Korva is a scroll of the internet’s best ideas.
Everything can be quantified
The objection
The usual objection is that some things can't be quantified: the chance of meeting the love of your life tomorrow, of still being with a partner in ten years, or of liking the dish ordered at a restaurant. They can, because everything is probabilistic. Nobody has access to an objective probability for any of them, and there isn't enough data to reach the certainty of a coin's 50 percent, but a range can still be estimated.
Wide ranges
Suppose the best estimate of whether a belief is true is somewhere between 20 and 80 percent. That is a very wide market, but it is narrower than 0 to 100, and trying counts as a success. A range also gives a reason to refine it: asking people, researching, looking things up. Each search narrows the range a little.
Without the habit
Without that habit, the future gets treated as guaranteed, and bad outcomes arrive as a surprise. People overreact to them, change strategies they shouldn't, reinforce strategies they shouldn't, and spend their time reacting instead of planning ahead. Life isn't chess, where playing better than the other person wins.
How do forecasters score their probability guesses?
Meteorologist Glenn Brier devised a scoring rule in 1950 to measure how good weather forecasts were. It compares each stated probability with what actually happened, rewarding forecasts that were both confident and right. The Good Judgment Project, led by Philip Tetlock and Barbara Mellers, used scores of this kind in a government-sponsored forecasting tournament that ran from 2011 to 2015. Forecasters in the project updated their numbers often as new information arrived.
Do phrases like serious possibility mean the same thing?
Sherman Kent, a CIA analyst, wrote about the problem in his 1964 essay Words of Estimative Probability. A 1951 intelligence estimate had called a Soviet attack on Yugoslavia a serious possibility, and the analysts who read it understood that phrase to mean odds from about one in five to four in five. Kent proposed assigning numerical ranges to such phrases. Intelligence agencies later adopted standard tables of terms and odds.
Why is poker a better model of life than chess?
Chess is a game of perfect information with no element of chance, so both players see the whole board. Poker has hidden cards and luck, and mathematician John von Neumann used it as a model when developing game theory. He and the economist Oskar Morgenstern published Theory of Games and Economic Behavior in 1944. Von Neumann's analysis of bluffing in poker was one of his early examples.
More from Talks at Google
