Welcome! Log In Create A New Profile

Advanced

respectfully disagree

October 15, 2016 11:08AM
Quote
Coy Bacon
I've worked quite a bit with statistics. If anyone who reported to me were to say, "I have a sample size of five but I'm discounting one because I feel it is an outlier," I would tell him he doesn't know what he's doing.

Except that's too narrow for me. It doesn't account for many many things. Remember this is football. People who use stats right in football are tracking something they also SEE. So to come down with stats theory from other domains is often irrelevant.

First, DID THEY CHANGE AFTER SF YES OR NO. If yes then the numbers ought to differ. That just stands to reason.

And of course I put in all the proper qualifications. "If" appears a lot.

AND FINALLY who are those stats for? Why does one of the most reputable sports stats sites do "the last 3 games"? You want the simple honest truth? It's for people who bet on football.

People who bet on football DO KNOW that a team can change suddenly for any number of reasons (key injury for example) and they would simply not be disdainful of that information because they would want to know the difference it makes.

Plus as I said IF they did change after that game then it is bad analysis to IGNORE that.

And btw "sample size" has nothing to do with it. That's the most abused concept in football conversations I know of. It is always understood that NFL stats measure something that is a moving target and open to change. We know that. So is there such a thing as 3 game trends? Absolutely. Watching stats across a season is like tracking a chart in progress. It's only people who dismiss stats who think that people who use stats are acting like the numbers are set in stone dogmatic truths. But that's basically a strawman. (Eg. What was the offense like last year when Foles melted down. That was just 4 games but you do not dismiss the info you get because of that problematical "sample size" thing people use improperly in football discussions. Far more often than not when people say "sample size" what they REALLY MEAN is "I don't agree so I won't count this." It's seldom anything more than that.)



Edited 1 time(s). Last edit at 10/15/2016 11:16AM by zn.
SubjectAuthorViewsPosted

  Theo Riddick ruled out for Sunday

Rampage2K-681October 14, 2016 01:25PM

  Stars are aligned for Rams win...

max392October 14, 2016 01:34PM

  Re: Stars are aligned for Rams win...

Rampage2K-388October 14, 2016 01:46PM

  that's true...and demonstrable with the stats

zn385October 14, 2016 01:54PM

  Re: that's true...and demonstrable with the stats

Rampage2K-387October 14, 2016 02:04PM

  Re: that's true...and demonstrable with the stats

zn339October 14, 2016 10:46PM

  Re: that's true...and demonstrable with the stats

zn430October 15, 2016 12:16AM

  Re: that's true...and demonstrable with the stats

Coy Bacon325October 15, 2016 10:22AM

  respectfully disagree

zn324October 15, 2016 11:08AM

  Whew ...

alyoshamucci401October 14, 2016 02:39PM

  Re: Whew ...

JYB378October 15, 2016 02:34PM

  Rams could and should win this game...(NM)

PaulButcher59286October 15, 2016 08:42AM

  Re: Theo Riddick ruled out for Sunday

RFIP273October 15, 2016 10:51AM

  Re: Theo Riddick ruled out for Sunday

moklerman300October 15, 2016 11:10AM