The 2-4-6 game has two players: one takes the role of Nature and the other takes the role of Researcher.

Nature writes down a rule that a triplet of numbers may satisfy or not. (triplets such as 1-3-5, 1-2-3, 15-11-2, …)

Researcher wants to figure out what the rule is, and is allowed to ask Nature about specific triplets if they fit the rule or not. Nature answers "Yes, that fits the rule" or "No, that doesn't fit the rule".

After a few such tests, whenever Researcher feels done, she says what she thinks the rule is.

---

(DON'T READ THE BELOW if you haven't played the 2-4-6 game before; find someone who has, and let them play as Nature for you. There's a fun lesson!)

---

…

---

…

---

Trick is that Nature wrote the rule "any integers in increasing order", and Researcher will tend to guess a much more convoluted rule.

The game demonstrates the importance of seeking disconfirmations of your hypothesis (Attempt to falsify) immediately, which would narrow down the search-space quickly, but people rarely do. This is "positive bias".

Example:

• Form an idea: "Hmm, 2-4-6 matches, so maybe the rule is +2 each time?"
• Bad next step (wasted test): "Let's see if that's correct. Does 4-6-8 match?"
• Good next step: "Hmm, how do I find out that my idea is wrong as quickly as possible? Does 4-6-9 match? 3-2-1?"

Positive bias (doing only confirming tests) can be regarded as distinct from confirmation bias (trying to preserve the belief you started out with), even though many times when there's positive bias, there's confirmation bias too.

This game is an exception: no one would say that Researcher is trying to "preserve a belief" – this is just a game – so the problem is nothing so sinister, only that she uses a naive methodology.

---

One may be lectured on positive bias for days, and yet overlook it in-the-moment. Positive bias is not something we do as a matter of logic, or even as a matter of emotional attachment. The 2-4-6 task is “cold,” logical, not affectively “hot.” And yet the mistake is sub-verbal, on the level of imagery, of instinctive reactions. […] You have to learn, wordlessly, to zag instead of zig. You have to learn to flinch toward the zero, instead of away from it.

—Eliezer Yudkowsky

Another game: en.wikipedia.org/wiki/Wason_selection_task

What links here

• Human heuristics, biases and fallacies
• Confirmation bias

www.greaterwrong.com/posts/3XgYbghWruBMrPTAL/leave-a-line-of-retreat

From Sun Tzu, one translation: "When you surround an army, leave an outlet free. Do not press a desperate foe too hard."

Equally applicable for your debate partner and yourself.

For yourself: When there's an unpleasant idea and you're trying to judge whether it's true, you can ask yourself how the world would look if it were true and what you would do then. Now you have an "outlet" – a line of retreat – it's no longer illegal for the unpleasant idea to be true and you can seriously consider it.

In a debate, this Sun Tzu quote also maps to the courtesy of giving your partner a way to avoid feeling embarrassed, and that's part of How to Have Impossible Conversations.

What links here

• Rationality techniques