Andrew Luck's 443-yard comeback win over the Chiefs is likely his first of many legacy-building games.

Luck also threw three picks earlier in the game, two of which were terrible decisions - short outs with little upside and lots of risk and on first down where there was no urgency to complete a short pass. But in some ways it made the flawless comeback - where the Colts were able to get down the field so rapidly on drive after drive - more remarkable. Few players have the mental toughness to shake off mistakes so easily and stay focused so consistently.

Luck also showed great pocket instincts, sliding away from the rush and throwing accurately on the move. And as usual he scrambled well when he needed to, getting 45 yards on seven carries. Russell Wilson has been great for two years, but is also playing with an elite defense and top running game. Like Ben Roethlisberger his first two years, Wilson has excelled while facing a lot of "hitter's counts." Luck, on the other hand, has had to deal with a weak running game, an average offensive line and a weak defense. If I had to pick one to start a franchise I'd take Luck by a small margin.

T.Y. Hilton had a monster game against the Chiefs, and is Indy's version of DeSean Jackson - a small, lightning-quick game-breaker. Most of the league's elite receivers are big, but there's still room for the guys who are hard to stop in two-hand touch.

Trent Richardson looks like he could get cut after the season. It's incredibly odd the Colts were so terrified of playing Donald Brown that they signed Ahmad Bradshaw (despite having Vick Ballard) and traded a first-rounder for Richardson once Ballard went down. Brown (who also fumbled) but was bailed out by Luck, is fast, stout and has a first-round pedigree.

The Bengals probably won't go deep in the playoffs with Andy Dalton at quarterback, and though they've made the postseason three years running with him, they need to find his replacement in the draft. If you think finding a replacement quarterback as good as Dalton would be tough, it's not. Currently 13 QBs are clearly better: Peyton Manning, Aaron Rodgers, Tom Brady, Drew Brees, Philip Rivers, Ben Roethlisberger, Russell Wilson, Andrew Luck, Tony Romo, Matt Ryan, Nick Foles, Cam Newton and Colin Kaepernick. At least five more: Alex Smith, Joe Flacco, Eli Manning, RGIII and Jay Cutler are probably better. That makes 18. Then Matt Stafford, Carson Palmer and Ryan Tannehill are at least as good. That makes 21. Finally, Mike Glennon, Matt Cassel, Kyle Orton, Sam Bradford, Michael Vick and Ryan Fitzpatrick certainly aren't much worse. Basically, there's very little risk to replacing Dalton.

Speaking of the Bengals loss, it's pretty odd Cincy was 8-0 SU and 8-0 ATS but lost by 17 to a seven-point dog at home in the playoff game. It just goes to show how worthless trends are as standalone analysis. While previous performance is predictive to a large extent, final scores of games are a heavy-handed way of capturing it. Of course, even if you had a method that captured past performance perfectly, it would only be a strong indicator of what will happen, not a cause. This is obvious, but I think it's something our minds easily confuse - that somehow Seattle will beat NO "because" Seattle is so good at home and NO bad on the road. That might be reason to think Seattle will beat New Orleans, but it's not why they will. If they do it's because on that particular Sunday they'll have to execute their plays more successfully. Once you realize the past has zero bearing on how a subsequent and totally independent game will transpire, it's a lot easier to see how the results of subsequent games might diverge widely from their precedents.

One other note re: playoff drafts: It's kind of crazy I penciled in Dalton and A.J. Green for at least two games while I considered Kaepernick and Vernon Davis 50/50 to play a second one. After all the line in the Bengals game was only seven which roughly translates to a 75/25 edge, while the line in SF-GB was three or 60/40. When 60 is rounded to 50 and 75 to 100, it forces you to push Dalton-Green way up the list. While the difference between 60 and 75 is significant, it's obviously a lot closer, and now that SF is favored in CAR this week (while Cincy would surely have been a dog in NE), their prospective games played were probably close to equal. At some point, though, when the odds cross 70 or 75 percent, it's easy to treat the more likely outcome as a done deal, and the more ambiguous outcome as totally uncertain. I'll have to re-remind myself not to make that mistake next year.

I honestly was only half paying attention to the Eagles-Saints game, given the huge emotional investment I had in the Chiefs-Colts. But there were some interesting strategic notes toward the end. Up three with eight minutes left in the game, Sean Payton settled for a field-goal on 4th and 4 from the 17 to put New Orleans up six. Some of the sabermetric football guys I follow on Twitter criticized the decision. For one, if they make the first down, they have a good chance to go up two scores, which would be huge. And second, the difference between being up three and six is negligible.

On the first point, I agree, and think maybe that makes it worth going for it. But on the second, I think they're wrong, and this makes it a closer call. If there were three minutes left in the game, then yes, being up six is almost worse because it forces the other team to go for the win rather than the tie and use all four downs to get there even when they're in FG range. While they still have a smaller chance of succeeding, if they do, it's game over rather than a 50/50 overtime.

But with eight minutes left the equation changes. If the Eagles, down six, score a TD - and they did - then NO is only down 1, rather than 4 with a couple minutes left. That puts them in a position of only needing a FG rather than a TD, and being able to run the clock all the way down - which is exactly what happened.

Of course, it didn't have to be that way. After Darren Sproles' return and the horsecollar penalty gave the Saints great field possession, it was time for the Eagles to consider letting the Saints score a TD. Not right away, but at some point with two-plus minutes left and Philly down around the 25, the ease of NO making the game-winning FG starts to outweigh the difficulty of PHI driving for a game-winning TD after letting NO score. But you have to do this before it becomes a no-brainer. By the time the other team has 1st and 10 inside the 20 with a minute left, it's too late. You can let them score, but they might stop at the one-yard line, making the FG even easier.

The Packers-49ers game was a very tough call ATS, and now I know why. It turned out to be a push. I thought Aaron Rodgers played well, and the Packers had no problem slugging it out with San Francisco. The killer - like last year - were Kaepernick's seven runs for 98 yards, many of them devastating third-down conversions. As in the Eagles-Saints game, the Packers should have let Colin Kaepernick run into the end zone when he scrambled for 11 yards and a first down at the Packers 27 with 1:07 left. It was a long shot for GB to score, but it did have two timeouts and Aaron Rodgers. Instead, they let the 49ers run out the clock and gave an elite kicker an easy shot to win it.