So I modeled the season (and you aren't going to like it)

[Happy Panther]

I'm in a contest where you pick every game for the whole NFL season. Instead of trying to slog through 256 games I built a stochastic simulation model that chooses a victor for each and every game during the season. My assumptions are below but the results are as follows:

 

(Remember computers have no feelings)

 

 

Highlights and lowlights from the Panthers: Some amazing familiarity here

 

1) We lose our opener 19-16 in overtime at Tampa

2) Lose to Detroit and Pittsburgh by 7 points total

3) Blowout Baltimore 43-18 to start 1-3

4) Lose by 1 to Chicago

5) Blown Out by Cincinnati with the first defensive hiccup giving up 35 points

6) Lose by 6 to Green Bay, 2 to Seattle and 3 to New Orleans. Now 1-8

7) Blowout Philly 35 -21

8) Lose to Atlanta 21-20... 2-9 on the season 

9) Much needed by week

10 And then a 5-0 to end the season missing the playoffs by 2 games. We average 32 points per game after the bye beating MINN, NO, TB, CLE, and ATL

 

[Jakob]

There is no actual way this team starts out 2-9.. sorry.

[rayzor]

afc east and north plus nfc west kind of all make me take your computer's projections with a few grains of salt.

[Happy Panther]

Assumptions if anyone cares

 

1) I assumed that each team in each games scores independently of each other. While this obviously isn't true, trying to correlate scores in individual games is too complex and probably doesn't make a difference.

2) I assumed the average score is 24 and there will be roughly 1 team that scores under 12 and 1 that scores over 35 on average. I generated 10,000 scores (normal distibution) and each game randomly picks 2 scores from those 10,000

3) I assigned a relative strength to each team based on the previous season's results plus an assumption that good teams will get slightly worse and bad will get slightly better.

4) For each individual game the better team gets spotted a certain number of points based on strength. As an example Carolina gets 4 extra points in the opener (vegas line is -2.5) while Cincinnati gets 10 against Houston.

5) While i wanted the simulation to be random I also had to set some filters to make the choice itself random. So I demanded that 6 teams have a record at least 4 games different than the previous year and that both Seattle and Denver have winning records.

5) Once every game is simulated you just add up wins and losses.

 

[PhillyB]

[Pejorative Miscreant]

New England is last in their division. Conventional wisdom says you loses big time credibility with that one. Confidence in your model is suspect. Sorry

Sent from my iPad using CarolinaHuddle

[Happy Panther]

Hey if last year I told you guys Houston would go from 12 wins to 2 wins, Redskins would go from 10 wins to 3 wins, the Chiefs would go from 2 wins to 11, and the Falcons would go from 13 to 4 wins you would have laughed as well.

 

 

[RoaringRiot]

You and your computer can go to HELL Happy Panther.

Kidding.

Sent from my iPhone

[sharkkiller]

So how many times did you run this simulation?

 

Doesn't really matter, because having any faith in computer simulations is not very smart. No computer simulation would have had us winning 12 games last year.

[Happy Panther]

If it would make you feel better I can hit the button a few more times until it spits out 13 wins

[TheRumGone]

fug you

[cardiackat88.]

Lol @ Bills 11-5 and Pats 8-8

[sharkkiller]

If it would make you feel better I can hit the button a few more times until it spits out 13 wins

 

And that proves that it is bunk.  If it changes that much with each time you run the program then it obviously has some major flaws. Unless it is consistently putting out fairly similar numbers, you have some very big flaws in your formulas.

[Chief Keek]