Re: Week Seven NFL Picks and "Game of the Week"
posted at 10/23/2009 7:58 PM EDT
In Response to Re: Week Seven NFL Picks and "Game of the Week"
[QUOTE]In Response to Re: Week Seven NFL Picks and "Game of the Week" : Paul; Sounds like a well thought out system. Do you do it on Excel? I agree with all your principles except #2. First, statistically, the team ahead by the third quarter is usually going to win anyhow, so what's the difference netween using the final score of the 3rd quarter score as a factor? I disagree that teams play their best for 3 quarters and slack off in the fourth. Sometimes football seems a lot like BBall where the fourth quarter is the only one worth watching and the time when teams play their best. Also, counting the third quarter score only ignores the come-from-behind-factor. It would seem that if a team comes back to win a game, they should be credited with a win. Likewise the team that can't hold on to a thrid qiauter lead shouldn't get credit for a win.
Posted by unclealfie[/QUOTE]
Teams don't "slack off" in the fourth quarter. They play for different goals. That makes those plays noise, not signal, for the question, "which team is playing better most of the time?" I want to eliminate noise from the signal.
Every single play says something about the team under some circumstance or other. Almost all of the plays in the first 3 quarters (except the last minute of the second quarter) are about the team trying to gain points in general while trying not to lose points in general. I want that particular signal because next week, almost all of the plays through the end of the third quarter are going to be equally for that goal. In fact, in many games, all of the plays halfway through the fourth quarter are for that goal.
At the end of the fourth quarter, the situations change. Teams are playing prevent so that they give up yardage and give up field goals but not touchdowns. The losing team often looks tremendous all of a sudden against the Patriots' defense. It's not that real! Brady has gone something ridiculous, like 70-3 when the Pats are ahead in the fourth quarter, so BB has usually gotten what he really wants, a 1 point victory. The most egregious example was when the Pats, instead of punting, threw away 2 points by deliberately hiking the ball out of the endzone. They got a free kick from the 20, not from the back of the endzone, and went on to win the game, thank you. Another quite common example, when ahead by 9 or more points, is to let the other team drive for a slow touchdown 5 or 6 yards at a time. The other team gets 7 points but loses almost all of their time, and has to kick an onside kick which fails. The final completely bogus play is the winners' kneeldown, especially in field goal position.
On the losing side we see hail mary plays with many interceptions. Defensive players try to rip the ball out, miss the tackle and watch a touchdown take place.
Those are the plays that are noise. Yes, they often lead to points on the final score.
When games are pretty even, 4th quarter results are a valid signal. If I had time I'd sift through and use the plays from the close games, but they don't add that much additional solidity to my stats. I need an easy rule to get them all out of the way. So, 3rd quarter stats.
Overtime coin flips are an additional source of noise. The flip winner tends to get the points. I see some of the same desperate defensive gambles when one team is in field goal range, just before the other team kicks the winning field goal.
I agree that the fourth quarter (of a close game) is worth watching.
I regret that I have no come-from-behind factor. It's a blind spot in my stats. For example, Brett Favre would be a great come-from-behind factor, and certain teams have no such tools. However, much of this factor should equally be evident in the first three quarters. Brett, Payton and TB usually do their coming-from in the first three quarters, and then there's no deficit to make up. If I took every game play-by-play (which I don't have time for) I could have such a factor.
My simulation of a come-from-behind factor is the time of possession. The come-from-behind team has usually beaten up the other team's defense.
It may also be that some teams are more physical or just plain dirty than other teams, and that would be an under-the-radar come-from-behind factor. Also, certain teams may specialize in staying up all night rehearsing the magic gimmick play that perhaps wins the game. Finally, Miami and Denver are exhausting places for visitors to play, and that's a come-from-behind factor for the home team. However, there is precious little signal to mine for all of this. If there's no signal, it's not too usable here.
I never give any credit for wins. I give credit for next week's potential. That's my target.
Yes, it's in Excel for easy editing, with a Visual Basic routine in the background.