Minimizing Defection Might Be the Solution to the Tragedy of the Commons | Hacker News Reader

Minimizing Defection Might Be the Solution to the Tragedy of the Commons

getpolarized.io

21 pointsburtonator47 comments

1) I find it amusing that an article regarding the supposed "Tragedy of the Commons" uses several Wikipedia quotes 2) The Tragedy of the Commons was thoroughly analyzed, with a great deal of empirical, historical data, by Elinor Ostrom. https://en.wikipedia.org/wiki/Elinor_Ostrom Yes, one component of successful, long-lived common resources is related to group size. But you don't have to take my word for it, read her book, "Governing the Commons".

There's a huge tendency for smart people like computer scientists to think that, because they're smart, they can come up with "the solution" to systemic problem X without reading into the deep academic literature around it. I knew when I saw the title of the article on the front page that I (or someone) was going to need to preach the gospel of Ostrom in the comments.

That tendency is universal. Perhaps CS and technical people are more likely to publish their thoughts in blog posts, but it's natural for everyone to talk about what they find interesting, experts or not, and we all underestimate what we don't know. This is how human conversation works, and it's arguably harmless—phrases like "shooting the shit" communicate this.

I wonder if once a post shows up formally on something like HN's front page, we see it as implicitly claiming authority in a medium-is-the-message way, even if it all it's doing is thinking out loud. Then the article seems disappointing when it turns out just to be someone's notes on shit they were thinking about.

That's an unfair sweeping generalization about "smart people like computer scientists". Far more fair to say they enjoy thinking about, discussing, and attempting to solve problems. Trying to solve problems is a passion for all the good programmers I know.

And predicting / guessing at solutions without having done the "deep academic literature" reading is actually an excellent learning technique that connects you more deeply to the reading if you decide to go down that path. Or do the predicting in a blog post and get free answers from the experts correcting you.

I've "invented" a few things in my life only to later find the existing patents. I had so much fun doing it and learned so much, I'm glad I didn't know about the patents ahead of time.

I mean, great that you had that experience, and thankfully you didn't make a blog post talking about how you solved ______. It's fine to come up with ideas on your own, but if you don't know enough about the field to know whether you're on the right track, don't announce to the internet that you've found the solution.

That's a strawman. Author's own words, right in the intro:

"I might be wrong but would love some feedback. It’s an exceedingly complicated topic which has been researched for hundreds of years so it might be that this observation has been made before."

And that very important word "might", right in the headline too.

>There's a huge tendency for smart people like computer scientists to think that, because they're smart, they can come up with "the solution" to systemic problem X without reading into the deep academic literature around it.

People that do come to solutions to problems (including deep problems), often come to them "without reading into the deep academic literature around [them]" first, so there's that...

True that, but it's not only computer scientists: https://xkcd.com/793/

I’ll take analytical thinking that hasn’t found the right information yet over folksy intuition any day.

> What if we could solve this problem through core changes in our society?

Yes, we call the necessary changes 'the state'. It's exactly to solve a collective action problem that the state first appears: military defense and preventing actual defection in war. Strong states prevent defection but then suffer from bad principal-agent problems, since rulers tend to advance their own goals - like waging war.

I have to agree with some of the commenters that your write-up is pretty bad, because far too general for such a well studied problem.

Let me try to help you along in your thinking: The tragedy of the commons is not a generalized version of the prisoner dilemma. I guess you meant to say a prisoners dilemma is a more generalized version of the tragedy of the commons. This is however also not true. Logically because there is no sensible pay-off where everybody defects, but the other player. The commons would be largely destroyed way before everyone gets their pay-off so all the players would never collectively decide to do that (and decision is a integral part of the prisoners dilemma). You could make it into a NxN dimensional prisoners dilemma, but that is not useful for another reason. The tragedy of the commons is about future benefit (keeping the commons so everyone can extract more than they could individually at a single instance) a prisoners dilemma is an instant game. You could model the tragedy of the commons as a repeated prisoner dilemma. The problem is that free riding is not a rational option if you repeat it often enough. In the real world free-riding works, because we expect to play a limited number of rounds and in the long term we are all dead. The number of repetitions one expects to play is however not the same for all players.

Limiting defection is in this case not external to the tragedy of the commons, as any way to resolve the prisoners dilemma is, because it reduces the pay-offs the participants get. Preventing defection costs and there is a balance to be struck to the costs to preventing an additional free-rider and letting him free-ride. There is also a moral component to this. Limiting defection means also being able to limit entry. The commons has de-facto become 'our commons'. There is the description of a 'mutual protective agency and one of the most common is the state.

Anybody interested in "the commons" should familiarize herself with the work of Elinor Ostrom.

Unionization is a bad example. If all union members were unable to defect, the employer would still reject their demands for above market wages, because he could replace them wholesale with workers from the larger labor pool, at market rates.

If all workers in the world were forcibly enrolled in a global union and prevented from defecting, and this union demanded a doubling of wages, this would just lead to 1. some companies shutting down, and 2. other companies seeing profit margins decline to below the rate that the supply of labor versus the supply of capital determines as optimal for economic development, resulting in more unemployed workers, and a one time boost in wages for employed workers, as labor's share of income increases, at the expense of lower recurring increases in wages, as the rate of economic growth declines.

Anyway, smart contracts could indeed be a solution to defection. They are a potentially massively scalable coordination tool, allowing huge numbers of people to enter any given agreement at minimal cost.

>Interestingly enough, the optimum solution is always present when N=1. If there’s just one actor (you) then you’re not likely to destroy your own property.

Two objections:

1) Who said one person would necessarily understand it as "[his] own property"? Unless it is truly their property (e.g. their own house), it's conceivable that a single person could still consider the place they are in expendable (because e.g. they could move elsewhere, or they care more for short term profit for it and don't consider long term issues from their behavior).

2) Even worse, who said people respect their own property that much? People can't respect their own bodies (sedentary lifestyles, obesity, junk food, etc), their own families (domestic abuse, cheating, etc.), their own houses (tons of neglected, broken down houses), or even their personal lives (from alcoholism and drugs, to all kinds of subtler self-sabotage)...

Solving the tragedy of the commons. I wish you luck.

Interestingly, you hint at, but don’t really acknowledge one of the classic solutions (typically pushed by the libertarians): privatize the commons. This is equivalent to your intuition of trying to set N=1.

Of course, the other classic solution is regulation, which is not too far from your union example.

Of course, both the classic solutions have downsides. Privatization makes it harder for everyone to share the good, and also leads over time to stratification and unequal society. Regulation means bureaucracy, which easily becomes captive to politics and money.

so the non-libertarian answer to privatization of the commons is the pigovian tax. It's got most of the advantages (it uses the power of the market to maximize results per unit effort) and avoids some of the disadvantages inherent to privatizing things like air.

Take, for example, the problem of limiting CO2 output. This is a global problem; there is no way to realistically privatize it. but, at least the people under the same treaty could tax the common forms of CO2 output.

Charging per ton of CO2 output, many economists believe, is better than simply regulating how much CO2 a person can output because by giving it a dollar per tonne value, you are encouraging the market to eliminate CO2 output where it's cheapest and easiest to do so, while allowing the market to continue to output CO2 where it would be really expensive to avoid.

'cap and trade' is a variation on this that is similar but different; instead of setting the economic price of a tonne of carbon, for example, you set an absolute limit for how much carbon can be released, and then lets the market decide what to use that for.

That's exactly why we need a CO2 tax now and why the current "right" (who have largely been captured by the fossil-fuel industries) are fighting it tooth and nail.

There are known “solutions” to the prisoners dilemma, which is the broader problem. (I say that in quotes because it always involves changing the problem in some way like adding a mafia Don or playing multiple rounds).

There are a wealth of real world examples of communities working together to collectively manage the commons without either privatisation or state regulation.

If you want to know more you should familiarise yourself with the work of Elinor Ostrom.

That's great. If a group of people comes together to collectively manage a commons, that sounds like regulation to me. It's all semantics though.

There is a subtle technical difference in the literature, where regulation is externally imposed and codified, while communal management is devised, enforced, and most importantly, flexibly changed by those actually involved with the appropriation and management of the resource.

Thanks. That makes sense.

It seems to me that the problem with the Libertarian approach is finding a just way to actually privatize the commons. Since they are by definition common property, the rich have no greater claim on them than the poor. In future generations it wouldn't be just that some people had access to the former commons and some people don't due to the actions of their ancestors.

The only solution that I can see would be for those who own the commons to pay the profit that they gain from that ownership to those who do not own the commons. Which I guess is effectively Georgism.

>Interestingly, you hint at, but don’t really acknowledge one of the classic solutions (typically pushed by the libertarians): privatize the commons.

Then, as many countries have discovered, you have another problem...

No doubt many specific problems could be addressed by work on the tragedy of the commons.

But at the same time, theories around a healthy market _require_ defection in order to "guarantee" competition. So in its rawest form, we have an economic system that requires a system that is known to produce some bad outcomes (climate change) in order to avoid other bad outcomes (collusion). This poses large systemic challenge indeed.

I think the dichotomy between collusion and cooperation is a good one.

I've been thinking about the Hong Kong protests a lot lately which is I think what spurred this post.

I've been trying to rationalize defection as something inevitable and not just collusion. If it can be mathematically modeled then it's easier to think about rationally.

I had always though the difference was an assertion of the speaker's values: HK protesters are cooperating to stand up for their rights vs The Phoebus cartel is colluding to the disadvantage of consumers.

But maybe there is an objective detectable difference, perhaps related to defection reluctance. It would be great if there is some additional force at play that differentiates between them.

Toward minimizing defection, the article offers these example solutions:

> What if unions gave perks to employees including job references, health benefits, etc BUT to participate in the union you had to put forth a $5k bond. This could be deducted slowly over time from your paycheck.

> ...

> Consumers could do something similar for corporations as well. Have reserved payments given out over time but only if the companies don’t create any externalities for which the consumer doesn’t approve.

There are multiple practical problems with these approaches, which might be grouped into two broad categories:

1. How do you prevent collusion? I defect and pay you to keep quiet. Or I pay you to falsely accuse another member of having defected.

2. What due process determines whether defection has occurred?

In the constrained world of a system like Bitcoin, this can be easy, but only if the outside world doesn't enter. The minute we start invoking oracles to inject external state, then the entire system begins to revolve around how to prevent the same kinds of collusion/misrepresentation behaviors in the oracles.

There's been a trend lately of low-quality academic-lite blogposts being upvoted on HN. This, unfortunately, continues the pattern. I blame Wikipedia (and reddit-style intellectual superficiality); it's a double-edged sword that makes everyone think they're an expert. The Tragedy of the Commons is extensively studied in Game Theory (we spent half a semester on these types of problems) and has been at the heart of many social experiments over the past century (Ostrom was already cited a few times in the thread). Anyone that's done more than skim the Wikpedia article knows that group size is a primary element, but it's not the whole story.

To say that the blog post brings nothing new to the table is an understatement. I really wish OP just gave a newbie-friendly introduction to the problem (which would seem more appropriate given their exposure), as opposed to proposing a "solution" -- the hubris is palpable.

I mean, take a look at this quote: "Interestingly enough, the optimum solution is always present when N=1" -- it's wrong on just about every level. First, that's not a solution (and saying optimum is incredibly telling; the term of art is equilibrium, anyway), and second, common property (even if you're the only actor) is still not your property (think about a park in a city of population 1; there's a difference between that park and your back-yard). That's the whole point of the thought experiment.

Dude... I literally in the 3rd line said this:

"I might be wrong but would love some feedback. It’s an exceedingly complicated topic which has been researched for hundreds of years so it might be that this observation has been made before."

... so there's to be NO discussion on the Internet of complicated topics. Got it!

> ... so there's to be NO discussion on the Internet of complicated topics. Got it!

I think you're attacking a straw-man here. That's not what I said at all. I also write blog posts on complicated topics, but they are thoroughly researched and generally, though not always, within my purview of expertise. You literally only read the Wikipedia page.

> You literally only read the Wikipedia page.

This is exactly what I'm talking about. You have NO idea of my background. You don't know what I've read.

Do you speak to people like this in person? Common decency goes a long way to further academic discussion.

I dunno, the post's some real n-gate bait.

> Being that my background is in CompSci I tend to think about thinks in terms of probabilities, statistics, game theory, and distributed systems.

Well. OK.

But then a bunch of:

> This would need to be validated experimentally but I suspect that the probability of cooperation (the opposite of defection) would be inversely proportional to the size of the group.

:-|

And the whole post is of about that quality. Fine as a "dear diary" I guess but it's very low-value even as an amateur effort.

This is my point, but I’m getting downvoted and buried by the mods so I’m ejecting out of this conversation.

You've crossed into personal attack, as well as breaking the site guideline that asks: "Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."

Can you please not do that? We all have a bias to dismiss others as idiots on the internet, and it's necessary to consciously discount that in order to have any chance of being fair. Otherwise you end up posting a shallow dismissal while criticizing someone else for being shallow, which improves nothing.

If you know more, that's awesome, and in that case you could write a much better comment by teaching the rest of us some of what you know, so we can learn something.

https://news.ycombinator.com/newsguidelines.html

>I think you're attacking a straw-man here. That's not what I said at all. I also write blog posts on complicated topics, but they are thoroughly researched and generally, though not always, within my purview of expertise.

Doesn't seem like he attacked a straw-man. You literally doubled down saying (to sum it up): "There's to be NO discussion on the Internet of complicated topics [unless you're an expert of them, like I'm on the topics I write on the internet about]".

I think your characterization is unfair.

> it's a double-edged sword that makes everyone think they're an expert

I did not get the impression that the author thought he was an expert. By reading, I think it’s clear the author is aware that he’s not an expert.

More generally, I don’t mind articles written by people who aren’t subject matter experts. Mainly because I think most new ideas come from outsiders.

I guess I'll have to disagree. In my book, proposing a solution to X implies being an expert on X.

Why gatekeep on who can brainstorm about solutions? If a non-expert stumbles on a great solution, you would throw it out? I don't like ANY of the author's solutions. The article still got me thinking about negative externalities in a new way.

Let me put it this way: I can't take someone "solving" the Tragedy of the Commons without mentioning Nash equilibria, Pareto efficiency, or Ostrom's empirical studies seriously any more than I can take someone "solving" the Riemann hypothesis without mentioning prime numbers seriously.

So, if someone does formulate a working solution in a novel way, but they doesn't mention Nash equilibria, Pareto efficiency, or Ostrom's empirical studies (either because they're not a necessary part of a solution, or because they've came at it on their own, without having prior knowledge of these, and thus formulated was was needed of them on their own terms), you'd don't want to hear about it, because dogma.

Does it surprise you that a place named "Hacker News" is usually welcoming to non-experts who try to find ways to hack all variety of "systems" that experts claimed wasn't possible?

In my book, outside of hard sciences and engineering, "being an expert on X" is just a BS title, almost unrelated with competency, actual experience, and a history of good outcomes working on X, and worst of all without skin in the game.

The biggest blunders and worst outcomes, have historically come from "expert solutions"...

https://www.smithsonianmag.com/smart-news/why-experts-are-al...

On the other hand, tons of creative solutions and amazing inventions came from non-experts -- people like Mendel, for example (and in fact any new domain must by necessity be founded by a non-expert, since there's no domain to be an expert of beforehand).

I'm the author. I don't claim to be an expert in this field and I explicitly mentioned this multiple times in the post.

Third sentence I had a disclaimer and my solutions at the end had a disclaimer.

Discussion is INCREDIBLY valuable for learning about new topics. Further, many major contributions have been from people with different and unusual backgrounds.

>There's been a trend lately of low-quality academic-lite blogposts being upvoted on HN. This, unfortunately, continues the pattern.

First, what's "lately"? HN has had since forever upvoted blogposts on all kinds of things, and they more often than not had simple or simplistic content. It was never about academic standards or some such.

Second, who said academic-proper (as opposed to "academic-lite") material is necessarily "high quality" or of more interest? Most papers are drab affairs, to get someone a degree, or to pad someone's publishing numbers. That they might be full of the proper references, doesn't make them necessarily insightful (or appropriate for a general audience).

And in any case, HN is not a journal, nor is the HN crowd comprised mainly of scholars...

Is "Worse is Better" (a seminal post) academic? Are PG posts (the closest to holy scriptures for early HN) academic?

>If your company cares about the environment but all your customers care about is price then the companies that have a higher priced product will go out of business.

I can picture Yvon Chouinard's middle finger slowly sticking up

https://bcorporation.net/directory/patagonia-inc

Oh look, a justification for the Berlin Wall.

Reply on news.ycombinator.com