The world is divided in many ways, but every way that divides also compresses down to a strategic choice most do not consider.
Are you operating in a world dominated by a Gaussian distribution, or a power distribution?
We are all familiar with the normal curve, which is a gaussian distribution. Many, if not most things across the commercial, natural, and institutional worlds we live in operate on a normal curve. Human height, manufacturing variation, travel time to work, cricket scores, heads thrown in a game of two-up. All are plottable in a normal curve, and the larger the sample, the more ‘normal’ the curve becomes. It is the statistical basis of every political poll you have ever seen.
Italian mathematician Vilfredo Pareto noted the wildly differing income levels amongst Italians. This pattern did not comply with the normal curve leading to the insight that reversion to the mean did not apply at a micro level. The differences in income of individuals, 80% of which accrued to 20% of Italians, led to the 80/20 ‘Pareto rule’.
Wherever you look, you see the pareto rule at work. A closer examination often also reveals a power distribution playing a key role.
Cricket is a game where the scoring is constrained by the maximum of 6 runs. Hit the ball over the boundary on the full, and it is six runs. Whether the ball clears the boundary by an inch, or by 100 meters, the maximum runs scored is six. However, if you were to perform a power distribution of the batsmen who had hit a six in test matches, the 80/20 rule prevails. Of the 1,828 sixes scored by Australians in test matches over the last 25 years, the top five batsmen scored 79.6% of them, the other 448 batsmen combined scored the other 21.4%.
By contrast, Rugby league is not constrained by a maximum. The score in Saturday evenings semi-final game between Cronulla and Easts was 46 to 10. The commentators were discussing the chances of Easts breaking 50 in the last few minutes of the game, as there is no limit applied to scoring beyond the time of the final whistle ending the game.
These differing perspectives lead to the question referred to in the header: Are you in a commercial context that is constrained by a reversion to the mean, or are you competing in a ‘pareto-like’ context?
The obvious follow up question is ‘does that optimise the commercial outcome’
The domain you are in drives the nature of the strategic choices made.
A domain dominated by a power curve business model is one where many bets are laid, knowing most will fail, but the few that work will deliver a disproportionate share of the total value delivered by all bets. The challenge is that those that will deliver the disproportionate outcomes are unknowable when the bets are laid. Typically venture capital works in this domain.
To continue the cricket analogy, coaches might look for the characteristics that made Gilchrist, Smith, and Co the outliers amongst the long tail of cricketers who score none or few sixes in their careers, and coach them to hit those sixes. Most will get out even more quickly than they normally would if trained for defence, but perhaps one or two might join the elite group in the 20% club.
In the bell curve or gaussian analysis, incrementalism is the driving force. Coaching the average test bowlers to turn their average test scores from 6 or 7 into the teens over time will optimise their capacity to contribute to the total reliably.
Long term success depends on both approaches being applied judiciously, as trade-offs must be made.
Scott Boland has a bowling strike rate of 39 balls per wicket of Australian test bowlers over the last 25 years, and a batting average of 9. He struggles to find a place in the test team other than in the absence of Hazelwood, Stark or Cummins, despite having the best balls/wicket ratio. (let’s not debate the obvious impact that his balls/wicket ratio is heavily influenced by his astonishing 7 for 55 in the 2021 boxing day test)
As someone running a business, you must also make these trade-offs between incrementalism and ‘hitting for the fence’ in the ways that best suit the circumstances of the business and its competitive context.
Note: explanation of header graph.
- The top 5 scorers of the 1,828 sixes hit over a 25-year period, Gilchrist 5.5%, Hayden 4.5%, Ponting 4.0%, Warner 3.8%, and Smith 3.7%, together account for 79.6% of all sixes hit by all 453 Australian test batsmen over the last 25 years.
- 276 players hit no sixes, the average being 3.9 sixes each.
- Gilly, is 8 standard deviations from the mean, making him, as we knew, a freak!
