What momentum really means (and when it breaks)

"Momentum" is one of the most overused words in trading and one of the most rigorously documented effects in finance. They're not the same thing. Here's what momentum actually means, the evidence behind it, and - just as important - the conditions where it stops working.

Two kinds of momentum

• Time-series (TS) momentum- an asset's own past return predicts its near-future return. If it's been trending up, it tends to keep trending up. It's about each asset versus its own history.
• Cross-sectional (XS) momentum- rank a whole universe by recent return, buy the strongest, and (in long-short form) sell the weakest. It's about each asset relative to its peers. MasterBitcoin is primarily an XS, long-only system.

The evidence

The effect was formally documented by Jegadeesh and Titman (1993), who showed that stocks with the highest returns over the past 3–12 months continued to outperform over the following months. It has since been replicated across equities, bonds, commodities, currencies - and, more recently, crypto. Han, Kang & Ryu (2024) specifically validated a long-only cross-sectional momentum edge in cryptocurrency markets, even after accounting for funding costs.

Why does it persist? The leading explanations are behavioural: investors under-react to new information at first (the trend builds), then over-react later (the trend over-extends). Momentum tries to ride the middle of that cycle.

When momentum breaks

Momentum is not free money - it has a specific, ugly failure mode. Daniel & Moskowitz (2016) documented "momentum crashes": the strategy behaves like a short option. It earns small, steady gains in calm trends, then suffers sharp, violent losses during volatility spikes and sudden market reversals - exactly when prior losers rocket back fastest.

This is why disciplined momentum systems gate for conditions rather than always being on:

• Regime gates suppress entries when the broad market (e.g. BTC) is breaking down.
• Dispersion gates require genuine separation between winners and losers - momentum needs something to rank.
• Blow-off gates avoid chasing parabolic moves that are statistically prone to snapping back.

The honest summary

Momentum is a real, durable, well-evidenced edge - and a fragile one if you ignore the regimes where it crashes. The work isn't inventing a secret indicator; it's implementing a known effect with the discipline to stand aside when the conditions turn against it. That discipline - staying selective, even when it means an empty board - is the difference between harvesting momentum and getting run over by it.

Go deeper in the cross-sectional momentum methodology and market regimes.