# SAS loves Stats: Don Wedding

Don Wedding, SAS

Don Wedding played a baseball simulation game called Sports Illustrated/Avalon Hill Superstar Baseball back when he was in grade school in Toledo, Ohio. The game involved rolling specialized dice, and then referring to cards representing the performance of the greatest baseball players of all time.

The problem was Wedding knew absolutely nothing about baseball, while all his friends were experts.

“I got pounded every season,” said Wedding, thinking back to those days in the late 1970s.

Then he had an idea: to study dice rolls and the statistics on baseball cards to ascertain how well a player would perform. “I used statistics to find who the undervalued players were, and it was my version of Moneyball before Moneyball became famous.”

Everyone was amazed that he went from worst to first. “I loved it so much I had to tell everyone what I was doing,” he recalled.

So, all the baseball experts in the league started using his techniques. Then, coupled with their baseball knowledge, they regained their dominance and “I proceeded to drop into the cellar again.”

He quickly learned that “statistics beats expertise. But statistics and expertise beats statistics.” He also learned that he shouldn’t have told his friends about his statistical formulas.

When it came time for college, he majored in electrical engineering at the University of Toledo. “Two years in, I hated it, so I started taking electives in computer science and statistics,” Wedding said. “I fell in love with statistics all over again!” He loved the classes and often would work problems in the text books just for fun.

After getting a master’s degree in engineering, he worked as a software engineer before earning his PhD in engineering, with many elective courses in statistics. While he was in graduate school, a professor introduced him to machine learning techniques such as neural networks and clustering. “In reality I was working in the field of data mining even before it was called that.”

He describes himself as having a wide breadth of knowledge but not the same in-depth knowledge as someone with a degree in statistics. “I am a mile wide, but the pure statisticians are a mile deep.”

After earning his PhD in 1995, Wedding worked building statistical models in several small startup companies before moving into the banking and insurance industries, where he used SAS to solve numerous types of problems. These ranged from risk to fraud to customer retention to text mining.

During that time, he decided to take an online course in data mining from Central Connecticut State University just for fun. He enjoyed it so much, he took a second, and a third, and so forth. He never had any intention of pursuing another degree. But eventually he did graduate from their program when “just one more class” turned into “there aren’t any classes left to take.”

In 2006, he got a call from his future manager at SAS. “He had been given my name as a person who might be good at pre-sales,” Wedding said. “He heard that I had the domain experience in financial services, a statistical background, and good customer-facing skills.”

Now a Principal Industry Consultant for the Financial Services industry, Wedding works out of a home office in Cleveland, but travels frequently to customer sites. “If there is an analytic component to a sale, I’m involved 99 percent of the time.

“I’ll demonstrate SAS® Enterprise Miner™ or SAS/STAT® or some of the analytic tools,” Wedding said. “Also, because I am an industry consultant, I will ‘white board’ problems with customers and sometimes use their data to build preliminary models in proofs of concept.

“Companies have lots of data that they accumulate in their day-to-day operation of business,” Wedding explains. “My job is to extract information out of that data using mathematics and machine learning and statistical techniques. I give that information to other people so they can turn it into strategy and action.”

Most engagements usually last about three days, but they can last longer if it turns into a proof of concept. Wedding likens himself to field-goal kicker in football because he has a single, specialized purpose.

“They bring me in for a short period of time when they need someone to talk analytics,” he said.

His job requires him to stay current on modeling techniques and industry trends. “I’m constantly learning something new and trying different analytic approaches,” Wedding said.

When he considers the work he does, the people he works with and the technology he gets to use, Wedding said he has never been happier in a job.

“I love talking to people and I love talking about statistics,” he said. “As the saying goes, when you love your job you never work a day in your life.”

• The field of statistics is a journey not a destination. You will never know everything, but you should always continue learning.
• Never fall in love with a statistical technique. Try many different approaches to solving a problem. You will never know which one will be the best for solving a specific problem.
• Never underestimate the importance of domain expertise in solving a problem. (See interesting story below.)

• If you love pure statistics, keep an open mind about more practical approaches such as neural networks, decision trees or clustering.
• Predictive analytics in a particular field requires domain knowledge. Learn as much as you can about that industry because that knowledge will translate into better models.

DO YOU HAVE AN INTERESTING STORY ABOUT STATISTICS?

• In college, two of my most dreaded classes were Thermodynamics and Energy Conversion. They were both nightmare classes that I thought I would never use. I waited until I was nearing graduation to take them and I basically blew them both off knowing that I wanted to work in statistics. Unfortunately, my first job in statistics dealt with optimizing coal-burning power plants, which required extensive knowledge of, you guessed it, Thermodynamics and Energy Conversion. Moral of the story: You never know what you’re going to need from a domain expertise standpoint, so learn as much as you can!

• John Tukey because he advocated a practical approach to statistics based on Exploratory Data Analysis (EDA), which is the idea that many statistical techniques are robust and can still give good results even when the underlying assumptions are violated.
• James Bezdek and Tuevo Kohonen because of their contributions to cluster analysis.
• Benoit Mandelbrot because of his contributions to fractal geometry.

WHAT DO YOU LIKE TO DO OUTSIDE OF WORK?

• Drink wine and watch movies with my wife.
• Watch Star Trek (the old series of course!), Batman, and other classic television shows with my kids.
• Play tennis and chess.
• Coaching the design, and programming the robot for my children’s Lego League Robotics Team.
• Mess around with SAS Enterprise Miner!

Learn more about other statisticians in the SAS loves stats series, or check out our International Year of Statistics page to read how we're celebrating.