Antiprediction

Ruby
Eliezer Yudkowsky (+487/-826) moved the concrete example higher; eliminated some discussion that, while well-linked, didn't seem to directly address the central idea
Grognor (+83/-76) This page was left with a completely wrong opening sentence for so long.
Vladimir_Nesov
Vladimir_Nesov (+15) /* See also */
Vladimir_Nesov (+54/-40) /* Related concepts */
Zack_M_Davis (-21) byline removal
Vladimir_Nesov (+1227/-630) maxent is just methodology; antiprediction is still a claim "it won't happen!".
Zack_M_Davis (+26/-26) TechExpl and authorship
Eliezer Yudkowsky (+45/-34) related to privileging the hypothesis

An antiprediction is a statement of confidence in an event that sounds startling, but actually isn't far from a maxentropy prior. Antiprediction fightsFor example, if someone thinks that our state of knowledge implies strong ignorance about the tendencyspeed of peoplesome process X on a logarithmic scale from nanoseconds to clutchcenturies, they may make the startling-sounding statement that X is very unlikely to the last possibility, even where it's completely implausible that a certain event will happen,take 'one to a point where thinking about it becomes counterproductive or otherwise distorts the decision-making process. Just as prediction simplifies reasoning by allowing to assume that an event will occur and only planning for the case where it does occur, antiprediction simplifies the reasoning by allowing to completely disregard the case where the event occurs.

The reasons for keeping a hypothesis salient despite its unlikelihood are many; it may be emotional attachment, availability created either by experience or problem framing, or any other cognitive biasthree years'.

InstrumentallyPrivileging the hypothesis, it's better is our term for what happens when somebody clings to believea small area of the possibility-space despite the lack of any evidence distinguishing in favor of that winning is impossible, than that it's likely, if the actual probability is very low.possibility versus many, many other equally likely alternatives.

An antiprediction is a statement of confidence in an event that notsounds happening, as contrasted bystartling, but actually isn't far from a maxentropy predictionprior, telling that an event will happen.. Antiprediction fights the tendency of people to clutch to the last possibility, even where it's completely implausible that a certain event will happen, to a point where thinking about it becomes counterproductive or otherwise distorts the decision-making process. Just as prediction simplifies reasoning by allowing to assume that an event will occur and only planning for the case where it does occur, antiprediction simplifies the reasoning by allowing to completely disregard the case where the event occurs.

An antiprediction is a statement of confidence in an event not happening rather than happening, as contrasted by prediction, telling that an event will happen. Antiprediction fights the tendency of people to clutch to the last possibility, even where it's completely implausible that a certain event will happen, to a point where thinking about it becomes counterproductiveor otherwise distorts the decision-making process. Just as prediction simplifies reasoning by allowing to assume that an event will occur and only planning for the case where it does occur, antiprediction simplifies the reasoning by allowing to completely disregard the case where the event occurs.

The reasons for keeping a very uncertain hypothesis of near-salient despite its unlikelihood are many; it may be emotional attachment, availability created either by experience or problem framing, or any other cognitive bias.

Often, it is sufficient to see the privileged possibility as but one among many equivalently probable events, and use maximum entropy that nevertheless seems startling becauseprinciple to divide cognitive biasesprobability make us pay attention to certain outcomes way out of proportion to their actual probability. The confident proclamation that you will not will the lottery may seem to theequally among them, leaving each very naïve like a bold prediction that concentrates probability mass, but in fact this antiprediction is derived simply from the observation thatlittle. For example, each of the millions of possible lottery combinations are equally likely, and so any one person's ticket is very unlikely to come up. The flaw inup: winning the examplelottery should be seen as one possibility among millions, not one of the lottery may be particularly easytwo possibilities of "winning" and "not winning". Instrumentally, it's better to spot, butbelieve that winning is impossible, than that it's likely, if the same fallacy may emerge in different, harder-to-detect guises, whenever the availability heuristic makes some a priori unlikely outcomes unusually salient.actual probability is very low.

An antiprediction is a statement of confidence in an event not happening rather than happening, or a very uncertain hypothesis of near-maximum entropy that nevertheless seems startling because cognitive biases make us pay attention to certain outcomes way out of proportion to their actual probability. The confident proclamation that you will not will the lottery may seem to the very naïve like a bold prediction that concentrates probability mass,mass, but in fact this antiprediction is derived simply from the observation that each of the millions of possible lottery combinations are equally likely, and any one person's ticket is very unlikely to come up. The flaw in the example of the lottery may be particularly easy to spot, but the same fallacy may emerge in different, harder-to-detect guises, whenever the availability heuristic makes some a priori unlikely outcomes unusually salient.

By Eliezer Yudkowsky:
Load More (10/26)