The Lab

PF > 1.0 = profitable. > 1.5 = good. > 2.0 = rare. Max DD is peak-to-trough on the equity curve.

All three streams started with $1k. Flat stretches = bot idle. Slope > 0 = winning.

Open circle = entry; ▲/▼ = exit pointing the way the bot was facing. Dotted line = round trip · green = net win, red = net loss. Amber dashed = currently-open position extending to now.

Same trades, plotted as $ outcome vs time. Streaks and clusters jump out faster than on the equity view.

Fat right tail with thin left tail = good. Symmetric or fat left = the strategy is sampling noise.

Which exit type is doing the most work — and which is bleeding money?

Symmetry check. If one side drags the numbers, that's a regime issue rather than a strategy issue.

After the 2026-05-14 LIVE↔shadow swap, shadow runs the bare hybrid_v5.1 baseline (no regime gate, no reverse-exit, no grid-what-if). The split lets us measure whether the live-side mode flags actually improve outcomes vs the unmodified strategy. Promotion bar still applies for any new mode flag we may want to A/B.

Upper-left = clean winners with shallow drawdown. Lower-right = trades that touched profit but ended losing (give-back). Cloud shape tells you how much of the move the exit logic converts.

Short-and-winning vs long-and-losing? Tells you if the bot is cutting too early or riding too long.

Skim this when reopening the dashboard.