Page 1 of 1

A Review of A Statistical History of Pro Football

Posted: Wed Oct 20, 2021 6:26 pm
by RRMarshall
Here is my long promised review of A Statistical History of Pro Football by the recently departed PFRA member Rupert Patrick. Although I have an advanced degree in Statistics I usually tend to pass on works such as this as I find they can become tedious. However, while Rupert’s work touches on former work done in the field of analyzing football statistics, he lays out a new analytical way of ranking players and teams from different eras that is quite intriguing. He is very good at carefully laying out the steps he takes to derive his new statistical measures, and the end product should interest anyone interested in quantifying football statistics across the eras.

There are 10 chapters in his book, the first being Home Field Advantage. He develops two new statistical measures, HWA or Home Field Winning Advantage, to see which stadiums have the best winning advantage for their teams (hint: it’s not the Cowboys’ AT&T Stadium!).

Chapter 2 examines estimating winning percentage by using points scored and allowed. Rupert begins by using Bill James’ Pythagorean Theorem approach with runs scored and allowed in baseball, but then expounds on that by employing Bayes theorem to get a more robust measure. Lastly he uses standard deviations to calculate Z-scores to compare teams from different eras (the ’62 Packers came in 5th, their 1996 brethren claimed the top spot).

Chapter 3 is dedicated for developing a statistical way to analyze Momentum, and Chapter 4 has an analysis of Cinderella and One-Year Wonder teams. But it is Chapter 5 and Chapter 6 where the book really takes off. Chapter 5 tackles the issue of Rusher Ratings with a means to not only compare the best rushers across the years but also ranking each of the great running backs’ finest seasons. He first addresses the issues of the eras where more games were played during the regular season, creating a normalizing statistic called YRAA or Yards Rushing Above Average to normalize the issue. In addition to Barry Sanders and Jim Brown some surprising names pop up in the top rushing seasons of all time (any Beattie Feathers fans out there?). When examining the career rushing records again the familiar names pop up, but some of the names that appear among the all-time worst rushing careers include a few current members of the Hall of Fame. There are certainly a lot of discussion points on these findings.

Chapter 6 deals with Rupert endeavoring to formulate a better formula to calculate Passer Ratings, or as he refers to it, the PRS (Passer Ratings System). The final result is the NPR, or Normalized Passer Rating. The new measure is very favorable to the old time QBs like Graham, Baugh, and Luckman and their outstanding seasons. Modern era QBs dominate the career passer list, although I was a bit surprised to see my favorite John Unitas not even dent the top 20.

By far my favorite chapter was Chapter 7 where Rupert attempts to create statistics to measure kickers across the eras, a very difficult thing to do considering modern day kickers have such near perfect conversion rates. He begins by introducing K%, the ratio of how many points a kicker scored on his kicks compared to how many points he would have scored if he would have been successful on all his kicks. By computing this figure for the league in any given season a kicker’s performance can be judged by comparing it to the league average.

Rupert addresses the issue of the rising rate of kickers through the years by using PAL, or Points Above League. He defines PAL as the difference in points of what a kicker actually did in a given season compared to what an average kicker would have done given the same number of extra points and field goal attempts. The advantage of PAL is that is can be calculated for any kicker or team in NFL history. However, he does point out that the major drawback is that PAL weighs all field goal attempts and misses equally as if they were all kicked from the same distance. Hence the use of PAL2, which accounts for field goal attempts and misses at each particular distance.

In order to compute this new statistic not only is it necessary to know a kicker’s misses from every distance, the number of attempts and misses for all kickers from every distance for a particular season is needed. Rupert collected all of that data but displays it for just the 1983 NFL season and then uses PAL and PAL2 to rank the best and worst kickers and teams. His database for field goal attempts and misses at each distance by season would be an incredible addition to the PFRA library to encourage further research.

Chapter 8 takes a similar attempt for punters, although he only shows the results of his newly devised ratings for the 1961 AFL season and 2018 season. The end result is PUR (Punt Rating). As it turned out his new statistic validated Brett Kern of Tennessee as the AFC 2018 Pro Bowl kicker, as he finished tops in the ratings.

Chapter 9 explains the concept of Coding, his simplified method of comparing teams. He uses that methodology in Chapter 10 in a study to rank the Dynasties of the game. What’s your pick, 50s Browns, 60s Packers, 70s Steelers, 80s 49ners? (Hint: The Brownies fare quite well)

Rupert close his preface by hoping readers of his book will find the journey through it interesting and it will cause them to think about the game of football in a different way. I consider it a mission accomplished on both points, and I hope my PFRA brethren will take up Rupert’s challenge and use the new tools he has provided us to continue his research. I cannot recommend this book enough. Rest easy Rupert, job well done.

Re: A Review of A Statistical History of Pro Football

Posted: Thu Oct 21, 2021 12:45 am
by SixtiesFan
I am a Beattie Feathers fan. Does he get a mention?

Re: A Review of A Statistical History of Pro Football

Posted: Thu Oct 21, 2021 11:11 am
by RRMarshall
I am a Beattie Feathers fan. Does he get a mention?
Indeed he does!!

Re: A Review of A Statistical History of Pro Football

Posted: Thu Oct 21, 2021 2:39 pm
by RichardBak
RRMarshall wrote:
I am a Beattie Feathers fan. Does he get a mention?
Indeed he does!!
Have you posted your review on Amazon? (Might have to trim it a bit, though.) It'd be good for potential buyers to read an analysis of the book by someone who actually read it.

Re: A Review of A Statistical History of Pro Football

Posted: Thu Oct 21, 2021 4:12 pm
by Ronfitch
RichardBak wrote:
Have you posted your review on Amazon? (Might have to trim it a bit, though.) It'd be good for potential buyers to read an analysis of the book by someone who actually read it.
Yes, please. Love it or hate it, Amazon is the gorilla in the room and people still look there for reviews. The information on Rupert's book there is sparse at the moment - no sample pages nor reviews.

Re: A Review of A Statistical History of Pro Football

Posted: Thu Oct 21, 2021 5:01 pm
by RRMarshall
Thanks guys, I definitely will hit up Amazon with a review. Hope we get some more interest and feedback going here on the board as well.

Re: A Review of A Statistical History of Pro Football

Posted: Sat Jan 22, 2022 11:26 pm
by RyanChristiansen
Great review! Very helpful.