Make Sure That You’re Optimizing for What’s Important (to you)

What are you optimizing for? Are you sure that’s what you want?

I recently listened to an episode of Revisionist History (my favorite podcast of late) in which Malcolm Gladwell spends much of the episode elaborating on his ‘life-rule’ of “Pull the Goalie”.

Gladwell is referencing an academic article Pulling the Goalie: Hockey and Investment Implications that uses pulling the goalie at the end of a hockey game to talk about risk management. I was almost scared off by the ‘academic’ orientation of the article (so many footnotes!) and that it’s mostly concerned with investment, which isn’t a particularity interesting topic to me, but the article’s authors did an amazing job of making the concepts easily accessible through the lens of professional hockey.

You don’t need to know much about hockey to follow the arguments, just these 2 points will do:

1. At any time in a hockey game, you can take your goalie out and get an extra offensive player
   When you do this, you of course increase your chance of scoring, but also leave your goal wide open, increasing the chance of your opponent scoring as well.
2. A tie in hockey is substantially better for your team than a loss
   This is obvious of course, but it’s particularly relevant in that a team get ‘points’ in the hockey standings for a win or a tie, but 0 points for a loss. These points ultimately decide if your team gets to play in the playoffs, as well as your playoff seeding. Simply put, points are very important, perhaps to be pursued above all else.

The concepts in this article are so accessible and well put, I’m going to liberally reference it (read: copy/paste!) in the remainder of this post rather than try to summarize, as there’s no way I can do any better.

Here’s one of my favorite bits:

Now, let’s get to our real purpose (besides just loving both hockey and math). There are some important risk management and investing lessons to be learned by considering this optimal hockey strategy problem. The most basic lesson is to make sure you are thinking about the right risk. Pulling the goalie always increases the volatility of numbers of goals scored, and is a negative expectation in terms of the score. For those reasons it is often used as a metaphor for a high-risk, desperation move. However, the point of hockey is not to maximize the differential between the goals your team scores during the season and the goals it gives up (if it were, no one should ever pull a goalie). The point of hockey is to maximize the number of standing points — a team down by a goal with short time remaining gains a lot by scoring, and loses little if the other team scores — which argues for a different measure of risk and return. As we have shown, pulling the goalie actually reduces the risk of losing the game — it’s an insurance move — and this is the proper risk measure.

This reminds me of an issue that I think about all the time, both in work and in life more generally, and gets us to the subject of this post (finally!), what we’re “Optimizing” for.

So in this case the authors are pointing out that, in the choices that they make at the end of games where their team is losing, coaches are not optimizing for what most people might argue is “point of hockey”, which is “points”. I’d argue that they’re optimizing for something else, either intentionally or unintentionally…

If you choose to read the entire article, you’ll probably be convinced as I was that coaches should change their behavior and pull their goalies earlier when they’re losing at the end of a game. The authors point out 2 compelling reasons why coaches fail to act in this way:

Two reasons have been advanced for this failure to act. First, coaches are not actually rewarded for winning. They are rewarded for being perceived as good coaches. Obviously, the two are closely related but not exactly the same thing. If a basketball coach gets his team to execute crisp offensive plays with few turnovers that lead to two point baskets on 50% of possessions, he’s deemed an excellent coach. If his team still loses 100–102, well, his players just weren’t quite good enough. If the same coach encourages his team to “run and gun” threes, with lots of turnovers and misses, but scoring on 35% of possessions, he’s clearly lost control of his team. If they win 105–102 it’s perceived as just luck as everyone knows three point shots are risky. Essentially winning ugly is undervalued versus losing elegantly; and losing ugly can be career suicide. Once again, the way you measure risk matters in making the optimal decision.

…

The second reason coaches shy away from actions with short-term risk is that sins of commission are far more obvious than sins of omission. The hockey coach who pulls his goalie down 0–2 with ten minutes to go and loses 0–5 will face harsh criticism from every quarter. A coach who quietly loses 1–2, pulling his goalie only in the final minute, can hold his head up and say his guys played hard but the puck didn’t roll their way tonight; it was a close game, and they’ll work even harder to get the breaks tomorrow (giving 110% of course).

Wow! I love these reasons in that they illustrate out that hockey coaches might actually be primarily concerned with something else besides winning, and it basically boils down to how well they appear to be doing in their goal of ‘winning’, and not how well they are actually accomplishing that goal.

However, the article mentions one more possibility toward the end, which I think is probably the most salient point:

Also, our model absolutely ignores that hockey is entertainment. Perhaps, more speculatively, fans enjoy the last 1–2 minute drama very much, and even will stay tuned to a game with a late two goal differential. But, pulling the goalie very early when down two goals leads to a very large chance of being down three. That would kill the entertainment. Might coaches feel a desire, or even tacit pressure, not to ruin the fun? In other words, our model is myopic, only caring about maximizing the points from the game. Perhaps coaches, and the league in general, are better long-term present value maximizers than our model as providing more entertainment maximizes “franchise value,” even if slightly costly in terms of points. Particularly, if all teams do it the overall entertainment value of hockey is increased with no ex ante advantage or disadvantage to any one team as they’re all acting alike. Now our analogy to business has drifted to collusion!

Is this possible? Is this giving hockey coaches too much credit? Perhaps, but perhaps not.

Even if the hockey coaches are completely informed and rational (not a given by any means), I’d even go so far as to argue that they are behaving exactly as they should by not pulling their goalies earlier, even though these decisions to leave their goalies in mean that they’re going to lose more games. In other words, I’m arguing that the coaches should not care most about winning (or tying) games. Their most important job is the same as that of the players, to support the goal of professional hockey to provide entertainment.

Let’s say for the sake of argument that this is true, that hockey teams (including their coaches) are optimizing for entertainment, and not for winning games, making the playoffs, or wining the Stanley Cup. If you’re a hockey fan, would that matter you? Does that change your expectation of hockey or the way that you engage with hockey? Would you stop watching hockey altogether as it ceases to be entertaining to you if your favorite team’s end-goal is entertainment (and not competing or wining)? Can entertainment in professional sport only be fully realized when entertainment is not the end-goal?

So I don’t really care about hockey, and maybe you don’t either, and even if you do maybe this entire argument isn’t a big deal to you and feels like splitting hairs, and that’s fine. But regardless of how you feel about hockey, hopefully you can see how in other areas subtle distinctions in optimization choices could have important consequences.

• In how you spend your money, are you optimizing for getting as much as you can while spending as little as possible? Or perhaps supporting local businesses? Or giving to charity?
• Are big tech companies (just to name a few: Facebook, Google, Uber, Airbnb, etc.) optimizing for making your life easier trough their services? Or are they optimizing for innovation? Or for building up data stores to secure their place at the top? Or for profits today? Or profits in the future? Corporate directors have legal obligations to their shareholders, how much does that factor into their decisions on what they’re optimizing for?
• In your politics are you optimizing for your ideals? Or for pragmatic goals that might require compromising some of those ideals?
• In your diet, are you optimizing for long-term health or for enjoyment? (🍩🍩🍩)
• In your work, are you optimizing for immediate productivity or for learning new skills (future productivity)? Or for salary? Or for work/life balance?

I’m not arguing here for any of these over the other (that would be another post!), and not all of these choices mutually exclusive (but at least some of them are). What I am arguing for is that we need to be asking these questions both of our ourselves and the organizations what we’re involved with.

Take the tech company example. It should matter to us what these companies are all about. That information should have an impact on which services we use and how we engage with them. Will we optimize for our convenience by using the given service, trading off our privacy for that convenience?

Yes, the privacy/convenience trade-off is a real thing and it’s nothing new. In the book World Without Mind: The Existential Threat of Big Tech, Franklin Foer points out that Louis Brandeis was concerned about this back in the early 20th century:

But Brandeis hated the prospect that society might elevate efficiency to the highest value. Convenience was nice, but we shouldn’t sacrifice ourselves to achieve it. His fear was that the benefits of efficiency might lure us to surrender our liberty. That’s the authoritarian temptation: that liberty comes to seem a small price for trains that run on time. To update the thought — it’s not worth having free email if the price is our privacy; next-day delivery is nice, but not if the consequence is a sole company dominating retail, setting the market price for goods and labor.

I think we’re starting to learn this lesson, but we still have a long way to go. We can start by asking, even if just to yourself: “What as a company are they optimizing for?”.

But the next question is just as important, if not more so: “What are we optimizing for”. Is it price? Or convenience? And at some point, we’re going to cross a threshold where the cost of what we’ve been prioritizing has become too great and we going to have to change what we’re optimizing for.

The same goes with anything you care about personally or professionally: work, money, politics, whatever. Upon reflection, we can sometimes discover that, without knowing it, we’ve been acting in a way that prioritizes X when we really care more about Y, and that we need to reprioritize. This isn’t really anything new, but it does take a conscious effort, it doesn’t happen on it’s own.