top of page
Search
Writer's pictureJason L

Introducing The Lawrence Index: How Does Each Game Impact the Race for the NFL Draft

Throughout this NFL season I have been giving a probability forecast on the race for Trevor Lawrence through my simulation model, Trevor-thon 5000. The model has been tracking the race from a macro-level overview of all remaining games throughout the rest of the season. As I mentioned a few days ago when I did this week's update, I wanted to look into more of the micro-level view by estimating the leverage each game has individually on the race based on whether the competing team wins or loses.


Introducing The Lawrence Index...

Each line shows three numbers for each team that is still mathematically in the running for the first overall pick:

  • The estimated average first pick probability assuming they win this week (blue)

  • The current first pick probability after last week's games (black)

  • The estimated average first pick probability assuming they lose this week (red)

For instance, my beloved Jets (for whom I created this model for so that I could justify cheering wildly every time they lost) enter this week against the New England Patriots at 53.3% on Trevor-thon, would move to 62% on average when they lose, and fall back to 33% on average when they win. So the total leverage for the game is about 29%, i.e. the swing between losing and winning this week. The leverage is much bigger for teams at the top than it is for teams at the bottom because they have more share to lose.


Note that the "on average" qualifier is important because these are just aggregates of all the possible outcomes of this week's schedule. With 14 games this week, there are 16,384 (2^14) unique possible outcomes to the entire schedule this week. For instance, this week could have the Jets, Jaguars, and Giants all lose; or just have the Jets win; or have both the Jaguars and Giants win; etc. The model attempts to weight all of these possibilities, but next week's probabilities would likely shift some when the actual results are known. The Jets would be a lot more than 62% for first pick if we knew the Jaguars and Giants had both won, but the model still has take into account the uncertainty of the results of unplayed games, hence the "on average" qualifier.


What I like about The Lawrence Index is the ability to read into what if scenarios. If the Jets won and everybody else lost, then we estimate the Jets (1-8) dropping to 33%, Jaguars (1-7) at 31%, the Giants (1-8) at 17%, and everyone else moving up 1-2 points (the rising tide lifts all boats). I think on its face that sounds about right because the Jets would lose 20% from where they started the week (53%) with 13% going to the other 1-win teams. The Jets would still maintain that 2% lead because of the loss in hand over the Jaguars but would otherwise be about equal. Meanwhile the Giants would still lag behind because of their easier strength of schedule (and quite frankly they just haven't played as badly as the former two, even with the same record).


Just a nice gauge into what is at stake in each game this weekend. I will post The Lawrence Index with the overall Trevor-thon probabilities for the rest of the season, starting next week.

0 comments

Commentaires


bottom of page