The Bayesian: probability theory and uncertainty in yacht sinking

It is a tragic irony of fate that the name of Mike Lynch 's boat, The Bayesian, and its entire corporate strategy, refers to one of the most important tools of probability theory, Bayes' theorem, to the pervasiveness of which we owe a large part of the functioning of our technologies, including, above all, artificial intelligence, precisely in its most astonishing version, that based on neural networks and on - precisely - Bayesian networks. In fact, to be precise, it is not an actual theorem, but a formula, attributed to the Reverend Thomas Bayes (1702-1761) to whom we owe an initial exposition, although the current version dates from 1774, due to none other than David Laplace.

Bayes' theorem is the quintessence of anti-dogmatism and Bayesian statistics is an invitation to a continuous updating of the knowledge underlying human decisions.Unlike frequentist statistics (another irony: the one used in so-called weather 'forecasting'), Bayesian statistics treats probability as a degree of belief that is constantly updated in the light of new information, allows the incorporation of pre-existing knowledge (a priori information) into data analysis, and is particularly well suited to deal with situations characterised by high uncertainty, where information is limited and subject to rapid change. This is demonstrated precisely by the speed of our electronic tools.

Do you want the simplest example? As I am writing this article, my smartphone, at the beginning of each word, suggests several hypotheses as to how the word will continue. Each time I add a letter, it will verify or falsify the previous hypothesis and then, if necessary, suggest something updated in the light of the new evidence - i.e. in this case, a new letter making up the word. It is no exaggeration to say that Bayes' theorem (and the accompanying philosophy of probability) should be the starting point of any good financial literacy programme. It is no coincidence that Mike Lynch is known for his innovative approach to business management, based on a strong analytical component and the use of data to make strategic decisions. The name of his boat suggests that continuous learning is key. The company is a living organism that must continuously adapt to a constantly changing environment. Bayesian statistics, with its mechanism of continuous updating of probabilities, fits this view perfectly. In particular, because of the importance it attaches to assumptions and a priori information: past experience and existing knowledge, although revisable, are crucial for making informed decisions.

Bayesian statistics makes it possible to incorporate this a priori information into analytical models and to manage uncertainty. In a business environment, and in everyday life in general, uncertainty is inevitable. Mike Lynch, the Bayesian , clearly suggests what is the most useful and effective type of statistics to quantify uncertainty and make decisions under conditions of incomplete information. Even in extreme situations. Not even a tornado, however violent, can take away the solidity of this important theoretical and empirical framework.