Ed Feng, PGSS 1994

Q: Why did you decide on chemical engineering as an undergraduate major (and subsequent doctoral program)? 

A: During my senior year of high school, I think I was visiting Case Western, and I was talking to one of the professors about chemical engineering.  I asked him, “Is this the hardest major?”  And he said, “Yes.”  And that’s when I decided that this was what I was going to study, because that’s what an arrogant 18 year-old thinks and does.

It ended up being a great background for me because I studied a lot of practical engineering, but also got experience in basic science and a good mix of chemistry and math.  I pretty quickly realized that the engineering and operations side wasn’t really my thing, but the major was so broad that it led me to what I was truly interested in, which was applied math, statistics, and more of the theoretical side of chemistry and physics.

My doctoral program was definitely theoretical.  I worked under a chemist on problems in statistical mechanics, which is the science of using applied math to understand how bulk properties of matter are connected to the microscopic properties of atoms.  We worked mostly on problems involving the movement of polymers in liquids.

Q: What is your involvement today with chemistry?

A: Well… none [laughs].  But I feel like it’s going to come back into my life when my kids reach high school and they take chemistry.  One of the things that irked me about my own education is that I wasn’t given a reason to want to understand things.  I feel like no one told me why I should care about thermodynamics or calculus or anything else, really–we were just told to learn it.  

But I envision a future in which my kids understand chemistry via food.  You can’t get away from food.  Frothing a latte is a problem of lipids.  How long it takes to cook a turkey is a physics problem–heat laws, diffusion, things like that.  Now that I have kids, I’m passionate about rethinking education and making it more interesting than it was for me back in the day.

Nonetheless, I’m still in touch with my dissertation adviser, and I’ve never known anyone better than him at understanding difficult problems on the fly.  He was a really fantastic influence.


Q: Professionally, you made a big leap from statistical mechanics to sports analytics.  When did you realize that this was actually a career option? 

A: I started in grad school writing some code for a simple model to fill out my bracket [in the annual NCAA Men’s Basketball Tournament.]  Eventually, people stopped inviting me to their pools.  I was always interested in sports analytics, but I had this hard-headed view that I needed to make some unique contribution.  I think that came with my academic mindset.

Around 2008, when I started coming up with my own algorithms, I thought that I might take an academic or journalistic approach and just write interesting, educational articles–kinda like what [predictive analytics blog] 538 does.  Then I realized pretty quickly that sports gamblers were interested in my work, and I made the slow transition from purely academic pursuits to producing content that gamblers wanted.  By 2012, my algorithms were sophisticated enough, and I had seen sufficient year-over-year growth in my services, that I realized I could make a full-time business out of it.  I also secured an investment from some friends that allowed me to quit my job six months earlier than I had planned.

None of this was obvious at all when I started–it was just a hobby.


Q: Which skills and tools do you use on your job? 

A: Python.  All of my coding is in Python.  Since I developed my own algorithms, I had to write all of the code from scratch–there were no open source packages available to me.

I have a bunch of scripts that collect the data each morning.  Say, for the NBA–they grab all of the previous night’s scores, use them to calculate rankings, and push everything to my site, thepowerrank.com.  For building the models, all of the computing power I need is on my laptop.


The Power Rank uses data and analytics to make accurate football and March Madness predictions. The methods started with a Stanford Ph.D. in applied math and have developed over 8 years.


Q: Do you have any advice for aspiring sports analysts?

A: Yes: you have to be social.  You can do the most brilliant things on your own, but if no one knows you, you’re not going to make the impact that you want to make.  Like any field, you need to go to meetings with the practitioners, learn what they’re interested in–figure out the problems that need solving, and go solve them!  If you specifically want to get into sports analytics, you go to the annual Sloan Sports Analytics conference at MIT.

The same thing applies in any academic field.  If you’re in grad school, you talk to the wise elders, and hopefully they’ll give you an interesting problem to work on.  Again, it starts with being social.