That's a matter of semantics. Whether you want to state what statistics "say" or "mean," it's the same difference to me. The point is that statistics almost never tell the entire picture. They are a tool for those that are educated with background knowledge of the situation and have empirical experience to make inferences based on observed probability. Most of the time, pointing to raw numbers as absolute proof of anything is reaching unless the same numbers have been found in controlled randomly assigned situations many times by many independent entities. Even then, the "facts" may not hold up if a single minor component is changed among endless random variability and biases. Beyond that, it's just taking numbers and trying to fit them into a narrative or an opinion. Sure, the inferences made MAY be right sometimes, but they're very likely to be incredibly wrong many times, as well.
As I said, statistics and numbers are just an inference tool, but in the case of the NFL, very rarely, if ever, do they "prove" anything at all, imo.
Originally Posted by: doddpower
If you don't know what stats mean, you wouldn't know the difference.
There is a huge difference to me.
For example, a QB throws for 400 passing yards in a game. That says lots of passing yards. You would think that means the team that put up those yards was great.
What does that really mean?
It means they were either in a shoot out and have a 50% chance to win or were getting blown out and had no chance to win. Teams putting up 400+ yards a game actually lose about 75% of the time.
If you run a correlation of passing yards to wins, on a scale of 0-100, with 100 being a direct 1 to 1 correlation, passing yards would be a negative 1.4, or essentially no correlation.
That is the difference between knowing what they say and what the mean.
So would you say a QB throwing for 5000 yards in a season was great? I wouldn't.
I would look for a stat that correlated to wins. Like Passer rating. Which was about a 90 correlation.
A stat that when you are leading, you are winning.
Now the stat itself isn't important. But doing the things that increase your passer rating have a direct impact on wins. Throw lots of TDs, don't turn the ball over, get few incompletions and get a lot of yards per attempt. So that means an efficient QB is going to win more games than one who throws for a lot of yards.
Which is why ratios mean so much. Comebacks don't mean anything without a per attempt.
6 comebacks says a lot but means little until you find out that it was out of 13 tries.
34 career comebacks sounds like a lot until you find out that it was out of 100 attempts. More than twice as many attempts as the next guy with 38 comebacks.
Was he really good at comebacks? Or did he make up for sucking by sheer volume?
I want to go out like my Grandpa did. Peacefully in his sleep.
Not screaming in terror like his passengers.