Protocol Regret

Regretting following a protocol once the outcome becomes clear

If you’re looking for something to read this summer, I highly recommend my book Thinking About Leaving, several people have told me it makes for a great read by the river or on vacation.

---

I was talking to a British friend while up in sunny Northern California for Edge Esmeralda. I’ve noticed that British people love to talk about Brexit whenever they are given permission to—especially when they’re outside of Britain. It’s as if they can finally express all that pent-up grief which, previous to landing in the United States, had been reduced to dry one-liners delivered in a pub that serves only flat beer. Americans welcome grief, but only if it makes them feel better about themselves, something that has been getting harder over the years.

Among other things, my friend called Keir Starmer a wet towel of a man, a phrase that I’ll have in my back pocket for the future. He also mentioned that “the enthusiasm that people had for Brexit lasted only two weeks, and it just so happened that the referendum happened in those two weeks.” I found this pretty relatable since I’ve noticed some corners of American society express regret at voting for Trump, based on whatever short-lived, cocaine-addled enthusiasm they had in November 2024. I labelled whatever he was expressing regarding the British referendum as protocol regret—the feeling in the pit of your stomach when you realize that you’ve followed all the right sequence of steps, and adhered to protocol but made a grave mistake.

Genres of Protocol Regret

You’ve probably experienced protocol regret every time you have gone through the TSA, a protocol that became monstrous and untamed in the wake of the 2001 terrorist attack. On my way back I saw this conversation unfold between a man whose bag had to be searched and a calm, deadpan TSA agent:

agent: I can’t give your shaving razor back with the blades. Either I can dispose off the whole thing or escort you outside so that you can take out the blade and keep the razor.
man: Why can’t you take out the blade and give the razor to me?
agent: It’s just protocol. These are the options you have.

I’m sure even the TSA agent wishes he had better things to do and worry about. Protocol regret for everyone involved.

There were several protocol regrets at the beginning of the global pandemic. Among them was the CDC’s directive to not use masks, and instead focus on washing hands and wiping down groceries. This directive was later reversed, and the CDC admitted it had been issued fearing a shortage of PPE for medical workers. A willful misdirection which set public trust lower for any future directives issued by the CDC.

In the three examples above, the protocol failed during an unexpected event (they used to call this black swan events when I was growing up, but when there are too many black swan events it’s not just that useful), something that was not considered when the protocol was initially designed. The inciting incident in these cases was a shock to the protocol, and probably a lot of people realized that the circumstances may be beyond what the protocol could handle (can only speculate), but decided to go ahead with what was accepted as truth or the “way we do things.”

Then there are protocols where the regret is slow and the effects only come into sight long after the damage has been done. In 1996, the American Pain Society (APS), with backing from the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), launched a campaign to treat pain as a “fifth vital sign,” alongside temperature, pulse, respiration, and blood pressure. Clinicians were urged to ask patients to rate their pain (often on a 0–10 scale) at every encounter, and to document and “treat to target” if the pain score was above a set threshold. By the late ’90s, pain score documentation became a routine process when a nurse checked you in at the clinic. Simultaneously, there was a push from insurance and pharma companies to increase prescription of opioids. From 1990 to 1999, total opioid prescriptions in the U.S. skyrocketed from approximately 76 million to 116 million annually, making opioids the most commonly prescribed drug class by the end of the decade. It took another 10 years and several thousand opioid overdose deaths before the medical community admitted that designing treatment protocols around treating pain as a “fifth vital sign” was not a good idea. By then, the damage had almost become irreversible¹.

Encapsulated vs Systemic Complexity

Most systems don’t encourage protocol regrets being visible because protocols are most useful when you forget about them, and treat them as truth²—the way things have always been. This is useful for protocol adoption but, as seen in the case of the pain guidelines, easy to get hijacked by white-collar indifference.

Protocols that have immediate outcomes and fatal failure modes usually have a culture of documenting how things can go wrong. A couple of years ago, a friend mentioned how she and her ex-boyfriend were on a climbing trip, where she had to belay him. The route was a bit unusual in that he was out of her sight for some part of it, and this led to an accident (no one was seriously injured), which made it to the American Alpine Club’s annual magazine Accidents in North American Climbing. Other professionalized and complex fields seem to have similar near-miss or fatal-accident reports—Aviation has FlightFax (focused on army reporting and fun to read), and CALLBACK. The fire service has something called the National Fire Fighter Near-Miss Reporting System (NFFFN). You can see these reflections and case studies as a Protocol Regret Library, an encyclopedia of edge cases, accidents, corrections, improvements and warnings.

But most of the above are protocols with ³, where most of what can go wrong is captured within a closed system and there is clear accountability for different parts of the system. The system may be a complex one, but root causes of the problems are identifiable. For example, in the case of climbing, there are always clear actors (climber, belayer), equipment, instructions, and clear differentiation between what’s external to the system and what’s internal. Thus it makes sense that documentation of what went wrong works as an efficient tool for correcting the protocol.

Protocols such as referendums, elections, security, and pain-pill guidelines, etc. have more systemic complexity, where the influences and impacts of different parts of the system can’t even be clearly separated. In the case of Brexit, while voting remain or leave was simple, what each option entailed—particularly leave—was not clearly defined and was left up to interpretation post–referendum.

Recently, in a group chat, we read the first chapter of Adam Przeworski’s Democracy and the Market⁴, which had a great definition of democracy as a process in which the losers still have a stake in the infinite game and a chance to win in the future. It did not matter who won an election, as long as it was clear that it was not the end of the game. In that sense, a referendum on something that is hard to reverse is perhaps a subversion of what the protocol was designed for? Maybe there should be an assessment of protocol fitness in such situations? A meta-protocol that asks the question, “Is this thing even supposed to work this way?”

Protocol Regret Library

*Don Draper Voice* Now with LLMs and the ability to translate inordinate amounts of text into interesting data, I think it would be interesting to have something like a protocol regret library, particularly for protocols that operate in environments with high systemic complexity. It would simply be an app into which someone can enter anonymized information that expresses regret over a protocol in the format “I regret that ______.” Such as, “I regret that I had to tell someone rushing to their flight that I could not take their blades out of their razor but he had to do it and I had to escort him for it. It made me look dumb and evil.”

It will be a slow burn that may be fruitful, or may yet be another protocol to be maintained, but I think slow-burn regrets are good to track because often by the time people notice something is wrong with a protocol, it is too late to do anything about it.

1. Overdose deaths have decreased by 27% from 2023 to 2024.

2. From

Monomythical

research on Dangerous Protocols: Because protocols abstract away complexity, they’re most effective when we don’t think about them. In earlier stages of entrenchment, however, protocols are still explicit. All participants know what the protocol is and willingly enter under its governance. Explicit protocols frequently arise in connection with social institutions—school, family, work. They are comparatively easier to identify and manipulate, because they are stated and enforced by physical constraints or a central authority.

3. Vitalik on Encapsulated vs Systemic Complexity: Encapsulated complexity occurs when there is a system with sub-systems that are internally complex, but that present a simple "interface" to the outside. Systemic complexity occurs when the different parts of a system can't even be cleanly separated, and have complex interactions with each other.

4. From the first chapter: In sum, in a democracy all forces must struggle repeatedly for the realization of their interests. None are protected by virtue of their political positions. No one can wait to modify outcomes ex post; everyone must subject interests to competition and uncertainty. The crucial moment in any passage from authoritarian to democratic rule is the crossing of the threshold beyond which no one can intervene to reverse the outcomes of the formal political process. Democratization is an act of subjecting all interests to competition, of institutionalizing uncertainty. The decisive step toward democracy is the devolution of power from a group of people to a set of rules.