The Lions have well and truly plumbed the depths of how to blow a promising team’s playoff chances. I guess on the plus side, they’ll get a slightly better draft position than they would have and hopefully they can correct what’s gone wrong so far this year. In terms of what that is exactly, my subjective feeling is that Stafford isn’t as accurate as he was last year; they need to get rid of that side-arm thing that he’s picked up. They also seem to run in inopportune places, which is kind of amazing for a team that passes so much in the first place. Objectively, the offense may not be the problem. Advanced NFL Stats has Detroit in 7th for total offense in the league. It does fit with my impression of the run game, which is in the bottom third of the league in EPA, EPA per play, and WPA. Run success rate isn’t bad though, which maybe implies that they’re picking ok times to run but simply don’t get any big plays. I believe I heard an announcer say that the Lions had their first run of 20 yards or more on Thanksgiving, so perhaps that’s the issue. The passing game certainly is explosive though, as the Lions are top 10 in every category except for success rate, where they are 11th.
Instead, the Lions are failing on defense. Their overall success rate is ok, but their EPA and WPA is bottom third. The blame lies equally on the run and pass defense, as both are bottom third, although pass WPA is particularly bad. It appears the Lions have been giving up particularly high-leverage passes. So the improvement for next year needs to come on the defensive side of the ball. I’m willing to believe in the front four for another season, so I’d personally like to see some help come in to the secondary.
I might have to look into some help for my models for next season as well. The picks have not been up to their standard from the last two years. As I talked about in a previous recap, it isn’t easy to tell if something has gone wrong or if it just happens to be an off year. But I have some ideas of changes or potential improvements to look into, so we’ll see if they turn out to be actual improvements and/or necessary in the future. With that, here are the results from the weekend.
The over/under picks were average, which is an improvement for them. Luigi was 9-7, Yoshi 1 8-8, and Yoshi 2 10-6. That puts their season records at 62-92-6, 78-76-6, and 68-86-6. They aren’t usually this bad, but one reason that the over/under picks tend to be basically guesses might be that my models rely largely on rate stats, like pass yards per play. If I included or limited myself to using the actual counts, they might do better. Intuitively I would say that it makes sense that you would do better knowing how many plays a team runs in general, since more plays or possessions (coupled with a decent yards per play) should mean more scoring. That’s one thing I’ll definitely look at in the offseason.
It was a rough week for picking outright winners, as Luigi was 9-7, Yoshi 1 was 10-6, and Yoshi 2 was only 7-9. On the other hand, as I’ve said before, you expect there to be upsets sometimes; you don’t want to be 100% correct. The season records are 99-60-1, 94-65-1, and 93-66-1. Unfortunately, the upsets didn’t quite add up enough to be profitable this week. A couple of close calls, like Tampa or San Diego, would have swung it. Luigi was 6-8 for a season record of 44-100-1, Yoshi 1 was 4-10 for a season record of 48-91-1, and Yoshi 2 was 4-10 for a season record of 45-99-1.
Picking every game against the spread, we would have had two pushes. Detroit and Houston pushed with the overtime field goal, and Tampa and Atlanta pushed when Atlanta managed to win by one. So overall Luigi was only 4-10-2 for a season total of 66-87-7, Yoshi 1 was 7-7-2 for a season total of 86-67-7, and Yoshi 2 was 5-9-2 for a season total of 70-83-7. Fortunately, some of those games would have been skipped for being too close to the spread. In pickable games, Luigi was 4-6-1 for a season total of 57-59-6, Yoshi 1 was 6-6-1 for a season total of 69-54-6, and Yoshi 2 was 4-6-1 for a season total of 56-61-6. The consensus picks were 2-4-1 for a season total of 40-35-3.
As you would guess given the generally poor spread picks this week, the Hilton SuperContest picks weren’t especially great. The predictions also weren’t particularly far away from the Hilton lines; Luigi only had 9 games that would have been considered ‘pickable’, and only one was more than three points off. So Luigi would have gone with Kansas City +10.5 (win), Jacksonville +3 (win), Miami +3 (win), Philly +2.5 (loss), and New Orleans +2 (loss). Another 3-2 week puts the season total at 29-24-2. Yoshi 1 would have gone with Oakland +8 (loss), Rams +2.5 (win), Washington +3.5 (win), KC +10.5, and Minnesota +4.5 (loss). That 3-2 week puts the season record at 27-27-1. Yoshi 2 would have gone with KC +10.5, Jacksonville +3, Miami +3, New Orleans +2, and Philly +2.5 (the same slate as Luigi but in a slightly different order). 3-2 puts the season record at 29-24-2.
Finally we have Bill Simmons. There was one push with his picks instead of two; he had the Lions game at +3.5. I also had a typo in his picks; Philly was getting 2.5 points instead of giving them. So he was correct to take Carolina but all three models would have gone with Philly. All in all Bill was 8-7-1 for a season record of 85-71-4, Luigi was 5-10-1 for a season record of 73-83-4, Yoshi 1 was 7-8-1 for a season record of 82-74-4, and Yoshi 2 was also 7-8-1 for a season record of 70-86-4. With five weeks left, I have 80 games in which to make up 12 picks on Bill. That’s a tall order, but I’ll be happy enough with rattling off a series of even decent weeks, like 9-7 or 10-6. Here’s to hoping.