VCP Backtest — Trailing Stop

Methodology: walk-forward simulation using the same vcp_detect() logic (tightened, volatility-adjusted, 8%-capped stop) the live screener uses — entry is the first close above the buy point for a valid, non-extended setup. From there this page never sells at a fixed target at all:

• Trailing ratchet: once a day’s close first reaches +2R unrealized, the stop jumps up to lock in +1R of profit. From there, every further +1R of close-based unrealized gain ratchets the stop up another +1R, always trailing exactly 1R behind the highest level reached — a big gap-up day can cross several levels in one step. Exit is always the (rising) stop being hit; there is no take-profit sell, only a floor that only ever moves up.
• Opportunity-cost exit: before the ratchet ever activates, held more than 14 trading days without reaching +1R unrealized yet — sold at that day’s close, tagged stagnant, same safety net the fixed-target pages use.
• Trades open past 252 trading days close at that day’s price, tagged time_exit. One open position per ticker at a time — no pyramiding, no capital/position-sizing model at this layer (see the Portfolio Simulation below for that), R-multiples only.
• Runs against the current watchlist projected backward — survivorship bias applies (a stock that would’ve been delisted or dropped isn’t here to test).
• Use the tabs above to compare against the fixed 2R, 3R, 4R, 5R target pages — same entries/stops/opportunity-cost rule everywhere, only what happens once a trade is winning differs.