Unit 3 · Level 4 · Factors & beyond the index
The factor zoo
Researchers have published hundreds of market-beating 'factors'; economists call it the factor zoo. The trick is that if you test enough random patterns against past prices, some will fit by pure luck. Studies that re-checked published factors found many don't replicate at all, and premiums tend to shrink sharply once a paper is published and money piles in. A backtest is a rearview mirror that's easy to polish.
Free to play. No ads, no token, no account needed to start.
What you get asked
What is 'data mining' in the investment-research sense?
Torture the data long enough and it confesses to anything. With hundreds of researchers testing thousands of patterns, 'significant' results appear by chance alone.
Order the life cycle of a mirage factor
Research found factor returns drop substantially after publication: partly luck evaporating, partly crowds arbitraging away whatever was real.
Testing hundreds of patterns until one fits the past is called data ___ .
The polite academic term for finding faces in clouds. The market's past is a big cloud; it contains any shape you go looking for.
Which factor deserves the MOST benefit of the doubt?
Luck rarely repeats across unrelated markets and eras. Breadth of evidence plus a sensible 'why' is the closest thing factor research has to a truth filter.
An ETF advert shows a strategy that 'would have returned 19%/yr since 1990'. Your trained reaction?
Backtests don't pay pensions; only future returns do. If a strategy truly printed money, it would be quietly run, not loudly sold. 🐜
The rest of this unit
Value, momentum, the factor zoo, and the marketing machine built on top.