Bounce here or there is the NFL. They've done a ton of numbers work that shows that some important NFL stats are totally random and are based on luck. For example, they do not believe that forced fumbles are a skill but do have a huge effect on outcomes. That typically, teams that have a lot of forced fumbles tend to win more but that their numbers tend to regress to the mean the next year and they get much fewer forced fumbles the next year. There's a number of NFL statisticians who have actually tried to quantify how much of the NFL is luck, and in a short 16 game season where players get hurt for no reason other than chance, a ton of it is luck.
They do not believe the winning close games is a skill that is associated with winning but rather has a huge luck component. They believe that the best predictor of winning is crushing bad teams. That the teams that win the Super Bowl tend to be teams that crush the bad teams versus teams that win close games against good opponents. Crushing bad teams regularly is an indicator of dominance.
Lots of interesting things over at the FO website that challenge commonly held assumptions. I don't buy into all of it, and know the limits of numbers.
I don't think Bill has a big opinion on what the Texans need to do to take the next step--but yeah, he was skeptical of them doing well this year because of the roster youth, lack of continuity of players on defense, his belief that the Texans were likely to have more injuries on the offensive line because last year they were unusually healthy. He does say that young defenses tend to get better all at once, but he didn't see that happening for this year, and he will need more data to make a prediction for next year.