Module VII: Psychology and reality
35. Cognitive biases
Gambler’s fallacy and hot hand: misreading independence
The gambler’s fallacy is the "it has to come now" feeling: after five losing tips the sixth is "due". The reality: independent events owe you nothing, the coin does not remember. The binomial maths of chapter 20 stated exactly this with a formula, yet your brain sees a pattern in the noise, because it is wired for it.
The hot hand is the same error in the other direction: "I’m in form, now I should raise the stake". A winning streak is part of the normal deviation band (chapter 20), and the stake increase comes exactly when the band is statistically about to pull back. The two biases together are the bettor’s idle mode: on a loss you see "due-ness", on a win "form", and both produce a bad staking decision.
The antidote: the streak never decides the stake size, the system does (chapter 24). If the run influences your stake, one of the biases is working on you right now.
Confirmation bias, hindsight bias, anchoring
Confirmation bias is the researching bettor’s main enemy: after you pick the tip, your brain automatically collects the confirming information and dismisses the refuting. That is why the order of chapter 11 is set in stone: first your own numerical estimate, only then the odds and the arguments. Whoever starts from the odds has an "analysis" that is anchored: it is built around the first number seen (this is called anchoring bias).
Hindsight bias is the retroactive "I knew it": the match played out, and your brain rewrites the past as if the result had been foreseeable. This is what makes learning hard: if you always "knew it", you never learn. The remedy is prior written fixing: you write down the reasoning of your tips BEFORE betting (chapter 45), and afterward you measure against that, not against your memories.
Know the market-level biases too: the favourite-longshot bias (chapter 4) partly lives on the fact that the crowd systematically overpays for dream multipliers. Your biases are costly not only to you: they are priced in.
Resulting: a good decision is not the same as a good outcome
Resulting is judging the quality of a decision by the outcome: the bet that came in was a "good decision", the one that lost a "mistake". After Module IV you already know why this is false: in the short run the outcome is a lottery. The bet played at 2.20, closing at 2.00 and lost, was a good decision with a bad outcome; the tip played at 1.80 and won was a bad decision with a good outcome, and the second is the more dangerous, because it reinforces a bad process.
The instrument against resulting is ready: CLV (chapter 12) measures the quality of the decision independently of the result, and the journal’s luck-skill panel (chapter 43) breaks your whole result into a decision part and a luck part. After a bet the question is never "did it come in", but "did the decision come from a good price, a good process".
The summary rule of Module VII: your feelings and your memories are bad instruments, the written prior reasoning, the CLV and the calibration curve are good ones. Decide from an instrument, not from an echo.