In the first 'Winning Stats' post on Wednesday, we looked at a number of NFL stats four weeks into the season.
Some of those stats I labeled 'Ho-Humm Stats', because they appeared to have little to no correlation with winning so far this season. These were passing yards, first downs, sacks, penalty yards and passing yards per attempt. The question I asked that led to these results was: If you are better than your opponent in these stats, does that make a difference to the game outcome?
Today we'll take a different approach to these stats by asking a slightly different question: How much better than your opponent do you have to be in these stats to see a difference in the game outcome?
1. Passing Yards: The teams with more passing yards than their opponents have won only 28 of the 62 games played so far this season (45%). Looking at the passing yards differential (own passing yards minus opponent passing yards), shows that simply passing more is no guarantee that you'll win.
|Winning probability by passing yards differential, week 4, 2010
|Passing Yards Differential||1-50 yards||51-100 yards||101-150 yards||151+ yards|
The most striking example of that this season was the Colts' week one loss to the Texans. The Colts threw for a whopping 433 yards, while the supposedly pass-happy Texans only threw for 107, a difference of 326 yards - yet the Texans won the game 34 to 24.
2. First downs. First downs are a tricky stat, because they are likely more indicative of the style of offense than anything else. Big play offenses generate a lot of yards with a low number of first downs, dink-and-dunk or ball control offenses will embark on long, drawn-out drives with multiple first downs, but in neither case are the first downs indicative of a teams' ability to score. The breakdown below confirms this. There's not a lot of difference in winning percentage if your team has up to eight more first downs. At more than eight first downs, the number of first downs is probably more an indication of overall superiority.
|Winning probability by first down differential, week 4, 2010
|First down Differential||1-3||4-5||6-8||9+|
3. Sacks. On sacks we run into problems of sample size. In 46 of 62 games this season, one team recorded more sacks than the other team. The team with more sacks won 33 of those games (72%), but there is no clear trend when you look at the breakdown below:
|Winning probability by sack differential, week 4, 2010
4. Penalty Yards: The team that has more penalty yards in a game lost 56% percent of the time so far this season. This is not a lot, but looking at the penalty yard differential (own penalty yards minus opponents penalty yards), reveals why this remains a stat worth watching.
|Winning probability by penalty yards differential, week 4, 2010
|Penalty yards Differential||1-10||11-20||21-30||31-40||>40|
If your team gets penalized for up to 30 yards more than your opponent, do not worry. Statistically, this has little correlation with the game outcome. However, when you cross the 30-yard threshold, you should start worrying, as the odds of winning that game decline rapidly with every extra penalty yard tacked on.
5. Yards per pass attempt. YPA is a passing efficiency measure and it was a little surprising for me to see that teams that won the YPA battle only won 56% of their games so far this season. Nowhere is the question "By How much?" more relevant than here:
|Winning probability by YPA differential, week 4, 2010
An efficient passing offense, coupled with a good passing defense (or just an an opponent with an inept passing offense) gives you a pretty solid shot at winning.
Not so Ho-Humm after all
Raw passing yards and first downs don't seem to have a strong correlation with winning any way you look at the numbers. For sacks, penalty yards and YPA a different picture emerges when you look a little more closely.