Tit for tat
Thoughts on cooperating (and competing) with others
Trust, But Verify
One afternoon in 2019, I Venmo’d a random stranger $50 for his entry spot to the Broad Street Run. Two minutes later, I received a confirmation email that his bib was officially mine.
He did not thrift me for my $50, though he easily could have. Cooperation ruled the day.
This makes me feel a bit torn about cooperation.
On one hand, I’m happy that cooperation is a key component to the wealth-generating, win-win transactions of the modern economy. When chimps take turns grooming each other, they both end up with less parasites. You scratch my back and I scratch yours. Humans probably figured this out a few hundred thousand years ago when Neander shared his extra meat with Thal one day, and Thal shared his with Neader the following. They both were better off by cooperating (and trading) with each other.
On the other hand, I also realize that cooperation is not always the optimal move. In certain game theory contexts, sometimes the selfish incentives to screw over the other person outweigh the more altruistic incentives to cooperate. I imagine one of the boroughs in New York City invented the term sucker for someone who consistently gives up too much for too little in return.
Reputations matter, and no one wants to be known as a sucker, or even worse, a cheater.
This raises the critical question: in which situations should we cooperate with others, and when should we just worry about our own interests?
On their awesome chapter on negotiation in Venture Deals, Brad Feld and Jason Mendelson argue that it depends on whether the engagement is expected to be a short-term, winner-take-all battle or whether it has potential to be a long-term working relationship. (1)
Short-Term
In a short-term, winner-take-all battle, because there can only be one winner, there often are fewer incentives to cooperate with others because there is no possibility of a win-win outcome.
There are two classic examples of this type of game:
1) The big lawsuit that will bankrupt the loser and proudly vindicate the winner. Think Hulk Hogan suing Gawker for over $100 million.
2) A nasty divorce in which reconciliation is no longer possible.
Because there are scant incentives to cooperate, the default behavior is often total war: hire attack dog lawyers, go on a smear campaign, and win at all costs.
While it would be nice if everyone “could work things out”, unfortunately, it’s not smart to bring a knife to a gunfight. In a game such as this, Feld and Mendelson warn that if you’re not prepared to march to the sea, you want to avoid getting involved in the first place.
These short-term battles are interesting, but I think they are relatively rare. More commonly, we are involved in long-term cooperative games, where reputations matter and there are opportunities for win-win situations. A venture capitalist may be more cooperative with startups because she does not want to develop a reputation for taking advantage of founders, even if she could win via schemey negotiation in the short-term.
Let’s dig into another example.
Long-Term
Suppose there is a new hire at the office working directly in your department. Option one might be to smack him in the back of the head while he is using the Keurig to assert your territorial dominance. Or if you are more passive aggressive and/or went to an elite MBA program, to take advantage of his naïveté by sabotaging his work or convincing him how much fun it is to make copies and/or paint the white fence outside the office.
"'Like it? Well, I don’t see why I oughtn’t to like it. Does a boy get a chance to whitewash a fence every day?'
That put the thing in a new light. Ben stopped nibbling his apple. Tom swept his brush daintily back and forth—stepped back to note the effect—added a touch here and there—criticised the effect again—Ben watching every move and getting more and more interested, more and more absorbed. Presently he said:
'Say, Tom, let me whitewash a little.'"
—The Adventures of Tom Sawyer (1876), Mark Twain
Option two is to be more cooperative. If you can share knowledge with him as he gets up to speed, he can share his new knowledge back to you, and you both can be better off. You can develop a working relationship, perhaps covering for each other while on vacation. In the United States, this is (fortunately) how most people operate in professional environments.
Let’s just hope that this new co-worker has not just read The 48 Laws of Power or is otherwise one of those Machiavellian types that is expert at sucking people dry for intel and favors while giving back little in return.
Over time, which option is best? Should you cooperate or compete with this new person? Clearly it seems to depend on the situation and the other person. Is it a mix of both?
Indeed.
In the 1980’s, political scientist Robert Axelrod organized a tournament among researchers to get at this answer. Which submitted computer algorithm would be best at repeatedly playing a specific type of strategic game, the prisoner's dilemma? Like our situation at the office, a single round of the prisoner's dilemma involves a choice to more altruistically cooperate with the other player, or more selfishly cheat them for (potentially) more points.
The winning strategy for repeated cycles of the game, submitted by Anatol Rapaport, was the simplest algorithm of all: Tit for Tat. (2)
Tit for Tat always cooperates with the other player first. After that, it simply mimics what the other player did last. If the other player cooperates, Tit for Tat cooperates again in the next round. If the other player cheats when Tit for Tat cooperates, Tit for Tat cheats them in the next round as “punishment.”
Tit for Tat achieves an almost ideal balance between trusting pro-sociality and more circumspect, anti-sucker street smarts.
Where Tit for Tat is sub-optimal is in one very specific situation: when the other player is a full-time cheater. Tit for Tat will get cheated in the first round and then will repeatedly cheat to mirror the other player, never to recuperate that initial loss.
Can you negotiate with Vladimir Putin, who famously stole Robert Kraft’s Super Bowl ring after Putin asked him to try it on? Unless you’re comfortable in playing the same Machiavellian, psychopathic games, the answer is probably not.
As Naval Ravikant has said:
“If you can't see yourself working with someone for life, don't work with them for a day.”
With a permanent cheater, the best option is usually to stop playing the game altogether.
Returning to our office example, the advantages and limitations of Tit for Tat suggest you should reach out and “cooperate” first with your new co-worker but remain more circumspect if he seems to be taking advantage of your pro-sociality. But don’t hold any grudges; in case he reaches out with a new opportunity, be ready to re-engage for a potential win-win.
Personally, and I’m sure you’d agree, I think it’s a bit cold and Zuckerbergian to robotically apply Tit for Tat to social interactions. Instead, my intention is to introduce it as a helpful framework for navigating the often highly political and competitive workplace environments in which long-term reputations matter. With that being said, there is also a different option altogether: find the people you know you can trust all the time, so you can worry (less) about when to cooperate. This can be a huge advantage at the group-level. As researcher E.O. Wilson has put it,
"Cheaters may win within the colony, [...] but colonies of cheaters lose to colonies of cooperators." (3)
Finally, recall that Tit for Tat doesn't apply to short-term, winner-take-all games. In these contests, such as a bidding war, if you choose to be involved, be prepared for the ruthless, uncooperative behavior that is incentivized by the game's very nature.
In general, it’s crucial to know the type of game you’re playing, and more importantly, who you’re playing with.
References
1) Venture Deals, Brad Feld and Jason Mendelson (pp. 191-196)
2) Behave, Robert Sapolsky (pp. 342-354)
3) The Social Conquest of Earth, E.O. Wilson (pp. 162-163)