No single indicator triggers a trade here. Each layer below answers a different question about a stock. On its own, any one of them can mislead. When several independent layers point the same way at once, that agreement is what creates a real edge. That stacking is called confluence, and it is the core of this system. Every idea card shows a Confluence score (how many of these line up).
Weinstein Stage Analysis splits every chart into four stages: 1 basing, 2 advancing, 3 topping, 4 declining. We only buy Stage 2 (uptrend) and only short Stage 4 (downtrend).
Trend template (Minervini): price above a rising 50, 150 and 200-day moving average, stacked in that order, within range of the 52-week high.
Why it matters: the large, durable moves happen in Stage 2. Fighting the stage (buying basing or falling stocks hoping for a turn) is the most common way swing traders bleed out. This is the foundation every other layer builds on.
Relative Strength (RS) measures how a stock has performed against the S&P 500 over ~6 months, then ranks it as a percentile. Longs want RS leaders (top ~20-30%); shorts want laggards. We also read 1-month and 6-month momentum.
Why it matters: leaders tend to keep leading and laggards keep lagging (the momentum factor, one of the most robust effects in decades of market data). Buying the strongest names in the strongest groups is where the outsized swing gains come from.
We read each of the 11 S&P sector groups through its SPDR ETF (XLK tech, XLF financials, XLE energy, XLV health, and so on): is the group above its 50 and 200-day averages, with the 50 over the 200, and outperforming the S&P? That makes the sector leading (bullish); broken-down and lagging makes it weak (bearish). Each card shows its group's read as a "Sector" chip.
Why it matters: roughly half of a stock's move comes from its sector. A breakout backed by a leading group has the wind behind it, so a strong sector adds a confluence factor to a long (and a weak, breaking-down sector adds one to a short). Used as a confluence layer, not a hard gate.
We define a pivot (a prior ~40-day high, the edge of a consolidation) and only enter as price clears it, using a buy-stop so the order fills only if the breakout is real. Shorts mirror this with a sell-stop below a breakdown level.
Why it matters: anticipating a move ties up money in names that never go. Buying the breakout means the trade is already "in gear" with momentum confirming, which raises the hit rate and shortens dead time.
Volume surge on the breakout shows institutions buying. Volume dry-up (the quiet, contracting volume of a tight base, the VCP signature) shows sellers are exhausted right before a move. In our own backtest, volume dry-up was the single filter that improved risk-adjusted returns, so it carries real weight in the score.
Why it matters: price without volume is a weak signal that fails often. Volume is the footprint of the big money that actually drives sustained moves.
ATR sets the initial stop beyond the stock's normal daily range, so noise does not shake you out but losers are still cut fast. Conviction sizing then sets the position: a base of about 1.25% of the account at risk per trade, scaled up for high-score, tight-stop setups and down for weaker / wider ones (roughly 0.5-3%), capped so no single name dominates.
The exit is where the returns are made: a winner scales out half at the first target, then lets the runner ride the rising 50-day moving average for the full trend, rather than being clipped early by a tight stop.
Why it matters: sizing and exits drive long-term results more than entries. In our backtest, letting winners run on the 50-day MA was the single biggest lever (it roughly doubled the compounded return versus a tight 20-day trail), and conviction sizing concentrates capital where the edge is strongest.
A CAN SLIM-style read of earnings growth, revenue growth and margins flags whether a technical breakout is backed by an actual business catalyst. Quality earnings add to confluence; weak fundamentals lower conviction (and, on the short side, support the thesis).
Why it matters: the biggest multi-month winners almost always pair a strong chart with accelerating fundamentals. The chart shows when; the fundamentals suggest why it can keep going.
The Market Weather score (0-100) blends leading indicators (VIX, breadth, the yield curve, defensive vs growth sector tilt, credit, futures) into clear bands that set the posture: 62+ take full setups at full 1.25% risk; 45-61 Neutral be selective (A+ only, half size); below 45 stop new longs and raise cash. A primary downtrend (S&P below its 200-day MA) blocks new longs outright. Shorts arm only when the tape weakens, the S&P loses its 50-day or 200-day average, or the market stalls at major resistance.
Why it matters: roughly three of four stocks move with the market. Trading with the tide, and sizing to it, is one of the largest, cheapest edges available. The same bands drive the cards, the alerts, and the autonomous paper engine, so everything speaks one language.
From confirmed swing pivots we read the structure: higher highs + higher lows (a real uptrend) vs lower highs + lower lows (a downtrend), plus range compression (a tightening base before expansion) and failed breakouts / liquidity sweeps (a false break that snaps back, often a trap). Pivots are only counted once confirmed, so nothing repaints.
Why it matters: it confirms the breakout sits inside a sound trend and flags the false breaks that trip up chart-pattern traders. Used as decision-support context, not a hard gate (see Validation).
On the last impulse leg we mark the retracement levels and the golden pocket (the 0.618-0.65 zone where strong trends most often find support), plus extension targets (1.272 / 1.618) for where a move may reach. When a pullback lands in the golden pocket while the trend is intact, that is a high-quality re-entry.
Why it matters: it gives precise, repeatable levels to buy a dip in an uptrend instead of guessing, improving entry price and reward-to-risk.
RSI / MACD divergence (price makes a new extreme but momentum does not) warns a move is tiring. The exhaustion flag catches names that are overbought or stretched far above their 20-day average, which is exactly when chasing a breakout backfires.
Why it matters: this is the one advanced layer the backtest promoted to a hard rule (the no-chase filter). Skipping exhausted breakouts cut the strategy's worst drawdowns meaningfully (see Validation).
From the same confirmed pivots we detect classic XABCD harmonics (Gartley, Bat, Butterfly, Crab, Cypher, Shark, ABCD) when their Fibonacci ratios line up, and mark the PRZ (potential reversal zone) with a confidence score. Inspired by harmonic / Fibonacci trading.
Why it matters: when a harmonic PRZ overlaps the golden pocket and the broader confluence, the reversal read is much stronger. Treated as a confidence aid for the human read, not an automatic trigger.
Each idea also carries a defined-risk option play, and which one is picked is driven by implied volatility (IV) vs the stock's realized volatility — the single biggest edge in options:
• IV rich (implied >> realized) → SELL premium: a bull put (bullish) or bear call (bearish) credit spread. You collect a credit, time decay (theta) works for you, and a roughly 30-delta short strike means about a 70% probability of profit. Defined risk = the spread width minus the credit.
• IV cheap → BUY premium: a bull call or bear put debit spread for cheap, leveraged, defined-risk direction. Lower probability, higher reward.
• Own the shares & up → covered call for income.
Every card shows the IV regime chip, the legs, net debit/credit, max profit, max loss, breakeven, and (for credit spreads) an estimated POP, so the trade-off is explicit before you act.
How they're traded well (and what the paper engine does): ~30-45 days to expiry; size each spread so its max loss matches your per-trade risk, capped by a separate options budget; scale out like the stocks do — bank half at T1 (50% of max profit on credit spreads, 65% on debit) and let the runner ride to T2 (75% / 90% of max), with a time stop near expiry to dodge accelerating theta/gamma (single-contract spreads can't be split, so they close whole at T1); trade only liquid contracts; and avoid holding short premium through earnings. Note: unlike the equity rules, options aren't backtested here (no historical option-chain data), so they run on paper as a live experiment riding the validated stock entry signals.
The advanced layers above were not added on faith. Each was A/B tested over 16+ years against the baseline strategy, holding portfolio mechanics fixed and measuring win rate, expectancy, profit factor, Sharpe, Sortino, drawdown and the year-by-year beat-rate vs the S&P 500 across bull, bear, sideways, high-volatility and low-volatility regimes.
What it delivers (16+ yr backtest): about 16% a year compounded with a worst drawdown near -18% — versus the S&P 500's ~14% a year at a -34% drawdown. So it beats the index's return with roughly half the drawdown, and finishes ahead of the S&P in a slight majority of individual years (it wins big in trending and bear years, and gives ground mainly in calm low-volatility grind-ups). Honest caveat: that means deeper, real drawdowns than a pure low-risk setting; past results never guarantee future returns.
What was promoted to hard rules: (1) letting winners run on the 50-day MA — the single biggest return lever, roughly doubling compounded return vs a tight 20-day trail; (2) conviction sizing (~1.25% base, scaled by score and stop-tightness), which lifts both return and the annual beat-rate; and (3) the no-chase (exhaustion) filter, which skips over-extended breakouts and cut the worst drawdown from roughly -35% to -18% while keeping nearly all the return. All three held up across regimes.
What was tested and rejected: piling on positions did not help (the strategy is signal-limited, not slot-limited), and using market structure or the composite quality score as hard entry gates degraded returns. Those stay as decision-support (shown on cards and charts to read confluence), not automatic triggers; harmonic and Fibonacci reads are confidence aids for the same reason.
The discipline: nothing is bolted on that does not pay for itself in the test. This is how the strategy stays robust instead of curve-fit to one backtest. Honest caveats: the test uses a fixed modern universe (winners and laggards) with realistic slippage, so some survivorship bias remains, and past results never guarantee future returns.
Each layer above is independent: trend, leadership, timing, volume, risk, fundamentals and the broad market are different questions with different data. A weak signal in one can be noise. But when a name is a Stage-2 RS leader, breaking out on a volume surge, near new highs, backed by real earnings growth, in a healthy market, that is many separate confirmations all pointing the same direction at once.
When unrelated signals agree, the probability edge does not just add, it compounds, and the false signals tend to cancel out. That is why the engine acts only on actionable, high-confluence setups with defined risk, and why the Confluence score on each card is the quickest read of conviction. Shorts use the exact same logic inverted: Stage 4, below falling averages, RS laggard, breaking down on volume, weak fundamentals, in a weakening market.
Bottom line: confluence turns a handful of ordinary indicators into a disciplined, repeatable edge, and removes the guesswork and emotion from when to act.