Predicting QB Success in the NFL

Last year I wrote and submitted a paper for the MIT Sloan Sports Analytics Conference. While my abstract was accepted my paper was not. The title of my paper was Reducing Risk in the NFL Draft: Using Machine Learning Algorithms to Predict Success in the NFL. You can read the full paper here. 

In it I describe a decision tree model that predicts a college QBs success in the NFL. To train the model I used over 40 variables including college stats, school competitiveness, combine performance, and text mining of pro scouting reports. Ultimately, the final model used 4 variables: college win %, body mass index (BMI), college games started per season, and age. The final model was 88% accurate in predicting whether a college player would be a success or a bust in the NFL. This model can be used to predict whether the top prospects in this year's draft will be successful in the NFL.

Below is an interactive version of that final QB model.

Read More

NFL Combine & Triangle/Ternary Plot

Triangle/Ternary plots are a good way of displaying the relative positioning of points across three variables. This can be used to cluster and classify points based on these three variables. After the break I have outlined a way of creating an interactive version of a triangle/ternary plot in Tableau.

Here is an example of a triangle/ternary plot on the performance of collegiate players at the 2017 NFL combine across three variables: size (BMI), speed (40 time) and strength (bench press).

hot to make it using R & Tableau
To create this using R and Tableau is relatively easy. First, you need a file with just three variables scaled appropriately. For example, in the above chart, a high 40 time is not good or representative of high speed. So I normalized all of the variables using min-max normalization:

You can do this is in R, Alteryx, Excel or whatever. Then inverted the value for 40 times:

So I have a table of 3 normalized variables (40 time, BMI, and bench press) by player. I also add 3 anchor points to determine the corners of my triangle plot. So I have three extra rows where the value is 0 for each row except 1 for each variable.

I then save this as a csv file. I import these variables into R in order to create the triangle/ternary plot. Here is the R script below:

install.packages("ade4")
library(ade4)
tri_data <- read.csv('C:\\Users\\name\\Documents\\2017_Combine4.csv')
x <-triangle.plot(tri_data, label = as.character(1:nrow(ta)))
write.csv(x, 'C:\\Users\\name\\Documents\\2017_Combine_Tri_w_anchors.csv')

This will create X and Y coordinates associated with each point (player). I append this to my player data set and bring this into Tableau. To create the triangle plot in Tableau I place the X coordinates in the column shelf and the Y coordinates in the row shelf and convert the measures to continuous dimensions.

Finally, I add a triangle as a background image, fix the X and Y axes, and then remove anchor points (in the corners). I also do some formatting removing the gridlines, adding color, and increasing the transparency.

hot to make using it just Tableau

The talented Mike Cisneros (@mikevizneros) recommended an approach that skipped the R step and allows you to create this entirely in Tableau. Here is link to Mke's version on Tableau Public. It only requires two simple calcs and is much more clever than mine.

Essentially, Mike imported the same data, normalized in Tableau, and then created the X and Y coordinates in Tableau rather than R. In the below Benchn, Bmin and 40n are the normalized measure values for Strength/Bench, Size/BMI, and Speed/40 Time respectively. Here are Mike's calcs in Tableau:

See details on these calculations here on Wikipedia. You again place the X and Y values into the column and row shelves respectively and repeat the other steps above.

You can download my Tableau Public file above. It has both versions (R and Tableau) included. I hope this is helpful. If you have questions please leave them in the comments box below. Thank you.

Read More