Chili Topped With Statistics

Data tools, Pareto analysis help win a cook-off

by Paul Stepan

My daughter calls me a data nerd, as I am always using Microsoft Excel to document my life. I maintain data on the time my wife’s van arrives in the carpool lot, data on the performance of my investments and data on my various collections. There is a certain sense of satisfaction of knowing within two standard deviations what time I need to be at the van pool lot to pick up my wife.

I recently used data and Excel in a new way—to increase my chances of winning a chili cook-off contest.

Earlier this year, my company in Houston held its annual chili cook-off competition. I love to cook, and I enter the contest every year, typically using a different recipe. I am constantly searching for that perfect combination of ingredients that would get me the coveted trophy.

Unfortunately, my chilies have, thus far, failed to impress the judges enough to garner sufficient points—if any at all—to win.

The early chili cook-off years were spent making northern-style chili, which typically contains beans and tomatoes and is light on the spices. Those years produced only one third-place trophy. At some point, I learned that true Texas chili does not have beans and should have sufficient spices to peel paint.

Back to the drawing board.

The next idea was to combine the ingredients from two successful recipes to see if that would work. Again, no such luck.

A pinch of statistics

This year, I decided to use a different approach. Statistics would be my savior.

Knowing my desire to win a chili cook-off, a co-worker recently gave me a cookbook as a birthday present. Inside the book were nine chili recipes, each of which had won a prestigious chili cook-off competition in Texas.

Certainly, I thought, I could come up with a recipe using these nine samples. Because the nine recipes had each produced championship chilies, they must have something in common.

Using an Excel spreadsheet, I first entered the ingredients used in each of the recipes along with their weights or volumes. Next, I analyzed which ingredients were most commonly used, and I calculated the average weight or volume of the ingredient. Not wanting to use all the ingredients, I then used a Pareto analysis to select only the top ingredients for my recipe.

Armed with my statistical chili recipe, I cooked up a pot. When it was complete, I tasted the results. I had little confidence my pot of chili would win because the taste was so different to me. I was determined to test the hypothesis, however, so I entered my chili into my company’s contest. After the steam had cleared, the judges scored my chili first place with 11 out of a possible 15 points.

Second opinion

Because one data point is not a trend, I had to validate my results in another chili cook-off. As the winner of my company contest, I was the company representative at an Institute of Electrical and Electronics Engineers student group chili cook-off at the University of Houston.

To my delight, my chili tied for second out of a field of 14.

As the manager in charge of Toshiba’s continual improvement program, I made a presentation about my journey to demonstrate that statistics can be useful outside the workplace. As a result, I have been recognized by those who liked my chili, as well as by those who saw the method as a creative use of statistics.

Paul Stepan is the quality assurance manager at Toshiba International Corp. in Houston. He earned a bachelor’s degree in manufacturing technology from the University of Houston. Stepan is a senior member of ASQ.

Average Rating


Out of 0 Ratings
Rate this article

Add Comments

View comments
Comments FAQ

Featured advertisers