Once a year, I throw a party dedicated to Guy Fawkes Day and the Scientific Method. That is, I invite a horde of friends over, and we perform many experiments. Where each experiment consists of drinking one (1) Irish Carbomb, and giving it a rating from 1 to 10. Each carbomb is mixed from a secretly selected mix of stout, whiskey, and cream. And with multivariate analysis, we see which ingredients got the highest ratings overall, in order to figure out the best recipe for the drink.

In previous years, Emmett's Irish Cream came out as the clear winner. Jamieson 12 Year Special Reserve had a narrow lead over Feckin' Irish Whiskey as the best there. The stouts were more contentious, perhaps because different batches of beer can be so very different, even within a single brand, from year to year.

Last year, the crew proposed a special test for the next time: Just run Guinness against Murphy's. The hypothesis: Murphy's milder flavor would let more of the Irish Cream through, compared to Guinness's tangy notes. So we tested that this year: Emmet's and Jamieson 12 Year for all the carbombs, only one variable: Guinness vs. Murphy's.

And the result? People couldn't really tell them apart.

Note that the "beers change a lot from batch to batch" principle may still apply. It seemed as if we had a not-particularly tangy batch of Guinness this year.