I find it interesting how prevalent analytics are becoming in the NFL. Sites like PFF have becoming increasingly popular and Cleveland went as far as to hire a numbers guy to run their football operations. Obviously the analytics community(multiple outlets) hated the Jets draft, and time will tell whether or not they are right. My opinion.... Analytics doesn't really work in the NFL like it does in many other sports. Why? It's pretty simple actually, the sample size is only 16 games. In college it's 12. There are also simply too many variables such as strength of schedule, and penalties to accurate predict an outcome. Sent from my iPhone using Tapatalk
Analytics are a tricky bird. The whole point is to identify an aspect of the game that other teams undervalue and then exploit it. You do this by looking at the data and constructing a simplified model to produce results. For example, should I take a risk on a QB with small hands? An analyst would see if small handed QBs over or underperfom their draft slot over the history of the draft. Unfortunately some people pretend they're analysts and fool owners who don't understand what analytics really are. For example, Sam Hinkie in Philly brought in shiny new tracker techs and tanked for picks. Sounds good except that if you look at the data, tanking rarely works. A real analyst like those in Boston or GS would know that. He then used those picks on centers, which is stupid because the data show that a team should drive and shoot threes, which centers don't do. Football is the trickiest sport of all to analyze. We'll see if the Browns can pull it off. They're right to trade down, but it's concerning that they already are overemphasizing combine numbers (Goff's hands, Corey Coleman the workout warrior). That's what a faker would do, overvalue the most quantifiable traits rather than the most relevant ones.
The problem is that analytics are taken to be something they aren't. They are not the be-all end-all, which makes sites like PFF completely useless. They should be complements to a discussion. They should never take the place of what you see with your own eyes. Unfortunately, as technology grows more and more people find it easier to punch numbers into a calculator then they do sitting down and watching football.
The NFL is more random than most sports due to the injury factor. The number and probability crunching works just fine right up until the moment your LT or QB or SS is lost for the season and then things can go to hell in a hand basket if in a hurry if you're not lucky. It's true that teams that do unorthodox things, particularly in the draft, tend to wind up screwed in the end. The Browns taking a redshirt sophomore QB in Johnny Manziel proved the point for us yet again.
If the person who has them isn't an idiot, they lie a whole lot less than a number derived from a convoluted formula.
We've never really had good luck. I'd rather rely on a quality GM that has put the right pieces in place to deal with an unexpected loss. I'm optimistic Mac is that guy.
I tend to agree to an extent. It's a low sample size, plus there are 11 guys executing a play at once. It's difficult to rate individuals in that manner, but I think it works for units like the oline as a whole. But sometimes it is definitely over analysis. I swear that in the long run we'll figure out that most of the game is pure luck anyways.
The problem with analytics in football is when the data relies upon a subjective judgment call made some some geek watching a computer screen (PFF). For example, how do you grade a pass that should have been intercepted? There is no objective formula for doing so, which makes the data inherently unreliable.
Apparently there is a dude here who can tell if a LB was not running full speed so he could only get a finger on it but if he had been running full speed he would have intercepted it. And if the LB was Usain Bolt he would not only have intercepted it he would have had an easy pick 6. It's all about "what's on the tape" and "facts". _
