Strategy for Humans

A computer once beat me at chess. But it was no match for me at kickboxing”


Emo Phillips

It’s common to the point of cliché to see some kind of chess symbolism illustrating a work on strategy (our own company website being a case in point). Chess seems to embody the very essence of strategy – a contested environment, a victory objective, fierce intellectual challenge, thinking ahead.

The ‘Chess Player’ as metaphor is also rolled out figuratively whenever we feel the need to characterise someone as calculating, competitive and even ruthless – the epitome of ‘the Strategist’.

Although he got his comeuppance in From Russia With Love, this cold, machine-like stereotype has probably been further reinforced by the rise of computers as more-than-able chess players. (When world champion Garry Kasparov was beaten by IBM’s Deep Blue supercomputer in 1997 there was a media storm. Today most good PC-based ‘chess engines’ are far stronger than any human player who’s ever lived(1), and this is scarcely remarked upon.)

The stereotype is misleading though.. Computers have become stronger than humans, yes, but they haven’t done it by playing like humans. What’s more, even when the human grandmasters could beat the best machines, they didn’t do it by playing like the machines, except better. According to Kasparov, they did it by playing their own strengths to the machines’ weaknesses, in three ways in particular(2).

Firstly, chess programs generally work by calculating several moves deep, more or less by brute force (this was certainly true of Deep Blue). They can do this to a greater depth and over a much bigger range of alternatives than a human player. What they can’t do is what a human can – think imaginatively. A human grandmaster may sense the opportunity to corral the opponent’s queen in a corner later in the game, and start working towards that outcome, without at that point being able to calculate exactly how to get there. Machine opponents could be beaten by working towards configurations which were outside the computer’s number-crunching horizon until it was too late for them to react effectively.

Secondly, although chess programs can be loaded with an ‘opening book’ – a database of tried and tested opening sequences which might help them to maneuver to a strong position over the first half dozen moves or so, they would have no understanding of why an opening line was a potentially good platform for the middle game. Once ‘out of book’, they’d be number crunching again from scratch. A human player learns not just openings but the principles behind those opening choices, which can be exploited strategically later in the game.

And thirdly the machines only play the objectively strongest line, without any thought for how difficult or risky that line is, or what impact it might have on the opponents mental equilibrium. A human might take an easier to calculate, and therefore safer, route to victory over a more spectacular win that requires a precise move at every step. Alternatively, a complicating move might be chosen over a straightforwardly good one, solely to make life more difficult for an opponent under time pressure, or to try to force an error.

This isn’t a free pass of course. Kasparov acknowledged the need to be able to calculate forward at the limits of his ability, on top of his quirky human advantages. And of course nowadays those advantages aren’t sufficient.

Be that as it may,  the skills Kasparov brought to bear to compete effectively against Deep Blue nevertheless illustrate the very essence of what strategy is. Translated into the business context, strategy for the merely human among us looks something like this:

  • It begins with a thorough understanding of current conditions, both internal and external, and what this implies for our capabilities. There are lots of tools to help with this. We don’t start with a goal.
  • We envision some future configuration of activities or circumstances which in principle would improve our positioning, typically by building competitive power, and which we judge to be potentially achievable, even though we probably don’t have a complete understanding of how to get there. There aren’t many tools to help with this, not least because, although our capabilities are finite, our objectives are limited only by imagination (3). Aligning these two conditions is in many ways the essence of strategic thought.
  • We proceed in stages, recognising that others may actively try to frustrate our intentions, and at each stage reviewing and potentially modifying either our approach or our objectives ( “adjusting ends so that realistic ways can be found to meet them by available means”(4)).

If this doesn’t look like a plan, that’s because it isn’t. A plan presupposes a sequence of defined steps, through predictable stages,  to achieve a defined goal. Strategy on the other hand is needed when our efforts are contested by others: the only thing we can be sure of is the starting point. This makes flexibility necessary, because we’ll often find that whenever we make a ‘move’, we don’t quite end up where we expected. To quote Mike Tyson, everyone has a plan until they’re punched in the face. At that point, we re-evaluate. We may have a plan to execute change in the business of course, but that isn’t the strategy in itself. We only plan what we can.

One of the most challenging areas in business strategy is the ‘envisioned future’ – defining strategic objectives. We tend to confuse potential derivative outcomes, ‘victory conditions’, with strategies (increasing profits by 30% isn’t a business strategy, any more than checkmating the opponent’s king is a chess strategy). We also tend to revel in setting objectives which are untethered from reality (or which become untethered when competitors, customers or even our own colleagues don’t react the way we assumed they would).

Business isn’t chess. There are many more potential variables to consider at any given point. The commercial world is infinitely more complex than a chess game, much less amenable to brute force calculation. But it’s precisely this point which makes our transcendent human characteristics come into their own – the ability to envision a future and move towards it without a complete understanding of how to get there,  the ability to understand deeply the principles underlying a strategic position in figuring out a way forward; a sense of when to take a risky or conservative path; and an instinct for the psychological and emotional consequences of a choice. Put like this, our human brains may well be uniquely suited to navigating the complexities of socio-economic life. Which, in an age of near panic over the rise of AI, is a cheering thought.

~

(1) As at the time of writing, the open source Stockfish program, which can be downloaded free of charge, has an ELO rating of well over 3,400. Current world champion Magnus Carlsen has the highest rating of any human in history, at 2,840, with 1,500 or so representing about the average for tournament players.

(2) Kasparov, G., Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins (2017).

(3) See Gaddis, J.L. On Grand Strategy (2018)

(4) Freedman, L. Strategy: a History (2015)

Leave a comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.