The Prisoner’s Dilemma and Kobayashi Maru

An Octopus

At group today, I found a way to Kobayashi Maru the prisoner’s dilemma.

The prisoner’s dilemma is an example in game theory that has no obvious solution. John and Frank were arrested for an alleged robbery, in our example, a warehouse. The idea is that the two prisoner’s are separated and interrogated independently by the police. The police say that if you confess and implicate the other person, you’ll only get 1 year in prison and the other will get 20, but if both of you confess, you’ll both get 10 years.

The prisoner’s dilemma that pits loyalty against self-preservation. There isn’t any clear way to behave. It’s a test of character and the relationship between two people.

Kobayashi Maru is a famous incident in Star Trek history. It’s a simulation of a battle with the Klingons. Basically, the captain can’t win and will end up dying along with the rest of his crew.

Kirk, famously, cheated and reprogrammed the simulator to let him win.

So me and my partner come up with a backstory to the alleged crime. The two of us had been friends since 7th grade. When were were 20, we went to prison because of a automobile theft. Now the pair was 35. About a month before the arrest and warehouse incident, we’d had a falling out because of a mutual love interest.

So, my out-of-the-box solution: I confessed that I had done it, but the other person was innocent. He had just been in the wrong place at the wrong time and it was actually a third person who was my co-conspirator in the robbery. In my mind when we played out the game, I was telling the truth and not just making up a story for the friend. But, regardless, I think it would blow out of the water any deal that the police had proposed.

In the experiment that we did today, the other person confessed as well, so according to the rules of the dilemma, we would both get 10 years in prison. The group leader didn’t reveal my actual note….

In summary, I had used my loyalty to my friend to help him out. It threw the whole situation in the air so that a prosecutor be required to sort it all out and the police wouldn’t have the tidy mutual confession.

It certainly would not be good for me in my solution. I used my loyalty to my friend to, potentially, put my life at risk. If the third person was vengeful, he could easily arrange for me to be killed.

So, I broke the dilemma, just as Kirk had broke the Kobayashi Maru simulation.

The Anti-Placebo Effect

There’s been plenty of information about the placebo effect the past few days. Some researchers are investigating the genetics of the placebo effect. Even the idea of drug-placebo interactions are on the table for discussion. I’ve seen the suggestion that it makes sense to have an additional treatment track in a drug trial: the drug, the placebo and a new “no treatment at all” track.

xkcd today has this amusing take on evaluating the placebo effect.

They work even better if you take them with our experimental placebo booster, which I keep in the same bottle.

xkcd May, 19, 2015

I had some recent experiences that might hint that there is an anti-placebo effect. With the placebo effect, the mind’s expectations modulate the properties of a pill. So, it seems to follow that normally effective drugs might have negative properties that are modulated by the mind.

This is different from the placebo/anti-placebo are imaginary or “all in your head.” A placebo is effective because there are biological consequences of the act of taking a medicine beyond the chemical properties of the drug.

I was talking to a nurse from my cardiologist’s office about headaches. I mentioned the headaches earlier to her earlier and she wondered why I hadn’t yet talked to my primary doctor about them. Almost immediately, my headache became worse until the phone call receded from my attention. A few days later, my doctor prescribed a medicine. Its effect of increased pain is similar to what happened after the conversation with the nurse. My symptoms get worse when I take the prescription medicine. An anti-placebo effect could even increase the likelihood of the side effects of a medicine developing.

So, with my headaches, I’m finding taking a small dose of Tylenol is more effective than the prescription medicine. I don’t think the dose I’m taking is enough to really have any direct effect and part of its effectiveness is a placebo effect.

How I discuss that with the doctor escapes me. “Hi doctor, I’m taking this tiny dose of acetaminophen and it working a lot better than the prescription medicine you ordered. I think part of its effectiveness is that it is having a placebo effect on my headaches.” I can only imagine that conversation going badly. “Err, we should try a different medicine that doesn’t have the same effects as the medicine I’m avoiding. Don’t take the Tylenol.” How do you talk to a doctor about placebos being effective in my case?

I’ve read that doctors think of the placebo effect as “cheating” or somehow improper. An anti-placebo for a specific individual might explain the appearance of side effects of a medicine.

Would an anti-side effect placebo be able to prevent the anti-placebo side effects of a regular drug? Perhaps an anti-side-effect placebo can be a good idea in patients that respond well to placebos.

It seems plausible.