How Troop reads a setup
TradingView is the pattern engine. Troop is the layer that records what it saw, lines it up across timeframes, and — over time — measures what actually worked.
Strategies & universe
Each strategy (top-left switcher) watches its own universeof tickers. A ticker in the universe gets 15m, 1H, 4H and 1D prices from a free feed (yfinance, delayed) and a status matrix. Removing a ticker keeps its history — Troop can always answer "what was in the universe on date X".
The status matrix
Rows are indicators (chart patterns + RSI), columns are timeframes. Pattern cells come from TradingView alerts; the latest alert for that pattern and timeframe decides the cell:
TradingView detected the pattern; it hasn't broken out yet.
Breakout (or breakdown for bearish patterns) — the entry-quality event.
The pattern was invalidated. Kept in history; scores nothing.
Price reached the pattern target — relevant for exits, not entries.
Bullish patterns score positive, bearish negative. RSI (14)is computed by Troop from the same price bars (TradingView's formula): above 55 is +1, below 45 is −1, and overbought/oversold extremes are flagged but not scored.
Signals age in bars of their own timeframe— a 15m breakout is stale after 16 bars (4 trading hours), a 1D breakout after 20 sessions. Nights and weekends don't count.
Timeframe verdicts
Each column's points are netted into a verdict:
Mixed means real bullish and bearish evidence on the same timeframe — often the most important thing to notice.
What these numbers are not
Setup points are a transparent ranking, not a probability of profit. Troop will only show probabilities once it has measured them: every alert is stored with Troop's own price at the moment it arrived, so the outcome of each pattern (did it reach its target? how fast?) can be counted — with sample sizes and error bars — before any of it drives a recommendation.