. There must have been something in the data that ticked off the model to predict that outcome. 

Overall, from a moneyline perspective, the model was 13-3, predicting the outcomes correctly for a vast majority of the games played. From the perspective of spread and over/under, the model did not fare nearly as well - pretty much tossing a coin at each option would have been equivalent to the model’s prediction. 

There are some tweaks I am thinking about making to the model, but we’ll see if that is something worth doing. Check back in a couple of weeks for this. 

Week 3:

For Week 3, I am feeling ready to see the predictions come to life. I am hopeful they will have some accuracy and edge over the games, but at the end of the day, there is so much randomness in these football games that make them so hard to predict. 

The Thursday Night Game was bittersweet… because of my moneyline performance last week, I decided to put money on the Jaguars playing against the Dolphins. Unfortunately, Fitzmagic had a different plan than my model, because the Dolphins absolutely destroyed the Jags by a score of 31-13. My model did predict Dolphins +3 but it also predicted the Jags Moneyline and the over 48 call, neither of which happened. 

Surprises:

This week the model is predicting a couple upsets that I’m not sure how to feel about. The first is the Buccaneers to lose to the Broncos 27-21. Is it possible? Sure, but I can’t imagine Brady and the Buccs falling to 1-2 after 3 games. The Broncos have been cold recently and are off to an 0-2 start, so maybe they’re due for a win. I’ll be cheering for the Broncos but we’ll have to wait and see about that one. 

The other surprises the model is predicting are not too crazy, but I’m not sure as a fan I agree with so much. The model has the Titans losing to the Vikings by 1 and the Rams defeating the Bills by 1. The Bills are playing pretty well through two games, but so are the Rams! So it’s not beyond imagination for the Rams to pull off a win. However, the Vikings are coming off two straight losses and will be facing off against the Bills who are 2-0. The thing that is keeping me interested in this potential upset is the fact that the Bills only beat the Dolphins by 3 last week. Either way, I hope the model predictions are close this week!

Check out the instagram for tidbits of machine learning knowledge and to stay up to date with prediction updates!

Have a great weekend everyone. 

And, as always, good luck bettors!