NFL 2003 Week Four Games

IMPORTANT: We do NOT recommend playing the overlays or the UPM picks this week! They are based on just three weeks of play and it is unclear how reliable they will be in week four events. Notes on this week's lines: there are several games this week as of Tuesday morning that are essentially "off the board" including Phi-Buf, Cin-Cle, SF-Min, and Det-Den. We have posted the data for these games with available lines, but be aware that the spreads you may find could differ significantly.

Week Four Game Pages Kansas City at Baltimore Philadelphia at Buffalo Atlanta at Carolina Cincinnati at Cleveland Jacksonville at Houston San Francisco at Minnesota Tennessee at Pittsburgh Arizona at St. Louis New England at Washington San Diego at Oakland Detroit at Denver Dallas at New York Jets Indianapolis at New Orleans Green Bay at Chicago

Week Three Wrap-Up

Week three games start to show a semblance of predictability from a statistical standpoint, but are still prone to exaggerated projections. Most of the tools that we rely on are deemed as ready to go from week five forward, and again we advise betting cautiously in the early stages of the season until we do hit the point where teams have all played three and preferably four games. The week as a whole was a good one for favorites, who went 10-4, moving the season-to-date tally to 28-16. Incidentally, home dogs were 1-5 and are just 1-10 on the year. Away teams were 8-6 winners, and hold a small 24-20 lead after three weeks. The unders were stong, finishing 10-4 and take a 26-20 lead in the season-to-date totals. Full Story: the third round of simulations could muster nothing better than a 55% confidence play. We are unlikely to see many good plays from the 'Full Story' until the week five mark is hit, at which point all of our tools can be brought to bear at full strength. Overall the results were 8-8. DC/TO: the DC/TO went 8-6 with all the picks (and 6-0 with the six highest overlays), but it's too soon to put much stock in the perceived value from these projected scores. Overlays: We stress that two weeks of data is not enough for statistical predictions, and the overlays in week three games served to prove this point, as the drive chart projections went 7-7, and the PBP projections also 7-7. UPM: the UPM was 5-9 overall, and 3-5 with the 65%+ plays. Again, we do not begin tracking the record officially until week five. Turnovers: The Turnover picks for week three performed well (4-2 with the full turnovers, 6-5 with fumbles, 6-4 with interceptions), but the long term history has shown only modest predictive value this early in the season. CSM: The Cumulative Spread Margin plays for week three went 5-5, but as with turnovers, the long term results have suggested this is not to be trusted. Trends: The Trend plays based on "week two outcomes" were 3-2. Angles: The angles plays were 1-1. Please note that this column has been discontinued. Inside the 20: The red zone handicapper elected to pass on the week three action. Injury Report: Injuries are one area where you can get interested a little earlier in the season we believe, and the injury picks went 1-0 on the week. There were no coveted "dual injury qualifiers" however. Over/Unders: The over/under columnist went 3-2 with totals plays. Best Bets: A mild 5-4 performance saw the DC/TO stake a claim as the best of the best bets in the first month.