The Average Story of the Average

We all want to be normal. The irony, though, is nobody wants to be average. Yet, this very benchmark, the average, is our constant companion from the moment we realise anything in life to the moment we stop realising everything. We compare ourselves and others against an average. Others do the same. The story of the average doesn't end here. From the moment we enter school, we're told that every one of us is a genius. We're told that each individual is different from others. If we work hard, all of us can be an Einstein or any other genius. Alas! Einstein was an exception to an average. He was what we call an 'outlier'. If all of us become an outlier like Einstein, we would all become the same. We would all converge to the benchmark, the average. The reality is each one of us is different. 

An average is the centre towards where data in a set converges. The average of 1, 2, 3, 4, 5 is 3. Once the average has been calculated, it can act as a benchmark through which we compare other numbers in the set asking a simple question: how far away from or how near to the average is the data we're observing? Different data sets will lead to different results on how well the average represents the data set. The problem or the challenge arises when the data set are performance or features of people being represented as numbers. When this happens, the average can become misleading. Think about the following two examples:

First: a doctor advises a drug to his patient with the cautionary warning – results suggest five percent of patients have fatal side effects. Ninety-five percent of the patients don't experience such side effects. Thus, the drug administration has classified the drug to be 'safe'. 

Second: The two greatest Sri Lankan players – Muttiah Muralitharan and Kumar Sangakkara – played their entire Test career without tasting a Test victory against Australia. Prior to the 2016 tour to Sri Lanka, the first and last Sri Lankan Test victory against Australia was in 1999, before Murali and Sanga made their Test debuts. Based on average calculations, there was no way one could have thought Australia would experience a Lankan whitewash of 3-0 in 2016. But it happened. 

In the first example, the benchmark of ninety-five percent success rate is based on a historical data calculation that does not include the patient. It would be useful if the patient could know which side of the ninety-five percent they fall in. The patient's survival depends on this information. In the second example, the performance of Australia and Sri Lanka in Tests is also based on historical data between teams that no longer exist and on past forms that are no longer relevant. Rangana Herath's magic was not calculated in the average performance between the two teams before the 2016 tour.

Do we give up the average? No. An average is certainly useful. It gives information on what we expect to see. However, it's not a good guide to predict individual data regarding individual people. The Sherlock Holmes creator, Arthur Conan Doyle, had a disastrous school record. He would barely pass in all subjects except for chemistry. His average performance was well below the average of the class or what the school expected of him. However, he was talented. History suggests it was all 'elementary, Dear Watson'.

Fortunately, a new science of the individual has emerged that argues that talent is not a straight line and CGPA's and other average calculations don't reveal many aspects of a person's talent. Google, Microsoft and other leading corporations are slowly adopting results from this new science. Echoes will 'echoe' on this new science soon.

Asrar Chowdhury teaches economic theory and game theory in the classroom. Outside he listens to music and BBC Radio; follows Test Cricket; and plays the flute. He can be reached at: asrar.chowdhury@facebook.com