SignalDeskPRICE ACTION LAB
BACKTEST ONLY
VERSION 13 · CONTROLLED THREE-WAY BACKTEST

Read the trend.
Test the price action.

Compare V12 unchanged against price action alone and indicator-confirmed price action. Same data, same risk model, separate frozen forward tests.

128Research candidates
148Model tests across 3 arms
30%Annual target*
*RESEARCH GOAL, NOT A FORECAST
30% annually ≈ +271.3% in 5 years$100k → $371,293

V12 earned 4.91% annualized with a 4.93% drawdown in your 128-symbol test. V13 keeps the original logic untouched and asks whether price-action confirmation adds a measurable edge. The 30% annual target is not a prediction.

01

Connect historical data

Use the Alpaca paper-account API keys you already created.

Not connected

Your keys stay in this browser session and are transmitted only through your HTTPS Pages backend to Alpaca. Nothing is saved in reports, local storage or source files. Never paste keys into chat.

02

Configure the V13 comparison

Use the same 128-symbol universe as your V12 study for a direct comparison, or run a smaller pilot first.

Unchanged V12 controlThe exact V12 portfolio engine and 48 candidate models are rerun as a controlled baseline.
V12
Price action onlyBreakout acceptance, confirmed swing structure, support rejections and retests. Also tests range-market support rejection.
20 MODELS
Indicators + price actionThe 48 V12 models compete with eight price-action-confirmed profiles × four exits, using prior evidence only.
80 MODELS
Three controlled portfoliosEach experiment has its own $100k cash account, max 5 stocks and historical entry-price limit under $200.
FAIR TEST
Starting capital $100,000Risk per entry 0.5%Max open stocks 5Sector concentration 40%Entry share price < $200Cost model 10 bps + fees

The three controlled tests use completed daily bars and may take longer than V12; use a desktop for 128/228 symbols, keep the tab open, and do not include pre-market or news AI in these historical results.