Research Lab

We publish the strategies that failed, too

Most trading research shows you the winner. These reports show every test that was run, every strategy that died, and the bugs that made results look better than they were — because a backtest you cannot check is not evidence.

Educational research, not investment advice. Every figure in these reports is a hypothetical backtest — simulated results from historical data, not money that was ever traded. Backtested performance is not an indicator of future results and carries inherent limitations, including hindsight in a strategy’s design. Nothing here is a recommendation to buy or sell any security. Mezan does not sell trading signals.

Report 03 · September 2026

The QMAC ladder — a scan we tested and failed

A moving-average setup from Lee & Tryde (2012), run over 21 years. It did not beat the market after costs, so we publish no signals from it. It did cut the worst drawdown of a trend book from -51.3% to -30.9% at every cost level tested, which is the only reason it is here.

Report 02 · August 2026

Every strategy tested, and what actually survived

Sixteen strategies run against nine years of daily bars on 8,227 US symbols — including 4,930 companies that no longer exist. Fifteen failed. The full log of all 13,160 positions is published so the result can be checked rather than trusted.

Report 01 · July 2026

Backtest & Monte Carlo — twelve classic setups

Twelve textbook chart patterns — Turtle Soup, Holy Grail, ID/NR4, 80-20 and others — backtested on nine years of daily and intraday data, each with 5,000 Monte Carlo simulations showing the range of outcomes luck alone can produce. Run on 69 large-cap symbols selected in 2026, so unlike Report 02 it is not corrected for survivorship — read it as illustrative, not as evidence.

Both reports open as standalone pages. They are static and self-contained — search a ticker, sort any column, and check the numbers yourself.