If I flip a coin 100 times, and get heads 70 times and tails 30 times, is the coin rigged? Not necessarily. Each heads/tails event is independent of the other, so although our expected value from tossing a fair coin 100 times would be 50/50, there is no force in the universe making it hit 50/50 instead of 51/49 or even 70/30. Actually there is a probability distribution, a bell curve, of outcomes by probability and 70/30 would just be another outcome out on the tail end of the distribution. 100/0 would be another possibility. Although the 70%/30% result would become less likely as the number of coin flips increased, it would not be impossible. Yet, certain statistical tests could prove that my 70/30 result is so unlikely to have occurred by chance that we should accept the coin is rigged. Such conclusions are based on an estimated chance of being wrong.
Suppose I was running a coin flip gambling scam. I approach someone, let them watch me flip my coin ten times, and then offer to bet my $80 against their $100 on the next flip.
When the ten flips are 5/5 or 6/4 or 4/6, the other person walks away because the risk of losing $100 against the risk of gaining $80 is a bad deal when the odds are 50/50.
However, sometimes a seemingly unlikely sequence of flips occurs, such as 8/2. When that happens, my victim watching this sequence becomes certain the coin is rigged. Then they take my bet, which is actually a 50/50 chance tilted in my favor due to the different payouts. The expected value of the bet to my victim is (80×50%)+(-100×50%)= -$10, and to me it is +$10. If I lose, I can count on my victim to play again, because it seems their narrative has been confirmed.
This example may seem silly and small-scale but the rationale applies to investing. Investors bid up the shares of companies and strategies on a winning streak, essentially shopping around for a street hustler whose coin flips seem unlikely - using computer filters and mass media to locate these lucky streaks amongst thousands of others. They develop a narrative in their heads about past performance predicting future results because of (insert reason here). They pluck down their money until mean reversion occurs yet again. Then it's on to the next narrative.