Stats: Maybe these should be used more sparingly in "discussions"?

[JSn]

Ok, check out the changes in the Pats stats in ONE GAME:

New England Patriots Blog - Pats soar in key statistical rankings - ESPN Boston

Points scored
7th (27.1 per game)
Last week: 18th (20.8 per game)

Points allowed
3rd (15.1 per game)
Last week: 11th (18.2 per game)

Third-down offense
8th (37 of 84, 44 percent)
Last week: 10th (32 of 73, 43.8 percent)

Third-down defense
19th (26 of 68, 38.2 percent)
Last week: 24th (24 of 58, 41.4 percent)

Turnover margin
Tied-2nd (Plus-8, 12 takeaways, 4 giveaways)
Last week: 9th (Plus-3, 7 takeaways, 4 giveaways)

Red-zone offense (based on TD percentage)
23rd (12 of 26, 46.2 percent)
Last week: 25th (9 of 20)

Red-zone defense (based on TD percentage)
Tied-16th (opponents are 8 of 15, or 53.3 percent)
Last week: Tied-20th (opponents are 8 of 14)

Some pretty amazing shifts for week six. Are going to hammer opponents like that every week? Not likely. I just think it's interesting and also a good reminder that Stats can be pretty misleading sometimes.

 

[Synovia]

Ok, check out the changes in the Pats stats in ONE GAME:

New England Patriots Blog - Pats soar in key statistical rankings - ESPN Boston



Some pretty amazing shifts for week six. Are going to hammer opponents like that every week? Not likely. I just think it's interesting and also a good reminder that Stats can be pretty misleading sometimes.

Total yards stats are useless anyways.

 

[BelichickFan]

Your point is valid - however the other side of the argument is that the winless Titans added a realism to our stats as every other team we'd played until then was undefeated when we played them (and all playing well at the time other than opening night) so our stats were skewed to be worse than we were probably playing.

 

[apple strudel]

NFL stats are a tough nut to crack because of the sample size. If there were more games played then it might be more interesting. Things like success of short yardage conversions over 10 years are more interesting, but stats like defensive ranking after 5 weeks don't really tell you all that much.

That said, it is still possible to have an intelligent discussion about them when used in context and when they are well understood. They work less well when they are used as an attempt to win one of the frequent pissing contests as though some poorly understood or out of context stat is a hard truth.

 

[Synovia]

NFL stats are a tough nut to crack because of the sample size. If there were more games played then it might be more interesting. Things like success of short yardage conversions over 10 years are more interesting, but stats like defensive ranking after 5 weeks don't really tell you all that much.

Which is why its better to used PLAY BASED stats instead of GAME BASED stats. It brings the sample size from 16 to roughly 1000 samples (on each side of the ball) per season. (which is one of the many reasons DVOA is so much more telling of a stat than Yards per Game)

 

[everlong]

Offense

Team  	  	                G  	 	Pts/G  	       Yds/G  	 	PassYds/G  	 RushYds/G  	 1stD/G  	 3rdM  	       3rdD%  	 	4thM  	 	4thD%  	 	Pen  	 	PYds  	 	TOP  	 
	 
Indianapolis Colts 	  	5  	 	27.4  	 	404.8  	 	326.0  	 	78.8  	 	20.8  	 	30  	 	50.8  	 	2  	 	40.0  	 	28  	 	202  	 	28:36  	 
New England Patriots 	  	6  	 	27.2  	 	404.7  	 	288.3  	 	116.3  	 	24.8  	 	37  	 	44.0  	 	5  	 	50.0  	 	34  	 	281  	 	35:00

Defense

 Team  	  	                 G  	 	Pts/G  	 	Yds/G  	 	RushYds/G  	 PassYds/G  	 Int  	 	IntTD  	 	FFum  	 	DefTD  	 	Tack  	 	PD  	 	Sack  	 
Indianapolis Colts 	  	5  	 	14.2  	 	295.6  	 	103.2  	 	192.4  	 	4  	 	0  	 	6  	 	0  	 	245  	 	26  	 	13.0  	 
New England Patriots 	  	6  	 	15.2  	 	293.3  	 	112.8  	 	180.5  	 	4  	 	0  	 	7  	 	0  	 	245  	 	26  	 	10.0

SOS does favor the Pats at 18-17 vs 10-17 (Colts playing one less game makes the gap seem large). I did find it interesting that the Pats are averaging more rushing yards per game. Even the penalties and penalty yards are the same once you factor in NE's additional game. The Colts also caught Seattle without Hasselbeck and befitted from playing the Titans as well and the Pats caught the Jets at the worst possible time.

So yes you cannot just look at last week but on the whole you are what you are.

 

[BradyFTW!]

Ok, check out the changes in the Pats stats in ONE GAME:

New England Patriots Blog - Pats soar in key statistical rankings - ESPN Boston



Some pretty amazing shifts for week six. Are going to hammer opponents like that every week? Not likely. I just think it's interesting and also a good reminder that Stats can be pretty misleading sometimes.

Of course there's going to be a huge swing. We just laid a historic beatdown early in the season. Small sample size + egregious outlier = skewed data. This is exactly why the people who know how to work with stats on this forum warn against small sample sizes pretty much constantly. Stats aren't misleading so much as they're often misapplied by people who don't know what they're doing.