Bayesian Prior

mechanism

New evidence should move belief in proportion to both its strength and what was already known. A weak claim should not erase a strong prior built from established science and repeated experience.

Rohit Sharma losing twelve tosses in a row sounds improbable enough to make an extraordinary explanation feel newly credible. Yet the streak itself does not make the Baba’s claim likely: the science and experience already bearing on that claim still count.

Evidence does not begin from zero

A Bayesian prior is your starting estimate before the latest observation arrives. That estimate may rest on established mechanisms, accumulated experience, and many earlier observations. New evidence then updates it; it does not automatically replace it. The size of the update depends on how diagnostic and trustworthy the evidence is. A surprising streak can deserve attention without carrying enough weight to overturn everything already known. This is the starting point that Bayesian Inference combines with fresh evidence.

Where it shows up

Twelve lost tosses

The streak is real and striking, but it is weak evidence for the Baba’s explanation. Its emotional force is larger than its power to displace the prior supplied by science and experience.

A prior is ballast, not an anchor

A strong prior can protect you from overreacting to noise, but treating it as untouchable would defeat the model. Reliable, sufficiently strong evidence must still be allowed to change your mind; the prior determines how much evidence is needed, not what conclusion is permanently permitted.

Price the evidence before updating

When an astonishing event is offered as proof of an extraordinary explanation, write down two things: what established knowledge made you believe beforehand, and how likely this exact observation would be if the explanation were false. Update only after answering both.

Episodes that teach this

Science Explains Rohit Sharma Losing 12 tosses in a Row? - Future IQ 5,873 views
The episode says existing science and experience make the Baba's claim unlikely, calling this "a prior in Bayesian inferencing."