For those unfamiliar with our turnover research, you may want to review the following prior articles for background on this theory:

• Turnover Difference Revisited
• Turnovers: Fumbles vs Interceptions

Week Two games (as we consider these week three events since the second week action was cancelled) are not included in all the data in the above research, so the first question is whether there has historically been any positive results against the spread playing as usual the WORSE net turnover team provided there is at least a half a turnover per game difference between the two teams:

The answer is clearly "NO" there isn't any predictive valye to the week one turnovers in attemptin to forecast second games of the season. We're not ready to call it a day just yet though, so next we'll break out the turnovers by using only fumbles and then only interceptions. Rather than produce the whole table, we will summarize the results as follows:

Category	All Plays	Win%	Comments
Net Fumbles (lost)	88-97	47%	18-6 in 10+ spread games, 29-47 in 5 to 9.5 spread games
Net Interceptions	109-100	52%	63-40 in small spread games (<5 pts), 15-8 with away favorites

We can dabble a little in angle creation with the above -- with fumbles concentrate on games with a big favorite. With interceptions look for away favorites or small spread games.

All right, let's take a look at the week 2 match-ups, but be warned that this is going to be complicated: the net numbers are expressed based on away stat minus home stat, and so if it's negative it means the home team is better, but the t/o method calls for you to bet on the poorer team.....basically negative=play away team, positive=play home team!

ANALYSIS: There's only one game that meets the 10+ favorite requirement for the "Fumble week 2 angle", which says play the Bills. There are many cases of interception plays on small spread games or away favorites.

How seriously can we take these picks? Not very. The turnovers off one game hardly constitute enough to be meaningful as the 114-113 record from '83-00 would suggest. Our angles are purely backfitted, so cannot be deemed reliable. Use the above only in conjuntion with other handicapping elements to either support or oppose a play.

Don't despair though, once we get a few more games into the season, the turnover analysis should provide more potent fodder for wagering!

Week One Results: 7-5 overall with Net Turnovers, 5-2 with Net Fumbles, and 1-1 on the "key plays". (Interceptions were not isolated since there has been no historical week one predictive value using them).