Here, hopefully, are stats you can trust

Sep 14, 2004 6:27 AM

For the past several seasons we’ve run a feature on the web site each week that looks at the "Week that Was." The statistics from the most recent slate of NFL games are used to project a "stats score" for the game based on the statistics.

Once in a while, the stats score has the team that lost projected as the winner based on what the numbers translate to in scoring. This is a situation we dub "The Wrong Team Won." In other words, the team that won the game might considered lucky since they didn’t have the numbers to back it up!

Then again, it could be a sign of a team that knows how to win even when things don’t go exactly as planned. Carolina last season had a number of games where the stats said they should have lost but the scoreboard showed them winning.

While ”˜wrong team won’ games can have significant implications for the following week, you can also find a lot of trends based on rushing/passing splits. At TwoMinuteWarning we like to go beyond the conventional stats. Here are key points we like to track:

PSR:”˜Play Success Rate.’ Every play is recorded as either a success or failure based on the situation (so gaining four yards on 3rd and 3 is a success, gaining eight on 3rd and 10 generally isn’t).

BIG: ”˜Big Gain Percentage.’ Track how often a team picks up a big chunk of yards, specifically 10+ yards on a rushing play, or 20+ yards on a passing play.

YDS: Adjusted yards per play. We include penalty yardage (so a 28 yd pass interference play counts as a 28 yard pass completion), and throw out non-plays like QB kneel-downs and spikes to stop the clock.

Play success rates were about the same whether teams ran or passed, but the ”˜adjusted yards per play’ was clearly better for passes.

With Week 2 coming up, here’s some history over the past five years on how teams have returned in the second game of the season after certain events:

Now some of these stats will be hard to calculate yourself (we do post a free recap of every game on the site), but you can get a reasonable sense of the adjusted yards, and can check in the detailed boxes for ”˜longest gain’ and find teams that had zero 10+ rushes or 20+ passes.

Some of the best ”˜comebackers’ in the second week appear to be:

”¡ Underdogs off good rushing games (whether 50 percent + PSR or 5.0 + adjusted yards per rush)

”¡ Go against teams that couldn’t break a single big gain either rushing or passing.

”¡ Bet on away teams that didn’t allow a single big gain either rushing or passing

”¡ Go against home teams that allowed a high passing success rate (50 percent +)

These sample size are of course very small, and should only be a part of your overall handicapping for Week 2 throughout the season ”˜digging into the detail’ of box scores and other statistics can lead you to insights that the typical sports bettor won’t have at hand.