Did you know that Friday, August 7 is International Beer Day?
This global celebration pays tribute to one of the world’s favorite beverages –– beer. It encourages beer lovers from all walks of life to sit back, relax, and raise a toast.
This year’s celebrations will undoubtedly look different than usual. Even so, International Beer Day presents an opportunity to celebrate by enjoying a pint of your favorite brew.
In honor of the holiday, we thought it would be fun to take a closer look at statistical distribution and the role it plays in the beer brewing process. Statistical analysis is a major part of the biotech and pharmaceutical industries, yet one of the most important continuous probability distributions –– Student’s T Distribution –– came about, thanks to the Head Experimental Brewer at Guinness.
Who is William Sealy Gossett?
William Sealy Gossett is a pioneer in the field of statistics, who spent his entire 38-year career at the Guinness brewery in Dublin, Ireland. He doesn’t have the same name recognition as other famous statisticians of the time, like Karl Pearson and Florence Nightingale, but his work still influences beer production today.
On October 1, 1899, William Sealy Gossett joined Guinness as a junior brewer. It may be hard to believe, but at the time, there was no sure-fire way to produce consistent, great tasting beer in large quantities.
Between 1877 and 1914, production doubled at Guinness, but a problem remained. How could the brewer increase output and reduce costs, while maintaining Guinnesses’ signature dry stout flavor?
Fortunately, Guinness had a nearly limitless supply of money and resources. As the world’s largest brewer at the time, Guinness regularly sought out and hired scientists who could assist them in staying one step ahead of the competition.
A challenge arises
Upon arriving at Guinness, William Sealy Gossett joined the scientific brewing team. His first assignment was to find an efficient and economical way to select the best, highest-quality hops. Hops are flowers of the Humulus lupulus plant and they play a crucial role in flavoring beer.
Guinness' chief scientific brewer, Thomas B. Case, developed an equation to assist in these efforts. It calculated the proportion of “soft” resins to “hard” resins in a batch of hops. The equation produced results, but Case had no way of interpreting their meaning.
Case decided to share his findings with Gossett, who had a reputation for taking on challenges with gusto; he had also studied mathematics at Oxford. Case challenged Gossett to develop a theory that would allow for accurate deductions from a small sample size. With a way to select high-quality hops consistently, the company could continue to grow.
Testing begins
Gossett’s first goal was to determine the error distribution of small samples. During the late 1800s, beer brewers would take dozens of samples and then use the standard normal distribution to gain insights.
Gossett figured there was probably a way to gain similar information, but with far fewer samples, say five or ten. He started by making observations of malt extract, with a goal of landing on a number that would accurately define the “degrees saccharin” within 0.5 degrees of a target of 133 degrees.
After conducting dozens of tests and collecting significant amounts of data, Gossett was able to confidently determine “degrees saccharin” within 0.5 degrees from two-observation samples 80% of the time. Adding observation samples proved even more fruitful. Three observations resulted in an 87.5% success rate and four observations, a success rate of more than 92%.
Studying with Karl Pearson
These findings thrilled Case and the rest of Guinness management, but Gossett remained unfulfilled. Unless he could understand the exact mathematics behind the inference of small samples, questions would still remain.
After meeting with Guinness brass, Gossett received permission to study under Karl Pearson at the University College of London. There, Gossett spent a year under Pearson’s tutelage and refined his equation for working with small samples. Ultimately, this work resulted in one of statistics’ greatest discoveries: “Student’s t-distribution.”
What is Student’s t-distribution?
Student’s t-distribution allows inference through small samples with an unknown variance. Compared to the normal distribution hypothesis, Student’s t-distribution provides more accurate results.
Even though William Sealy Gossett developed this statistical distribution, it doesn’t hold his namesake. This is because Guinness carefully guarded its brewing secrets. When Gossett decided to share his findings with the general public, Guinness management required him to use a pen name. This prevented competing brewers from knowing about the brewer’s extensive research apparatus.
Cheers, William Sealy Gossett!
William Sealy Gossett changed the world of mathematics forever. Today, Student’s t-distribution is one of the first concepts taught in statistics courses. Additionally, Gossett’s discovery made it possible for brewers to maintain iconic flavors, all while expanding their production efforts.
As you celebrate International Beer Day, raise a glass to the great William Sealy Gossett. Thanks to his hard work and dedication, your favorite beer will always taste the same, whether it’s an IPA, stout, or lager.
https://priceonomics.com/the-guinness-brewer-who-revolutionized-statistics/
https://www.york.ac.uk/depts/maths/histstat/student.pdf
http://faculty.fiu.edu/~blissl/GuinessGossetFisher.pdf
https://simple.wikipedia.org/wiki/Student%27s_t-distribution