Thinking Games and System 3
Since people began to imagine the possibility of intelligent machines, they have believed that chess playing would be one of the ultimate tests of that achievement. Why? Of course, part of the answer is that the game of chess is demonstrably and mathematically complex. Chess poses tough problems for programming, but it’s highly rule-governed play also fits nicely with a computational approach to modeling.
When we think about chess we also have in the grandmaster a model of a perfect opponent: a player as a mind. Mind is the seat of intelligence. Build an intelligent player, a computer that can beat a grandmaster, and you have succeeded at building an intelligent machine. So goes the narrative of a significant chapter in the history of advanced computer science.
There are all kinds of flaws in this reasoning. We could pick them apart, and we should. At the very least, they should make us far more critical and skeptical of Alan Turing’s “imitation game” test. For, however clever it is, it has placed arbitrary and misguided limits on the way we think about intelligence.
Let me offer a radical departure from the chess grandmaster paradigm of intelligence. It starts with the question: which of these people is a player of a thinking game, Gary Kasparov or Michael Jordan? For many, I expect, this is a nearly absurd question. Chess is a game of logic, reasoning and strategy, while basketball is a physical game, moreover a team sport, and though it is surely a game of skill, it is certainly not a contest of competing intelligences.
There are many reasons to accept this view of things, but let’s return to the question of comparing Jordan and Kasparov. The given is that they are players of quite different games. It is uncontroversial that basketball is a far more physical game than is chess. Moreover, basketball is, indeed, a team sport, one that depends on nearly “thoughtless” coordination of movement and action. Even if we were to grant, you might say, that people, including the players themselves, think about the game, that thinking, we imagine, takes place both and after, but not during play, whereas in chess, the thinking is the playing and vice versa.
We might also wonder about the different paths to becoming an exceptional player of either game. No doubt the routes to learning chess and basketball are quite different. But if we were to decompose how it is that people learn these games, how they come to understand the moves of the game, the rules they can follow and exploit, the opportunities for novelty and innovation in play, we might start to see more similarities than at first we thought. Now, of course, we might come back to the question of complexity and be tempted by the argument that the exponential complexity of chess settles the matter.
Complexity thinking, however, suggests that there’s a wrinkle in this apparently smooth picture of things. The mathematical modeling of complex, but determinate systems, like weather, has shown that while such systems are predictable up to a point, once we reach that point the “behavior” of the system becomes unstable, and radical change becomes both possible and devilishly difficult to foresee.
Another kind of complexity has been observed that makes systems not only nearly impossible to reliably predict, but that makes the systems themselves resist understanding, because the system’s dynamics are so infinitely changeable. This type of complexity has been characterized as giving rise to “wicked problems” and what makes these systems so hard to predict and model are the social dynamics that make up the system in question.
So while understanding the weather involves complexity, given how small, local changes in conditions can propagate and create massive change downstream in the system over time, being able to predict and understand the trajectory of the development of artificial intelligence, by contrast, poses a vastly greater degree of system complexity to grapple with. Because, not only do we have to reckon with engineering problems of software and hardware, along with the complexities of mathematical modeling, but also with the network effects of changable and changing human dynamics.
OK. What’s all this got to do with basketball and chess? With Michael Jordan and Gary Kasparov?
I am going to suggest that the complexity of the global weather system is analogous to the complexity of the game of chess. Computation must overcome significant hurdles, but the hurdles themselves are fundamentally knowable. Solving problems like the development of Artificial Intelligence and Climate Change, however, are more like the problem of how to create the conditions for a basketball team to reliably win the NBA championship year after year. The very shape and nature of the problem shifts and changes as we work on it, as for every advance we make, new, hitherto unseen problems emerge.
The documentary series, The Last Dance, dramatizes the years of the 1990s, when the Chicago Bulls and Michael Jordan won 6 NBA championships. The core of the show’s drama is how the team learns to function as not only a high performing unit, but as a learning and thinking machine of incredible power. There is much to learn about the dynamics, complexities and rewards of teamwork in The Last Dance, but to me the most powerful episode is the one that dramatizes the team’s adoption of the triangle offense. Assistant Coach Tex Winter who learned its principles as a player at USC under Hall of Famer, coach Sam Barry, had been marginalized for years by Bulls Head Coach, Doug Collins.
When, in a mid-season shift in 1989, Collins was fired and replaced by Phil Jackson, the Bulls began a dramatic evolution that saw Michael Jordan moved out of his role as solo superstar into that of linchpin into a complex and dynamic mechanism capable of incredible adaptation and innovation. In episode VI of The Last Dance, we begin to see this shift in thinking as they become open to the possibility that the triangle offense might open new doors of perception. As this shift unfolds we see every person on the team, from players and coaches, to management and ownership begin to harmonize their roles in new ways.
What this amazing basketball team makes evident is the incredible power of linking System 1, 2, and 3 thinking. The individual players each show the integrative power of System 1 as a deep seated system of dynamic response. Game analysis, competitive analysis and statistical analysis each and all show up as reflective aspects of System 2. The cohesion of the team, the advanced social dynamics, ability to consistently innovate in real time, and the capability to reliably outperform and defeat the other highest performing teams gives us a way to see System 3 at work within an integrated system of teaming.
There is a reason that we have long adopted the language of sport to capture the dynamics of teamwork and sought to instill the principles of successful and reliable models of high performance in our organizations. However, in all, but exceptional cases, the ability to do so remains elusive to most.
In future posts I will continue to examine and deepen this investigation of the intersection of team “play” and the integrative power of System 3 to create the ability to navigate complexity.
While I have not and likely will not follow an academically rigorous process of notation and citation in these posts, it is important to acknowledge the significant intellectual debt I owe to both key sources and conversations.
Thoughts about gaming and games are indebted to the theory of language of Ludwig Wittgenstein, Jane McGonigal’s Reality is Broken: Why Games Make Us Better and How They Can Change the World; Bernard Suits, The Grasshopper: Games, Life and Utopia; T. Chi Nuygen’s Games: Agency as Art; The Art of Game Design: A Book of Lenses, Jesse Schell; A Theory of Fun by Raph Koster; Gamestorming: A Playbook for Innovators, Rule-breakers, and Changemakers by Dave Gray, Sunni Brown and James Macanufo and Amy C. Edmonson, The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. The concept of wicked problems comes from Horst Rittel and Melvin M. Webber, “Dilemmas in a General Theory of Planning,” 1973.
Conversations that informed this post are too numerous to acknowledge completely, but I am grateful to Jen Rice, Richard Merrick, Dionna McPhatter, Dave Gray and Robin Uchida.
