Depending on how you arrange the digits of a date, today is either 20-10-2023 or 10-20-2023. Statistically speaking, it’s 10-20-2023 for about 88.5% and 10-20-2023 for the 11.5% of the world that puts the month in the first position.
In both positions, though, October 20 sometimes represents World Statistics Day in about 49.5% of the countries that celebrate it. About 1%% celebrate Statistics Day on the 29th of June, which is also the birthday of Prasanta Chandra Mahalanobis, and the remaining 49.5% celebrate on the 18th of November.
World Statistics Day is not an annual event; it’s only celebrated every five years. The Royal Statistical Society made an even bigger splash more than a decade ago when they launched their statistical literacy campaign on 20-10-2010. They kicked it off at exactly 20:10, of course.
There’s also an International Association for Official Statistics (of course there is), and they promote a “Global Statistical System” that has ten Fundamental Principles. But there’s very little mathematical — or, for that matter, statistical — about them; they’re just statements like “Official statistics provide an indispensable element in the information system of a democratic society…” and so on. Relatively boring stuff that 63% of the members of the Association weren’t even able to recite from memory on the last World Statistics Day.
Prasanta Chandra Mahalanobis, on the other hand, was interested in the statistical sort of statistics. Depending on how long you continued to take statistics courses (84.2% of all undergraduates in the US progress no further than Statistics 201), you may remember the Mahalanobis distance, which he invented.
If you know about standard deviations — the more unusual something is, the more standard deviations it is away from the mean — the Mahalanobis distance is pretty similar. It’s a way to measure how anomalous an anomaly is when you don’t have everything measured the same way.
If you see a cat that weighs about 8 pounds and has black and gray fur arranged in stripes, you can use the Mahalanobis distance to find out how unique a cat with that combination of characteristics might be. It works even though weight is a quantifiable continuous variable, fur color is a set of discrete values, and stripes versus spots is a non-quantifiable subjective evaluation. This is something only 12% of cat owners generally care about, though, which is why if you ended your statistics studies after Stat 201, you probably didn’t encounter the Mahalanobis distance.
When the Convention of 1818 was signed on October 20, 1818, establishing the border between Canada and the US, the surveyors didn’t encounter the Mahalanobis distance either, even though the treaty explicitly addressed locations and distances. This is because the probability that they were aware of it asymptotically approaches zero, since Mahalanobis didn’t come up with the calculation until 1936. As an aside, the International Association for Official Statistics has come out against the misuse of statistics a number of times, so they would definitely frown on anyone pointing out that since the year 1818 is 93.9% of the year 1936, anything that happened in 1936 ought to have a pretty good chance of happening in 1818 as well. That’s why the probability that the signers (or even the surveyors) being aware of the Mahalanobis distance only APPROACHES zero, without actually BEING zero.
The next paragraph should, obviously, present the formula for calculating the Mahalanobis distance along with at least one practical example. But instead of that, we’ll just ponder the fact that on October 20 in 1944, General MacArthur came ashore in the Philippines during the Battle of Leyte, the Soviet Army liberated Belgrade, AND storage tanks in Cleveland leaked so much liquified natural gas that the explosion leveled thirty city blocks. All on the same day. And what are the chances of that? Probably about the same as having 1.49% of the words in this article being “Mahalanobis” — but that just actually happened too.
