winning ... analytics, or gut & intuition?

My primary school math teacher would say, "the probability of a given number showing on a fair die is 1 out of 6". And, it's often quite the battle of semantics when someone argues that the odds are five to one against you (5 to 1) if you bet on any particular number. Or is it? Because with this experience, we tossed this die 3 times and the number 2 came up all 3 times (really). Then, a senior executive of a Fortune 50 Firm we were with said, "well, it depends - what is your intuition?"

Some may posit that the calculations of odds are just 'academic'. Bain & Co. surveyed executives at over 400 companies (US$1 billion and up) to produce its report entitled, "The Value of Big Data: How Analytics Differentiates Winners" and concluded "companies who use analytics the best are twice more likely to have top quartile financial performance", but also found that only "4% of companies are 'really good' at analytics". The MIT Sloan Review in the USA where the researchers surveyed over 3,000 executives and reported in its “Big Data, Analytics and the Path From Insights to Value" with similar conclusions i.e. top-performing organisations are thrice (3x) more likely to be sophisticated users of analytics, and are twice more likely to say that their analytics use is a competitive differentiator. We only to casually observe performance of performers like Coles, Woolworths, Seek.com, eBay, Facebook, BUPA and our banks (ahem!), here in Australia.

It is no longer an academic exercise when the likelihood of any single customer purchasing the next bottle of milk contributes to what appears on the shelves (optimum assortment), at what price (customer value at optimum margins), where (location & channel type), how (targeted promotions) and who (obsession on the customer' preferences and subsequent behaviours). Then also delivering the super human ability to understand & mitigate real risks to revenue and reputation (both vital components of any viable business).

Just how difficult is this to do? Well, if it was easy, we would all be using analytics as a matter of fact. But the more relevant question I suggest, should be how "real" are the answers from analytics ...

.. does probability actually measure the real, physical tendency of something to occur or is it a measure of 'how strongly one believes' it will occur, or; does it mean a bit of both (or simply ambiguous at best)?

The 2 research activities by the world's most prominent organisations I mentioned above were 'real'. The results we now experience when predicting customer behaviour and outcomes are 'real' and tangible; far from a philosophy, fad or a sci-fi concept. But very much founded on science and mathematics.

In the example of the die above, the odds offered perfectly reflect 'relative probabilities'. A fair bet that a fair die will roll a 2 will pay a potential gambler $5 for a $1 wager (and return the person who bet this, his or her wager) in the case of a 2 and nothing in any other case. On average, 5 rolls result in something other than a 2, at a cost of $5, for every roll that results in a 2 and a net payout of $5. The 'revenue' and 'expense' exactly offset one another and so there is no advantage to gambling over the long run. If however, the odds being offered to the gamblers do not correspond to probability in this way then one of the parties to the bet has an advantage over the other. 

The probability of rolling a three with a fair die is the single number 1/6, roughly 0.17. To express the (statistical) odds against, the order of the pair is reversed. Hence the odds against rolling a three with a fair die are 5 to 1. In statistics, the odds for an event is normally defined as a simple function of the probability of that possible event. Expressing the uncertainty of this possible event as odds requires a calculation. The (statistical) odds for rolling a 2 with a fair die is 1 to 5. This is because, one can expects a single "2" event for every 5 times the die does not show "2" (i.e., a 1, 3, 4, 5 or 6).

So whilst I agree it is not possible to perfectly predict random events, the behaviour of these events can be modeled, and outcomes anticipated as illustrated in the very simplistic example of the die. As a mathematical foundation for statistics, probability applies to descriptions of complex systems given only partial knowledge of its state, as in statistical mechanics. Analytics, along with probability theory, treats the concept in a rigorous mathematical manner by expressing it through a set of axioms. Typically these axioms formalise probability with measures between 0 and 1. Central to this are mathematical abstractions of non-deterministic or uncertain processes or measured quantities that may either be single occurrences or evolve over time in a random fashion.

Casinos often use this to advantage to position themselves for advantage, which is how they almost guarantee a profit (business objective). It is often, however, a much more complex task to judge the fairness of the odds offered in a wager on a sporting event such as footy (Australian Rules), cricket or the races; and significantly more complex to anticipate sales by customers in a retail environment over the next Christmas period. But we have far better tools than ever before, and access to more data points than ever before given the pervasive use of digital devices, and electronic inter-connectivity that we all subscribe to constantly. And today, with increasingly pervasive tools (the 'toys' I refer to constantly) founded on fundamentals of pattern recognition and cognitive capabilities (ability to see, understand and interpret) of artificial intelligence and machine learning the fundamentals that surround the concept of a bet or wager in a casino are about to be over turned.

It will be absolutely absurd to run a business without the use of Analytics or Data Science.

Just as it is no longer laughable when we predict the weather, accurate to the degree; or win at baseball because of data (moneyball). We no longer are surprised when amazon suggests a product for our next purchase, or Google anticipate the next word in the sentence you have yet to write. And it will be normal to sit in a car, and not need to drive.

So, this 'flight WIN101' is about to leave, and "its all aboard now!", or get left behind. But of course, bring along human intuition in any case.