Analytics Are Wrong All The Time

[NoGain]

Why do we put so much stock in them? According to analytics, every Hawk fan should have either left the stadium or turned off the TV set when we were down 30-14 in the 4th quarter against the Rams. Analytics has made Lions coach Dan Campbell look borderline incompetent as an in game coach this season and in last year's playoff. The Seahawks have been consistently proving analytics wrong all season long when it comes to winning the division and becoming the top seed in the NFC.

Games are not won or lost on a computer program. Not all games and situations within games are alike. Whether to go for it on fourth down or punt, whether to kick a FG or go for it are often best decided by a variety of other factors that have little to do with analytic models.

Sure, there is a place for analytics, but we can't seem to get through 10 minutes of a broadcast without the analytics of some such situation being brought up. Something tells me what John Madden would have thought of them.

 

[ruffENrowdy]

This is exactly what I've been saying for years man. Thats why the whole Ai bs is stupid. Sometimes you just need a little bit of common sense or a gut feeling.

 

[BirdsCommaAngry]

Statistically speaking, how often would you say analytics are wrong?

 

[MontanaHawk05]

Analytics can't take into account what the coach knows about each player, what he knows about the opposing players, the way he wants to play the game, the things he's trying to set up, etc. There's a million factors a coach is taking into account that the Ben Baldwins of the world don't understand.

 

[Chapow]

Analytics should be used as just an additional bit of information to take into consideration when making in game decisions. Or to put it another way, just one factor among many.

Analytics should definitely not be used as essentially a replacement for an actual coach making in game decisions.

What I don't like about analytics is the way it's presented. Like it is definitively wrong to punt on 4th and 3 if analytics says you should go for it. Whether punting or going for it is the best decision is dependent on a lot of different factors. If you're up against a very stout defense, and your team is really bad at converting short yardage situations (the Seahawks last year), you should probably punt even if analytics says you should go for it (but again, this depends on a lot of other factors).

 

[hox]

Analytics helped MM with better time management in the Colts game. It also factored into why we went for the 2 pt conversion instead of the tie.

Smart intel = good but is only as good as how coaches are able to apply it to their specific game situation.

 

[sutz]

Analytics is probably most useful in-week, when prepping the game plan. In game, it should always be tempered with the coach's experience and, frankly, his gut feeling of what needs to be done and what is possible based on his knowledge of his own team's capabilities and the current flow of the game.

 

[NoGain]

I bet analytics had the Packers overwhelming favorites to beat the Bears late in the game tonight. LOL

 

[OneLofaTatupu]

This is exactly what I've been saying for years man. Thats why the whole Ai bs is stupid. Sometimes you just need a little bit of common sense or a gut feeling.

100-600 million neurons in our gut. They communicate directly to the 3lb electrified meatloaf in our skull via the vagus nerve. A lot to be said about gut feelings.

Nassim Taleb would probably also agree.

 

[evergreen]

Like the other guy said it helped MM save time to beat the Colts. Maybe it just saved time for them? The two point conversion was a no brainer because of the previous loss. I think it’s like the lines on the road. It’s information to use in your decision. It’s not God.

 

[Sperrydogg]

Dan Cambell: it’s 4th and go
Dans OC: no boss it’s 4th and GOAL….
Dan Cambell: you’re fired

 

[GemCity]

As a data scientist, I agree with the sentiment in this thread.

There are far too many unaccounted variables when it comes to analytics in football. Crowd noise? Momentum? Fatigue? Degree of difficulty for a particular play? A players stats…can a RB that runs a 4.4 have less burst than a guy that runs a 4.7? How much does that guy weigh? What’s the statistics of the folks blocking that hole or the #’s of the opponents trying to shut it down?

Analysis is a valuable tool but, imo…best used coupled with subject matter expertise.

I take stats that pop up during the game with a grain of salt.