Prisoner’s Dilemma and New Types of Nuclear Reactors

By Meredith Angwin

viewfromVermontPrisoner’s Dilemma is a famous example of game theory. You can look at this example in quite a few ways. I especially think of this game when I am thinking about Liquid Fluoride Thorium Reactors (LFTRs) and advanced and non-traditional reactor development.

But before I begin looking at LFTRs, let’s see the classic example of the game.

Prisoner’s Dilemma, summarized

cartoon prisoner 90x153Two men are arrested. They are held separately and cannot communicate with each other. Prisoner A is told that if he gives evidence against Prisoner B, while Prisoner B does not give evidence against Prisoner A, Prisoner A will go scot-free and Prisoner B will spend three years in prison (and vice-versa). If A gives evidence against B, but B also gives evidence against A, they will both spend two years in prison. If neither A nor B gives evidence against the other, they will both spend only one year in prison.

In other words: If you give evidence against the other guy, you get a good deal, unless the other guy gives evidence against you.

The best deal overall is for nobody to give evidence against the other person: this leads to 1 year in prison for both people. However, this decision puts the individual in a difficult position. He fears that if he does not give evidence, but the other person does, then he will get three years in prison for having kept silent. The best outcome is for both people NOT to give evidence, but the most probable outcome is that most people WILL give evidence.

The most probable outcome is two years in prison for both.

Viewpoints on the game

This game can be looked at as simple strategy: Here are the rules, and therefore these results will be the consequences of those rules.

The game can also be looked at as a moral issue: Taking the high road (not attacking the other guy) leads to the best overall outcome.

Or a community issue: How can you persuade people to do what is best for the group, instead of what is best for themselves (and not-best for the group)?

Or a truth and falsehood issue: As a friend pointed out to me recently, the people judging the prisoners sound like very crooked people. There is nothing in this game to say that the evidence given is true or false, or that the prisoners are guilty or not guilty. If one person “gives evidence,” it convicts the other person. End of story.

xkcd valentine dilemma 132x140Or a communications issue: If the two prisoners could talk to each other, they could mutually decide to not-attack the other and get the best overall outcome. Most versions of Prisoners Dilemma make it clear that if the players are able to talk to each other, it is fundamentally not the same game.

Okay. Let’s get back to communicating about nuclear energy.

LFTRs and the Dilemma

I am happy to know several proponents of LFTR development. Chief among these people is Dr. Robert Hargraves, who wrote the book THORIUM, Energy Cheaper Than Coal, and has spoken on the subject worldwide.

cover thorium energy cheaper than coal 100x149I am also a strong proponent of LFTR development. I presented LFTRs years ago in a poster session at a high temperature chemistry Gordon Conference.

Hargraves finds it straightforward to be in favor of continued operation of currently operating LWR like the Vermont Yankee nuclear plant AND in favor of rapid development of LFTRs. I feel the same.

With the people I know, there is no dilemma, because there is no real contradiction between two goals:

  1. Keeping current reactors contributing to our electricity supply.
  2. Designing new types of reactors for the future.

But there is a dilemma

Not everyone is mellow all the time.

The people who want to see new types of reactors funded sometimes attack current reactors vigorously: You NEED our reactors for safety and non-proliferation!

People who work on current reactors sometimes attack new types of reactors as impractical: Your idea is just one step beyond a PAPER REACTOR and will be full of new problems when you scale it up!

To some extent, we are having the classic Prisoner’s Dilemma outcome: Both sides are attacking. The attacks lead to a worse result for everyone.

Communication as the answer

Supposedly, the fact that both sides can communicate with each other should change the game. The game shouldn’t even be Prisoner’s Dilemma anymore. We can learn by communicating.

For example, attacks on current reactors will translate easily into attacks on new types of reactors. Therefore, supporters of new types of reactors should not join the nuclear opponent attacks on operating reactors.

Similarly, no one should dismiss other reactor designs so vigorously that the implication becomes that operating reactors are the best reactors that could ever be designed or built. Engineering advances have improved cars, houses, television sets, and computers. Engineering advances will improve reactors.

In other words, both sides will win if we promote nuclear power as a whole.

Yet, like the prisoners who can’t communicate, we are afraid of the downside. We fear that new reactor types won’t get funding, or we are afraid that existing reactors will be shut down. Sometimes we can’t see that attacks on the “other” nuclear technology hurt the whole community of nuclear supporters.

But there is a way to win. The way to “beat” Prisoners Dilemma, the acknowledged way to change the game, is to communicate.

Let’s start communicating.


Endnote: I wrote about communication, but I think that the other issues I listed (moral, community, truth and falsehood) also come into play. I welcome your comments. Is the nuclear industry playing “Prisoners Dilemma”? And if we are playing, how do we stop?

I also recommend Rod Adams’ post Open letter to advocates of Generation IV reactors (IFR, LFTR, NGNP, PBHTR) in which he reminds us that the real battle is “fission versus fire”, not “types of fission.” The post also has a lengthy and very interesting comment stream.




Meredith Angwin is the founder of Carnot Communications, which helps firms to communicate technical matters.  She specialized in mineral chemistry as a graduate student at the University of Chicago.  Later, she became a project manager in the geothermal group at the Electric Power Research Institute (EPRI).  Then she moved to nuclear energy, becoming a project manager in the EPRI nuclear division.   She is an inventor on several patents. 

Angwin formerly served as a commissioner in Hartford Energy Commission, Hartford, Vt.  Angwin is a long-time member of the American Nuclear Society and coordinator of the Energy Education Project.  She is a frequent contributor to the ANS Nuclear Cafe.

8 thoughts on “Prisoner’s Dilemma and New Types of Nuclear Reactors

  1. Meredith Angwin

    I would like to thank you all for your comments, and especially thank Greg London. London is the friend who described the moral dimension of Prisoners Dilemma. When I wrote this blog post, he was still working on his analysis of the game. I am very happy to see his comment and his link on this blog!

    I have been thinking about Caplan’s posts and Schmidt’s post. Caplan is correct: current reactors are very safe. However, Schmidt points out that they do have failure mechanisms that the newer reactors can avoid.

    I think that it is always a bit of a balance when we talk about improving technologies. LFTRs still need significant development work and they need an NRC license, which may take decades to get. So, in this country, if we need to build the intermediate reactors that are currently approved. Or we won’t build reactors at all, and we will have mostly gas plants. The current fleet has done very well, world-wide, and the new reactors will also do well.

    Also, I think a little about disasters. The worst type of disasters come on very quickly: a train wreck and fire, a dam bursting, a chemical plant explosion. In contrast, it would take a few days before anything would happen at a nuclear plant, and people would know that it was going to happen. This may not seem like much, but it can seem pretty good to a first responder. That first responder isn’t even going to HEAR about a train wreck that releases hazardous material until AFTER a lot of people have already been exposed to fire and/or toxic materials.

    Today’s nuclear is safe, as Caplan said. And we can make it safer.

  2. Robert Margolis

    I have always thought that there is room for different reactor types as there is room in aviation for both jets and propeller planes. Yes we should have discussions on the different reactor types, but when we lose our perspective we get too busy fighting AGAINST ourselves instead of standing together FOR a better nuclear world.

  3. Milton Caplan


    So true. It’s about demonstrating the benefits of nuclear versus other forms of generation.

    We must not fall into the trap of always talking about safety. Current nuclear plants are safe – in fact – very very very safe. Period. Just yesterday there was a horrible train derailment in Spain killing close to 80. Or the recent unbelievable disaster in Lake Magantec Quebec where a train load of oil exploded effectively destroying the town and killing close to 50. Do we insist that the only answer is to develop a whole new generation of trains because the current generation is not safe enough? (Although they could use a lesson from the nuclear industry on safety culture).

    It is great to develop new reactors for the future but we need to focus on other benefits than just safety. We need to show the benefits of a new generation of reactor in term of cost, fuel utilization and other good stuff. But I have no time for those saying we need to move forward to be safer……we are not doing the industry any good with these kinds of arguments as it supports those who say the current fleet is not safe.

    So one more time – the current fleet of nuclear plants operating around the world is very very safe!

  4. Greg London

    One of the biggest lessons from Prisoner’s Dilemma is that sometimes operating from “What’s in it for me” can end up with everyone worse off than if they came from “What’s best for everyone”.

    In 1993, Switzerland asked one of its small mountain villages to support putting a nuclear waste repository nearby. About 51% of the people supported it. Trying to increase support, Switzerland offered all residents a small cash payment if they would accept. Only about 26% of the residents supported the plan after that.

    The government had invoked a moral decision for the residents and residents chose from a “What’s best for everyone” calculus. Everyone would be better off if someone accepted having a nuclear waster repository in their backyard. When the government added a monetary incentive to increase support, they unwittingly changed the decision making calculus of the residents from “what’s best for everyone” to “what’s best for me”. And suddenly, having a repository in your backyard didn’t seem worth the small amount of money they were offering.

    The calculus of “what’s best for everyone” was lost, and residents chose from “what’s best for me”. But at the same time, by refusing a repository, the entire nation lost out, and everyone was worse off. It seems counterintuitive to many, but sometimes “what’s best for me” ends up choosing what’s worse for everyone, including the person making the choice. This is the difficult lesson of the Prisoner’s Dilemma.

    One of the things I think that makes the Prisoner’s Dilemma so difficult for some to understand is that the narrative is inherently an evil or immoral narrative. Dirty cops are trying to put innocent people in jail by rigging the system. So a lot of people inherently want to resist this immoral act, and therefore they see the moral choice (cooperate with the other prisoner) as the only choice. They don’t believe anyone in their right mind would make the economic decision (betray the other prisoner).

    I’ve come up with a narrative that creates the same payoff matrix, the same decision conundrum, as the Prisoner’s Dilemma, but does so in a morally-neutral way so that the economic decision is more accessible to those having a hard time seeing anything but the moral decision of the Prisoner’s Dilemma. Folks can check it out here:

  5. Jeff S

    I think we should keep current reactors running, but I’m not exactly too excited by a “nuclear renaissance” of building more LWRs. I try not to attack them too much, but, a lot of the problems the public associates with “nuclear”, are really “LWR” problems. “New” LWR designs like the AP-1000, and I think the B&W mPower, don’t fully solve some of those problems. For example, the AP-1000 – ok, it has 3 days of non-powered emergency cooling. Great, that’s clearly better than a few hours of battery powered cooling which might fail if the batteries or electrical room get flooded.

    But, I can’t help but think about the AP-1000 . . . and what if they are unable to add more water for over 3 days? That’s not a huge amount of time. At Fukushima, an AP-1000 might have still failed, just because all the roads were destroyed, and basically nobody could move around the Fukushima area very well.

    Also, there’s still water in contact with ZircAlloy, so there is still fundamentally a possibility that free hydrogen can be generated, and explode.

    So, I look at something like LFTR, and that has NO HYDROGEN or any other potentially explosive elements or compounds anywhere near the reactor. Just fuel salts, metal piping (which MSRE showed didn’t react with the fuel salts), graphite moderator (Ok, there is some possibility that could hypothetically burn, but in an emergency, the hot fuel salts flush down into holding tanks, so there’s nothing to get hot enough to set the graphite on fire, and graphite does not easily burn, and will not explode in any circumstance).

    It’s all at almost atmospheric pressure, so there’s not a pressure-explosion risk.

    I feel like the public needs to know that LFTR removes essentially all failure modes which are likely to release radioisotopes into the environment. That’s a big deal. That is worth communicating. There is no way that communicating the benefits of one technology over another, won’t come off as an attack on the technology which comes out less favorable in the comparison.

  6. Joe Carson

    How does engineering ethics relate to one’s reasons/words/actions in this “dilemma”? My perspective is that there are significant and persistent deficiencies in scope and implementation of engineering ethics which significantly – and negatively -impact nuclear power safety and the public trust in nuclear professionals.

  7. donb

    The world needs large quantities of clean, inexpensive energy. So there is plenty of room for many types of nuclear reactors. Look at the world of “fire energy”. There are many ways of turning fossil fuels into electrical or mechanical energy — steam plants, turbines, reciprocating engines using gasoline or diesel fuel, etc. The world of nuclear energy is presently dominated by U235 fuel making steam. The reactor variants are pretty much light water types using either boiling water or pressurized liquid water, and the heavy water CANDU. It is as if the world of fossil fuel energy used only 2-cycle and 4-cycle piston engines, and the Wankel engine, but no diesels, combustion turbines, or steam plants.
    As new reactor types are developed, the present reactor types may fade away, much like reciprocating steam engines have largely faded away. Those steam engines didn’t fade away because they were bad or dangerous, but because the replacements were better. Indeed, the reciprocating steam engine was a major technology that made its replacements possible, just as today’s reactors are a major technology making tomorrow’s reactors possible.

  8. James Greenidge

    You’re so true! What supporters of Thorium and Fusion and molten salt and likewise non-conventional species of nuclear plants have to get a grip on is that unless the public becomes tolerant if not comfortable with _today’s_ plants by their unashamedly admirable record in normal operation AND accidents, their own pet nukes are just the same old nuclear promises-promises vaporware rose smelling under a different name to most the non-discriminating under-nuclear educated public. It behooves them greatly to support and educate public on the function and record of our current crop of real-life tangible existing nukes if they want to feather their nest for their more advanced stuff. Neither Indian Point or the latest Georgia reactors are going to switch to Th or salt anytime soon so best sow your wheat by getting on board the fight against FUD right now. Shunning or keeping mum about current plants only works against your own hopes and dreams and does the antis work. Human dynamics isn’t as logical and rational as engineering issues are.

    James Greenidge
    Queens NY

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>