[[Start here]] → [[expose to return drivers|drivers]] → [[strong stocks beat weak stocks|momentum]] → short term reversals --- There’s a trader’s rule that goes something like “if it moves too far, too fast, then fade the trade”. The data seems to back this up. Over short timeframes (a week or a month), share prices have a tendency to swing too far in response to flows or market sentiment - and then snap right back in the opposite direction. This effect is not due to news or fundamentals - it’s because of a lack of liquidity or due to noise. When too many investors rush to buy or sell at once, market makers adjust prices aggressively in order to clear their books. So stocks can overshoot. Anyway, back in 1990, **Lehmann**[^1] showed that buying the week’s biggest losers and shorting the week’s biggest winners earned a profit roughly 90% of the time. While **Jegadeesh**[^2] confirmed that these short-term reversals can last for about one to two months before the usual momentum effects take over. This pattern is one of the most robust and most arbitraged in finance. > [!INFO] **NB**: Sometimes short term moves do continue… > When there are price jumps due to **new information** (e.g. when there is an earnings surprise) the share price does tend to drift in the same direction - this is classic “post earnings announcement drift”. So while **medium-term momentum** (3–12 months) works because prices underreact to genuine information, **short-term reversal** is the mirror image - it’s driven by an **overreaction to noise**. They’re really two sides of the same behavioural phenomenon. Even intraday, you can see the same reflex. We found that [[profit warnings bounce intraday proportionally to their decline]] - a classic whiplash effect. **** ## **Rules of thumb** - **Don’t chase fast moves.** When a stock surges on no fresh information, be patient - liquidity spikes often retrace. These can be good sell points. - **Don’t panic-sell into capitulation.** After sharp drops, markets often rebound once short-term sellers are done. These are good counter-trend buy points. - **Know your jump type.** If the move follows a _real information event_ (earnings surprise, analyst revision), that’s PEAD territory - and those trends tend to _persist_, not reverse. ### Related notes * [[big jumps on earnings announcements predict future drift]] * [[momentum continues for three to twelve months]] * [[momentum reverses after three to five years]] ## **Sources** [^1]: Lehmann, B.N. (1990). _Fads, Martingales, and Market Efficiency_. _Quarterly Journal of Economics._ [^2]: Jegadeesh, N. (1990). _Evidence of Predictable Behavior of Security Returns_. _Journal of Finance._ [^3]: Chordia, T., Subrahmanyam, A., & Anshuman, V. (2005). _Liquidity and Asset Returns._ _Journal of Financial Economics._