Probabilistic 1-day price range for the system's AI Technical Stock Picks — expected range, P(high/low reached), and bullish/bearish breakout scenarios.
Pulls the current top picks from stock_picks.php (trend/RSI/MACD/relative-strength/volume screener) and/or the manual /watchlist, plus any symbols you add above, and layers a range forecast on top of all of them.
Ready.
Click a row to see the full narrative and layer-by-layer breakdown below.
Select a symbol above to see its full range forecast.
Layer 1 — Expected range. Blends up to four independent 1-day move estimates: (a) ATR(14) — the 14-day average of true range (high−low, gap-adjusted); (b) historical volatility — the annualized standard deviation of daily log returns over the last 20 sessions, converted to a dollar move the same way an options desk converts implied volatility: Expected move ≈ Price × Volatility × √(1/252); (c) a 4-hour-bar ATR(14), scaled up to a full session using the actual number of 4-hour bars in the latest trading day (range grows roughly with √time), so the estimate reacts faster to recent intraday swings; and (d) the options-implied expected move, when a nearest-expiration ATM call+put quote is available from Polygon's options chain — Expected Daily Range ≈ ATM Straddle Price × 0.85, the same forward-looking, market-priced 1-standard-deviation estimate an options desk uses (cross-checked against the quick Rule-of-16: Expected Move % ≈ Implied Volatility % ÷ 16, shown in the detail panel but not blended in separately since it's derived from the same quote). ATR-based estimates (a) and (c) measure a full high−low range rather than a one-sided move, so they're halved before blending with (b) and (d) so all represent the same one-sided sigma. When no options data is available, annualized historical volatility alone stands in as a proxy for implied volatility. The available estimates are averaged into a single 1-day sigma (σ), and the resulting high/low bounds are widened if needed so they never sit inside the range the stock has already traded through today (per the 4-hour bars).
Layer 2 — Direction. A −1..+1 bias score blends: trend (price vs 20-day average), momentum (RSI 14), MACD histogram, relative volume vs the screener's baseline, and price vs today's volume-weighted average price (VWAP). The bias shifts the expected range's center by up to half of σ instead of assuming a purely symmetric, driftless day.
Layer 3 — High/low probabilities. Rather than treating the closing price as the only random variable, the day's path is modeled as a Brownian motion with drift (μ = the Layer 2 bias × σ, volatility = σ). The reflection-principle formula for the distribution of that process's running maximum/minimum gives P(high ≥ target) and P(low ≤ target) directly, which is why the two probabilities in the table aren't simply mirror images of each other.
This is a statistical/technical heuristic, not a prediction — treat the output as a confidence band, not a guaranteed range.