The Evolution of Competition

Category: Game Theory

Death inspires me
Like a dog inspires a rabbit.
― Twenty One Pilots, “Heavydirtysoul” (2017)

Sometimes a cigar is just a cigar.
― Sigmund Freud, who probably never actually said this (but should have)

Over the past few weeks, I’ve had a fight with my wife, spoiling an otherwise wonderful night out, and email spats with two of my best and oldest friends. It was about Trump, of course. Not directly, but always on some silly tangential issue like “Should Sally Yates have resigned instead of countermanded an executive order on the basis of her personal beliefs?” or, better yet, “Is Lady Gaga authentic?” In each case, I didn’t recognize that we weren’t really talking about what I thought we were talking about, and by the time I did recognize the real issues, I was already too far down the path of combative Ben (think Bruce Banner but without the green skin) to care. Not my finest moments.

I suspect a lot of Epsilon Theory readers have had similar individual experiences of late. Certainly it seems that our collective experience as a nation and political society is breaking down this way.

What I want to write about today is not the specifics of this policy or that policy. It’s not to make an argument of any sort. It’s to write about argumentation itself, and the way in which the GAME of our politics and our society has shifted. Yeah, I know this is all very meta and has zero direct impact on your investing or portfolio decisions. But it’s actually the only thing that I think really matters for our social lives, including our lives as citizens and as investors, because it’s only by recognizing the game that we’re playing that we can survive it. Together. Maybe.

The most widely read Epsilon Theory note ever was “Virtue Signaling: Or Why Clinton is in Trouble”, published last September, where I wrote about why I thought Hillary could lose the election. The argument was that this was a turn-out election for a handful of swing states, and Democrats were all too keen to proclaim their political virtue by being anti-Trump in easy places like the Huffington Post or California metro advertising markets, where lots of like-minded Democrats would see them, rather than to barnstorm FOR Clinton in places where unlike-minded Democrats would see them, like Pennsylvania or Michigan or Wisconsin. Hubris, thy name is Debbie Wasserman Schultz and the rest of the DNC cartel.

But here’s what I wrote about Trump in that note:

Trump, on the other hand … I think he breaks us. Maybe he already has. He breaks us because he transforms every game we play as a country — from our domestic social games to our international security games — from a Coordination Game to a Competition Game.

Blowing up our international trade and security games with Europe, Japan, and China for the sheer hell of it, turning them into full-blown Competition Games … that’s really stupid. But we have a nasty recession and maybe a nasty war. Maybe it would have happened anyway. We get over it. Blowing up our American political game with citizens, institutions, and identities for the sheer hell of it, turning it into a full-blown Competition Game … that’s a historic tragedy. We don’t get over that.

Geez. Like anyone else with a public persona, I loooove being right. But I didn’t expect to be this right, this quickly. The election of Trump IS breaking us, and not because of the specifics of his policies or whether they’re right or wrong or anything like that. It’s breaking us because of the nature of repeated-play competitive games and the shifting meaning of cooperation.

That first bit — the nature of repeated-play competitive games — is a mouthful. All it really means, though, is that our real-life social interactions, whether in politics or markets or everyday life with our family and friends, are never a single, solitary game. We play the same core game over and over and over, each single interaction setting the stage for the next, and what we really should be concerned about is the overall pattern of the entire set of interactions. That’s real life, as opposed to some 2×2 matrix of Cooperate/Defect like you’d see in a game theory textbook.

And famously, repeated plays can help improve competitive games that otherwise end up in a sad equilibrium, like the Prisoner’s Dilemma. A political scientist named Robert Axelrod (not to be confused with David Axelrod of Obama campaign and CNN fame … this is a different guy) wrote a really influential book back in 1984 called The Evolution of Cooperation, where he showed that a cooperative but non-patsy player (i.e., willing to cooperate first and reluctantly forgive an opponent’s occasional defection) would, over time, find enough similarly “nice” players to create an ecosystem of cooperators and dominate, over time, those not-so-nice players who were looking to WIN BIGLY in every single interaction. Axelrod’s book was one of the most popular political science books of the past 40 years, and it spawned a cottage industry of academics looking to expand this insight in theory and practice. It’s a powerful idea because it’s a hopeful idea for nice people. If only us nice people can signal each other and band together, why golly, this proves that there’s nothing we can’t overcome together in this mean old world.

Unfortunately, the evolution of cooperation through adopting “nice” strategies is not a particularly robust finding. Or rather, it’s robust, but only in a particular subset of competitive games and only if the players agree on the meaning of cooperation. For example, if you’re playing a game of Chicken over and over again rather than a game of Prisoner’s Dilemma over and over again, being nice and forgiving doesn’t work very well. At all. Google “Sudetenland 1938” if you don’t believe me. In fact, the entire concept of repeated-play doesn’t fit neatly with the competitive game of Chicken, which is a problem because it’s the dominant competitive game form in the modern world, both internationally and domestically. It wasn’t always this way, particularly in our domestic politics. But it sure is now.

The fundamental reason that a repeated-play cooperative strategy doesn’t work in a game of Chicken is that the meaning of cooperation is different in this class of games. You see it in the title of the game itself. If you cooperate in a game of Chicken — i.e., you’re driving your tractor straight on at Kevin Bacon’s pick-up truck and you veer off from the looming crash, or you and James Dean are racing towards a cliff and you put on your brakes first — you are the LOSER. You are the COWARD. That becomes your identity and your reputation, which means that others will now treat you like a loser and a coward in the games that they play with you in the future. Compare that to the meaning of cooperation in a game of Prisoner’s Dilemma, where cooperation — i.e., you refuse to rat out your partner and cut a deal for yourself at his expense — means that you are STRONG and LOYAL. The words and the examples used to illustrate bloodless, mathematical game theoretic matrices are not accidental! If we believe that our identity is at risk in a repeated-play competitive game, we behave very differently than if it’s not. More to the point, we should behave differently if our identity is at stake. It’s the rational thing to do. If Trump inspires you like a dog inspires a rabbit, then you should never cooperate if it’s a game of Chicken with his tribe and you should always cooperate if it’s a Prisoner’s Dilemma game with your tribe. Maybe you’ll crash the car in this particular game of Chicken and maybe your partner will rat you out in this particular game of Prisoner’s Dilemma. But your identity and reputation will be strengthened, not damaged, for the next game you play with the other tribe or within your own tribe. And there’s always another game.

Okay, Ben, fair enough. We don’t want to be cowards but we still want to think of ourselves as nice. For the big identity-is-at-stake games, we should play nice strategies within our own mob and play mean strategies with the other guys. Got it. But how do we avoid crashing the car in our everyday lives? How do we avoid talking past or yelling at our friends, family, and fellow citizens with whom we share so much common ground on the really big ideas of what it means to be Americans or, more fundamentally still, a good human being?

Well … first off I’m going to suggest that we should all prepare for impact. The evolution of competition and the success of “mean” strategies in games like Chicken is at least as robust as the evolution of cooperation and the success of “nice” strategies in games like Prisoner’s Dilemma. Once you introduce, say, mustard gas into the trench warfare game, it doesn’t just un-introduce itself on its own. These bells are really hard to un-ring, and it typically takes a lot of car crashes on both sides before you get a peace treaty and a chance to rebuild a cooperative game structure. That’s at least four mixed metaphors, but you get what I mean. And unfortunately, all of these metaphors apply just as aptly to a social structure of family and friends as to a social structure of a political party or an entire nation. The evolution of competition is a powerfully contagious virus, and it hops easily from a big tribe like a nation to a small tribe like a family.

But I do have two suggestions to limit the damage that the evolution of competition inevitably spews in its wake.

First, whatever competitive social interaction we’re having, at whatever level we’re having it, the most important thing in that interaction is to figure out the meaning of cooperation for yourself and whoever you’re dealing with. Otherwise you’re going to find yourself playing a different game from the other person, and that never ends well. This is a tough piece of advice to follow (myself included!) because we assume that whatever our “identity weighting” might be for a given issue, the person or group we’re interacting with attaches that same meaning. So, for example, if you voted for Clinton as an affirmation of a personal identity that rejects the racism and sexism you see in Trump, your natural assumption is going to be that anyone who voted for Trump similarly did so as an affirmation of a personal identity, but one that accepts racism and sexism. Or vice versa. Or whatever. We’ve all seen a dozen variations of this theme over the past eight weeks, and we’ve all (yes, every single one of us) engaged in it, as well. This sort of projection is an innate behavioral bias of the human animal. I get it. But it is also entirely wrong-headed when it comes to complex and over-determined social behaviors like voting. Or buying a stock. Believe it or not (and many people reading this note won’t), behaviors like voting or purchasing or speaking or tweeting are not necessarily markers of personal identity. Maybe they are, and when they are they MUST be respected if you care about having a peaceful social interaction. But maybe they aren’t. And that must be respected, too.

Second, it’s crucial to recognize that not all political arguments or competitive games are really existential in nature or fraught with questions of identity. Not every tweet is a constitutional crisis. Sometimes a cigar is just a cigar. This is also a tough piece of advice to follow (also myself included!) because the ringleaders of the various Team Elite cabals, whether it’s the Trump Plutocrats or the Davos Globalists or the Central Bank Mandarins or the NeoCon Spymasters or whoever, are working diligently day in and day out to convince you that it is. That every action or statement by the other ringleaders is an OUTRAGE. That this is how Hitler got started or how liberty is lost.

Of course, the really scary thing is that this IS how Hitler got started and it IS how liberty is lost, it’s just not clear to me which of our contending factions or geographies is supplying the 21st century version. History rhymes roughly; it doesn’t repeat neatly.

Meanwhile the barrage of fiat news and alternative facts continues from all sides unabated. We are caught in the crossfire of the “mean” strategies implemented by the various factions as they quite rationally engage in a massive repeated-play game of Chicken, where winning means mobilizing the hearts and minds of the cannon fodder. And by cannon fodder I mean us.

It’s the oldest saying in poker, and one I can’t repeat often enough. If you’ve been playing poker for 30 minutes and you don’t know who the sucker is … it’s you. We are — all of us, without exception — being played. That doesn’t mean we stop playing the game, whether it’s the game of markets or the game of citizenship. It means, though, that we resolve not to be the sucker. That we turn a clear eye to the stories that others tell us and the stories that we tell ourselves. That we demand to be treated as the rightful, autonomous owners of our identities, and we extend that right to others.

Know thyself.

Treat others as you would have them treat you.  

Pretty good advice 2,000 years ago in some pretty hard times. Pretty good advice today.

epsilon-theory-the-evolution-of-competition-february-7-2017.pdf