Jump to content

Cynthia Frelund: Projected Final Win Totals for 2018


Recommended Posts

Projected Final Win Totals for 2018: Rams, Chiefs Riding High 

 

By Cynthia Frelund NFL Network Analytics Expert
 
0ap3000000981920.jpg&f=jpg&w=1280&c=80
 
Now that eight weeks of the 2018 season have passed, we have enough data about the identity and potential of each team to really start sorting out who will be contending in January and who will be picking near the top of the 2019 NFL Draft next spring.
 
Just as I did before the season began, I've used a model to generate projected final win totals for the 2018 campaign, providing a peek at how things could play out as we head down the final stretch toward the playoffs.
 
But first, here's a quick look at how my model works.
 
My model compares this season's games with 15 previous seasons of situational production metrics that led to wins and losses (between 2003 and 2017). Tracking personnel, matchups, play-calling and results from past seasons establishes historical "profiles." The results from the games that have already been played this season are then collected and analyzed in the same way, with the model revealing similarities between the current iteration of each team and its past versions. Then each remaining game is simulated. The reason every game isn't a 50/50 coin flip is because each team has different strengths and weaknesses, and the way they match up against each other has different historical references for "what happened most often." Because there are many different ways the situational aspect of football can play out, it's necessary to run many simulations for each remaining game, to see each of the involved teams' profiles stack up over a range of reasonable situations.
 
Just how many times is each game simulated? While I might ordinarily go with a figure like 10,000, I decided to up the number to 20,000, including each remaining game in the regular season (there are 135 left).
 
A quick note: Scoring is obviously up this year compared to past years (we're on track to set a new offensive scoring record). But that change hasn't yet -- and likely won't -- significantly affect the use of historical data, as the most influential production factors (things like turnovers and red-zone efficiency), taken in context of the game situation (think: down, distance, score and time), remain consistent with what we've seen before, in terms of what relates most strongly to the ability to win games.

 

2v3k7eb.jpg

 

 

Edited by 26CornerBlitz
Link to comment
Share on other sites

  • 26CornerBlitz changed the title to Cynthia Frelund: Projected Final Win Totals for 2018
Just now, SoCal Deek said:

Sweet!  Last time we picked at Number 3 we landed Dareus.  Flash forward and folks are once again dreaming of yet another D Lineman. Round and round it goes.

Yup. See what happens when ego gets in the way of building a roster.  The wheels just keep spinning. 

  • Like (+1) 1
Link to comment
Share on other sites

4 minutes ago, SoCal Deek said:

Sweet!  Last time we picked at Number 3 we landed Dareus.  Flash forward and folks are once again dreaming of yet another D Lineman. Round and round it goes.

The more things change the more they stay the same. 

Link to comment
Share on other sites

×
×
  • Create New...