The genius we called the “Old Man” said that “what are not accounts, are stories.” And so it inhibited us from issuing opinions that we could not defend with numbers. Thanks to that, every time I come across an interesting database, my learned reflex is to apply some statistical techniques to it to try to discover “truths”. I have no particular interest in basketball, but it is one of the sports with the most information available. What I was really looking for yesterday was to understand a problem related to the supply and demand of programmers and software architects and I came up with a theory that I wanted to test by analogy. With the data I found on nba.com and some other minor sources, I tried to answer the question: What predicts a person will play in the NBA? And what common characteristics do the stars of that league have?
After a few hours of (here is the part where I think if it is better to pretend that I am intelligent or to pretend that I am didactic, and if I have to write “machine learning, binary classification models and logistic regression” or better “do some little stories”) I came to the following conclusions:
- To play in the NBA, the best thing is… ..to be tall. Before congratulating me on my sagacity, let me put numbers on the claim. The average height of an NBA player, throughout the history of the league, is 2,015 meters. Considering age (there were never players under 18 or over 45) or nationality (almost all American and there was never a player from India,) height is the best predictor of whether a person will play in the NBA. And it is decisively so. Considering then an American man, between 18 and 45 years old: If he is 1.75 meters tall, the probability of playing in the NBA is less than 1 in 120 million. If you have the average height of the NBA (2.01 meters) the probability jumps to 0.7% (that is, a million times higher). If, instead, he takes to the bottom of the chocolate and ends up measuring 2.19 meters, the probability flies to 35%. Finally, if you measure more than 2.60 meters, you play in the NBA, “statistically” guaranteed.
But the second conclusion of the analysis is the interesting one. 2) Once you enter the NBA, height and success are no longer correlated. In fact, the best fifty NBA players of all time have an average height of… 2,014 meters, one millimeter below average, with Michael Jordan and his 1.98 meters as an example. This situation, where “the formula that assured you to reach a goal is useless to keep moving forward” occurs, identical, in the world of technology. If you ask any person in charge of a knowledge company, what is their greatest current concern, the most frequent answer will be “there are not enough people”. Notes are published every day that highlight the asymmetry between supply and demand for programmers, and the vacancies that are increasingly generated. However, surveys show that candidates say that, although getting a job is easy, growing in companies is very difficult. In a strained analogy, they are like 2.20 meter players, frustrated at not making it to the hall of fame, despite how easy it was to get into the league. I promise you that if you join me next Tuesday, I will tell you why this is totally true (and logical), why it will be more and more pronounced, and what simple things a young technology professional has to do to reach the top in his career.
