Trend-following Effect in Stocks

The three-gate gauntlet · genuine only if it clears all three and survives adversarial refutation

Gate 1

Survivorship-free

free

clean panel

Gate 2

Placebo ≥ P95

P96.5

outranked ~193 of 200 baskets

Gate 3

Cost-aware net

RF +40.51

positive, not certified

Failedfailed refutation

Worst 12-month leg (RF)-0.99

−1.00 floor0

Every strategy here — winners included — loses in its worst 12 months. Depth is honest context, not the verdict.

Cleared the numeric checkpoints it reached but failed adversarial refutation — the positive figure is market beta or a single-period jackpot, not a repeatable edge.

This approach applies classic trend-following to individual equities: enter on an all-time-high breakout and ride the position with a chandelier (ATR-based) trailing stop, one position per stock.

What we found

The headline profit figure for this rule is an artifact of fixed-dollar-per-position sizing — it is a gross-notional sum across roughly 8,311 trades, not a capital return, so we do not report it as a return here. On a like-for-like basis, passive buy-and-hold beats the rule about ninefold over the same window, and the survivorship-free selection placebo sits only at the 96.5th percentile. In other words, the result is dominated by market beta and entry timing rather than any durable stock-selection skill, and the worst-year risk-adjusted result is a loss. We reject it.

How we tested it

2005–2026 test windowmodelled liquidity-aware costssurvivorship free

Data: survivorship-free 1077-name US common-stock panel, 2005-2026. Realistic modelled costs.
Placebo / robustness test: real result vs random baskets or shuffled signals (real vs the 95th percentile of random)

Source: Wilcox & Crittenden (2005), "Does Trend Following Work on Stocks?", Blackstar Funds
Read the paper ↗

← The Academic Audit — all 54 studies

Research, not investment advice. “Validated” factor-legs are market-neutral diversifying building blocks with a losing worst year — none is a standalone tradeable strategy. Metrics are cost-aware and modelled (not live fills); the 2005–2026 test window is out-of-sample versus the source paper. Dollar figures are not returns and are omitted by design.