The Panthers entered Week 10 last in the league in offensive time of possession at 25:30 per game--more than two minutes behind the second-to-last team. Only one other team in the past decade (I didn't search farther back than that) has averaged less over a full season.
But last night, Carolina had the perfect matchup to right the ship. I had thought "perfect" would mean more P.J. Walker, not unlike what we saw two weeks ago when these teams last faced off. I also figured--as I shared in comments yesterday--that D'Onta Foreman was a good bet for 12+ points in a Carolina victory. A victory through the air and on the ground.
Instead, surely due in part to the weather, but also probably because it was a sensible alternative to their Week 8 slugfest, Carolina sought to control the clock. And they did, running their offense (literally and figuratively) for 34 minutes and 12 seconds. The strategy worked. If Atlanta had found the end zone in their opening two drives, it might have forced the Panthers to play more through the air. But the Falcons' offensive struggles (and Carolina's strong defense) allowed the Panthers to remain patient.
There's a trend I've been tracking since last year, and over the summer I ran the numbers back about a decade. (Yeah, a lot of my research starts by going back a decade.) My theory last year was that RBs who played on Sunday were more likely to regress if they played four days later on Thursday. And without going into too much detail, the data backs the premise. And the more Sunday touches an RB receives, the steeper his likely regression that next Thursday.
But there's another outcome that I wasn't expecting. I haven't documented it quantitatively yet. But for fantasy purposes, it's worth keeping in our back pockets:
Sometimes starting RBs earn considerably fewer touches on a Sunday when they're scheduled to play that next Thursday. It's essentially an inverse regression. I'm betting that teams are aware that feeding their bell cow 25 times one day, and then 22 times a few days later, doesn't often produce good results in the second game.
Late last season and into this season, I only considered the Sunday-to-Thursday regression. And when one of my fantasy RBs was playing on a Thursday, I factored the regression probabilities into my decision on whether to start that RB or go with a backup.
But then it became clear that inverse regressions are a "thing." Just last Thursday, Miles Sanders had 18 touches. Four days earlier, he earned only nine--the only time this season he's had fewer than 15 in a game. No doubt, he dominated on those nine touches. But the low volume is key. Boston Scott had seven carries for an inefficient 21 yards.
In that same Thursday night game, Dameon Pierce put the Texans on his back, rushing 27 times. Four days earlier? His only performance with less than 20 touches since Week 2. Yeah, Pierce still had 18 touches. Not bad. But it was a one-score game. I believe Houston made a decision to either work him hard on Sunday or on Thursday. Doing both might increase the risk of poor Thursday outing, or an injury, or both.
So last night, even with Chuba Hubbard returning, I was fairly bullish about Foreman. Carolina fed him only nine times on Sunday--well below his 22-touch average the previous two weeks. We could argue that it was a blowout loss, so he wasn't needed. And maybe that was part of the calculus. Maybe if the game had been closer, Foreman would have hit 22+ touches again, and then the team could look to the fresher Hubbard more on Thursday.
Regardless of the "if-then" possibilities, this needs to be tracked (and I track the results every week). It's not just about regressions. It's also about preparing for potential inverse regressions. The Packers and Titans are playing next Thursday. I'm betting if we see a lot of Derrick Henry this Sunday, we'll see more Dontrell Hilliard than usual four days later. The same goes for Aaron Jones and A.J. Dillon. Or . . . if Henry or Jones net, say, 16 or fewer touches, it might signal that the team is preparing to run them harder on Thursday.
OK, not exactly a recap of last night. But of everything that happened, I found the Foreman situation the most interesting. And I like when research is actionable. It isn't always. But in this case, the model played out roughly as expected (although Foreman did even better than I expected).
Finally, a big hand for Adam Noslekim. He was last week's TNF closest-score winner. And now he's the first back-to-back TNF winner in the history of closest-score competitions (I assume). He predicted 24-20 Panthers. And since a large majority of you picked the Falcons, you made my job easier tabulating the results. So thanks for that. And congratulations Adam.
But last night, Carolina had the perfect matchup to right the ship. I had thought "perfect" would mean more P.J. Walker, not unlike what we saw two weeks ago when these teams last faced off. I also figured--as I shared in comments yesterday--that D'Onta Foreman was a good bet for 12+ points in a Carolina victory. A victory through the air and on the ground.
Instead, surely due in part to the weather, but also probably because it was a sensible alternative to their Week 8 slugfest, Carolina sought to control the clock. And they did, running their offense (literally and figuratively) for 34 minutes and 12 seconds. The strategy worked. If Atlanta had found the end zone in their opening two drives, it might have forced the Panthers to play more through the air. But the Falcons' offensive struggles (and Carolina's strong defense) allowed the Panthers to remain patient.
There's a trend I've been tracking since last year, and over the summer I ran the numbers back about a decade. (Yeah, a lot of my research starts by going back a decade.) My theory last year was that RBs who played on Sunday were more likely to regress if they played four days later on Thursday. And without going into too much detail, the data backs the premise. And the more Sunday touches an RB receives, the steeper his likely regression that next Thursday.
But there's another outcome that I wasn't expecting. I haven't documented it quantitatively yet. But for fantasy purposes, it's worth keeping in our back pockets:
Sometimes starting RBs earn considerably fewer touches on a Sunday when they're scheduled to play that next Thursday. It's essentially an inverse regression. I'm betting that teams are aware that feeding their bell cow 25 times one day, and then 22 times a few days later, doesn't often produce good results in the second game.
Late last season and into this season, I only considered the Sunday-to-Thursday regression. And when one of my fantasy RBs was playing on a Thursday, I factored the regression probabilities into my decision on whether to start that RB or go with a backup.
But then it became clear that inverse regressions are a "thing." Just last Thursday, Miles Sanders had 18 touches. Four days earlier, he earned only nine--the only time this season he's had fewer than 15 in a game. No doubt, he dominated on those nine touches. But the low volume is key. Boston Scott had seven carries for an inefficient 21 yards.
In that same Thursday night game, Dameon Pierce put the Texans on his back, rushing 27 times. Four days earlier? His only performance with less than 20 touches since Week 2. Yeah, Pierce still had 18 touches. Not bad. But it was a one-score game. I believe Houston made a decision to either work him hard on Sunday or on Thursday. Doing both might increase the risk of poor Thursday outing, or an injury, or both.
So last night, even with Chuba Hubbard returning, I was fairly bullish about Foreman. Carolina fed him only nine times on Sunday--well below his 22-touch average the previous two weeks. We could argue that it was a blowout loss, so he wasn't needed. And maybe that was part of the calculus. Maybe if the game had been closer, Foreman would have hit 22+ touches again, and then the team could look to the fresher Hubbard more on Thursday.
Regardless of the "if-then" possibilities, this needs to be tracked (and I track the results every week). It's not just about regressions. It's also about preparing for potential inverse regressions. The Packers and Titans are playing next Thursday. I'm betting if we see a lot of Derrick Henry this Sunday, we'll see more Dontrell Hilliard than usual four days later. The same goes for Aaron Jones and A.J. Dillon. Or . . . if Henry or Jones net, say, 16 or fewer touches, it might signal that the team is preparing to run them harder on Thursday.
OK, not exactly a recap of last night. But of everything that happened, I found the Foreman situation the most interesting. And I like when research is actionable. It isn't always. But in this case, the model played out roughly as expected (although Foreman did even better than I expected).
Finally, a big hand for Adam Noslekim. He was last week's TNF closest-score winner. And now he's the first back-to-back TNF winner in the history of closest-score competitions (I assume). He predicted 24-20 Panthers. And since a large majority of you picked the Falcons, you made my job easier tabulating the results. So thanks for that. And congratulations Adam.