Monthly Archives: January 2014

Quarterback Evaluation

The single most important position, in possibly all team sports, is the quarterback. Since the early 70s, quarterback performance has been measured by a calculation called Passer Rating. The Passer Rating was commissioned by the NFL in 1971 to standardize performance and have the ability to compare quarterbacks over a single game, season, or career. A team of statisticians used all data from 1960 to 1970 to create this formula, which was approved by the NFL in 1973.

The formula takes five passing statistics into four parts. These are attempts, completions, passing yards, passing touchdowns, and interceptions. The four components are as follows:

c = (comp/att – 0.3) x 5

y = (yards/att – 3) x 0.25

t = (td/att) x 20

i = 2.375 – (int/att x 25)

Finally, bring it all together and the Passer Rating = ((c + y + t + i)/6) x 100

I have created an excel template to quickly calculate passer rating based on the five statistics. You can download it here.

This formula has two constraints. The four variables cannot be less than zero or greater than 2.375. So if any of the calculations yield a result outside those constraints, then the minimum of zero or maximum of 2.375 is used. A perfect passer rating is then 158.3. The creators of the Passer Rating scaled all 1960 – 1970 data between 0 and 2.375 with 1 being statistically average.

For decades, this was accepted as the measurement of quarterback performance. Is it perfect? No. It does not factor in fumbles, rushing, or sacks but, after all, it is called a “passer” rating. In 2011 ESPN released their own calculation called Total Quarterback Rating, QBR. For a detailed explanation of the QBR, please click here. The CliffsNotes version tells us that the QBR is a weighted calculation that incorporates game context and how those plays transfer to wins. The rating ranges from 0 to 100. Sounds great, show me the algorithm! Ah…that’s where ESPN says hold on, this is proprietary. The QBR formula is not public.

So let’s see how these two rating systems stack up against each other. I pulled “qualified” regular season data only for each rating from ESPN.com for 2008 to 2013. You can explore the dashboard below. Note, ESPN states qualified data for passer rating means a player must have at least 14 attempts per team’s games played. There was no qualified definition for QBR data. 

Select the statistic. This applies to Player Ranking table, the overall average, and the Distribution box and whisker plot. Year filter allows to view the Player Ranking for each single season.


Tableau 8.1 now has a one-click Box and Whisker Plot in the Show Me feature. This is a great function to analyze distributions. It is simple to read. In this case, QBs in the grey box are considered average, with the median being the value where the light grey and dark gray areas meet. QBs that fall above box and below the upper whisker are above average. Finally, QBs outside the whiskers are either exceptional or terrible, depending on the upper or lower whisker.

Looking at 2013, Nick Foles had the highest passer rating but Josh McCown had the highest QBR. But I think most will agree that Peyton Manning was the best quarterback this year. He’s number 2 in both statistics. And if you’re building a fantasy football team, you’d probably take 10-15 other quarterbacks before drafting Foles or McCown. My Washington Redskins had a dismal 2013 season. But 2012 was a great year. Robert Griffin III was 3rd in Passer Rating but 5th in QBR. He had an amazing regular season by any standards, incredible for a rookie. Let’s not talk about what happened in the post season. In general, a QB does not vary greatly between the two metrics. Upper tier players will have upper tier Passer Ratings and QBR scores and so on for mid and lower tier.

Since ESPN launched the QBR, they only use this metric in their broadcast. ESPN has it’s upsides and downsides. I find them only reporting a non-public metric they created on television to fall under the downside. Obliviously they do have Passer Rating on their website, where I got the data. Call me a traditionalist, but I prefer Passer Rating, mostly because I can calculated myself. How about you?

Share Button

One week in, what I have learned

It’s been a week since I launched this blog and already I have a few notes to pass on to future potential bloggers.

First, I’ll admit that I jumped in without researching blogging platforms at all. My wife has a WordPress blog so I figured I’ll just use that. WordPress.com is fantastic tool to get one up and running in minutes. The .com version has all the basics at a click of the button and it provides tutorial videos. I bought my custom URL and started drafting my first post. It was not until my second post that I realized that .com was not the platform I would need. It is a stripped down version of WordPress.org and most importantly, I could not embed Tableau Public vizzes.

I went to my Google account and created a test Blogger.com site. That platform did allow for embedding Tableau Public vizzes from the generated code. So the decision was Google or self-hosting? I’ll point out that I’m a big Google fan. I’ve had an Android phone since the original Droid launched in 2009 and am currently rocking a Nexus 4. But sometimes Google can be disappointing. They shut down popular products like Google Reader and not so popular products like Google Wave. I didn’t want to start a blogger site then have to move it later down the road.

WordPress.org it was! Now came my difficulties. I would need to convert my .com to .org. WordPress will do all the conversion for a fee of $129. Ouch! I consider my self pretty tech savvy. I’ve picked up decent SQL skills in my professional life and I love hacking my Nexus 4 (I root my phones and I’m currently running CyanogenMod rom and the Franco kernel). I decided I’d do it myself. I started reading the conversion steps about exporting your blog and importing and bought a Bluehost account. I did not know that I could not transfer my custom URL within 6 months of purchase. I would need to point WordPress nameservers to Bluehost’s nameservers. Honestly, I know very little about web hosting, so some of this was a little confusing. After I changed the nameservers, I crossed my fingers and it worked….took a day or two to actually be processed. I also saw that I lost my stats and subscribers from my first post. I mentioned this to my wife who quickly reminded me that she moved to self-hosting after a year and lost a lot more than I ever accumulated in two days. I got over it.

In creating the Bluehost account, I used a one-click install of WordPress.org but inadvertently created another WordPress account. Now I have separate accounts for .com and .org. I still need the .com because it has my custom url and it does not look like you can merge the two. All the features of .com that got me up and running so quickly are not present from the start in the .org platform. Everything needs to be installed as separate plugins. I installed a few plugins and I still could not embed a viz. I finally found an iFrame plugin that worked with Tableau Public content. It is still not perfect if viewing my vizzes from a mobile device, that’s a work in progress.

My advice to anyone considering starting a blog is do your research. There are a plethora of blogging platforms out there. Find the one that best fits your expected needs, then create your necessary account(s).

 

Share Button

Puppies…need I say more?

Ten years ago, Animal Planet began running Super Bowl counter programming called Puppy Bowl. A tradition was born. Discovery Communications is headquartered in Silver Spring, MD. We have a large building in the heart of Silver Spring and a few blocks away is our DCTC (Discovery Creative & Technology Center) building. As a Discovery employee (Animal Planet is one of our networks), I work in the DCTC building where Puppy Bowl was filmed for the first few years. It was always a glorious week in October where a plethora of puppies would invade our office and be shuttled in and out of our stage. Due to Puppy Bowl’s popularity, along with the likelihood of the high number of distracted staff, PB’s filming was moved offsite, then eventually to New York.

As the 10th Puppy Bowl approaches, I decided to create a dashboard of twitter data for tweets containing “Puppy Bowl” or #puppybowl. I utilized Andy Cotgreave’s  twitter template he posted on Tableau Public but gave it a few tweaks. For the data, again as shown by Andy, I used ScraperWiki to create a data set for my desired search terms. You can set it to track future tweets in order to keep tracking your results.

Deadspin’s tweet announcing this year’s ability to draft a fantasy puppy team has been the most popular so far. Let’s hope Yahoo Fantasy adds the Puppy Bowl to their platform!
I’ll periodically update the data until the actual event…or until the data limits are reached.
Share Button

Welcome to my blog…hope you stick around

Hello peoples of the information superhighway. I’m excited to announce that I have decided to create a blog about data. For the past few years, my involvement in the world of data visualization and analysis has grown exponentially…and I’ve loved every minute of it. So why did I decide to start a data blog? Well, there are a few reasons:

  1. I have discovered a wonderful community of data enthusiasts and wanted to become a part of it. I have learned a great deal from blogs such as DataRemixed, Tableau Love, Health Intelligence and VizNinja just to name a few.
  2. Tableau – A magnificent self-service Business Intelligence tool that has transformed my career. My own data visualizations in this blog (vizzes for us in the biz) will utilize Tableau Public. A free version of the software designated for public data and sharing.
  3. At the encouragement of my lovely wife Hilary and peers at Discovery Communications.
  4. Finally, because I recently found out that January is Data Blogging Month.

I can’t wait to get started. I hope you’ll enjoy what I have to offer and come back to interact with me and my vizzes.

Thanks for stopping by and have a nice data day. (see what I did there)

Share Button