TRADING PLAYBOOK
Chart
METHODOLOGY

Deployment

Snapshot ·

Live portfolio simulation across both strategies + today's recommendations.

Portfolio Performance

Slot caps + trigger semantics + cooldowns. Forward-test record from 2026-04-26. CORE=long-term (V9 ∧ Timing percentile-0.85 gate, 5 slots, V9-trail exit). SWING=short-term (rank_product ≥ 0.80, 6 slots, V9-trail exit).
CORE · Long-term v9_and_timing · percentile-0.85 gate · V9-trail exit · 5 slot cap
SWING · Short-term rank_product ≥ 0.80 · V9-trail exit · 6 slot cap
REAL MONEY Logged trades · ML exit overlay @ q25 (CORE threshold)
Open positions
EXIT signals firing
Last scored
REAL MONEY No held positions. Use the LOG TRADE tab to enter a trade, or add rows to data/held_positions.csv.

Today's Recommendations

CORE

Long-term

V9 ∧ Timing percentile-0.85 gate · V9-trail exit · 5 slot cap
Scan date
Timing date
ML threshold
Picks ⓘ
Gate: percentile · top 15% V9 ∧ top 15% Timing
Absolute-mode filters (inert)
SWING

Short-term

Rank product ≥ 0.80 · V9-trail exit · 6 slot cap
Scan date
Timing date
ML threshold
Picks ⓘ

Market Regime

Snapshot ·
Sector Rotation — Relative Rotation Graph
Where each sector's relative strength (vs SPY) sits and which way it's rotating, over the last weeks. A map of WHERE strength is and its direction of travel — it does not call the bottom. Descriptive, not a signal.
vs · wk
Sector rotation unavailable (insufficient history).
QUALITY WATCHLIST
Snapshot ·

Stocks ranked by a gradient-boosted model trained to predict the top 20% of 12-month forward performers from fundamentals, sector context, and cycle-age proxies. Independent of breakout setup timing — this is the "which companies to watch" signal.

Methodology →
Sector
Show top
Search ticker
TIMING — WHEN TO BUY
Snapshot ·

Stocks in the Quality top 25% universe, ranked by a gradient-boosted regression that predicts realized R-multiple under the production exit (6R breakeven + 21-EMA close trail, no partial), with a 20-bar swing-low structural stop. Higher pred R = better expected trade expectancy today. Independent of pivot breakout — the "right day to buy" signal.

Strategy
Sector
Show top
Search ticker

Watchlist

Add tickers directly or select from the Universe tab

Surge Watch

Fresh volume thrusts off 52-week lows or out of a long base. as of

Discretionary surfacing screen, not a backtested edge. Catches fresh moves: a name with an abnormally large volume thrust (≥3× average, up, real move, within the last ~10 sessions) that is near its 52-week low or breaking out of a long tight base — explicitly not names already trending up to new highs. This is the off-the-lows / base-emergence species — the speculative, fat-tail side (high failure rate; a sizing + selection game). Ranked by thrust size. Use it to find fresh moves to read, then size small.
shown
Trend + Liquidity Leaders above the 21-EMA / 50-SMA / 200-SMA, ADR ≥ 4%, ADV ≥ $100M
Presets fill the filters below — tweak any field to switch to Custom.
match · scanned · as of
Scanning universe…
No stocks match these filters.
Screener unavailable — check server log.
Page /

Bases

Names currently in a proper base, ranked by base quality. as of

Discretionary surfacing tool, not a backtested edge. A name shows only if it is actually in a proper base (structure = hard gate); base_quality ranks them (structure 55% · RS line 35% · near-high 10%). The RS line — not the lagging RS rating — is the leadership filter: a blue dot = RS at a new high while price isn't (it leads price out of the base). Depth is read as retracement of the prior advance + 50-EMA hold, not absolute %. Forward-edge validation is a separate later step.
shown

Pullbacks

Uptrend names pulling back to a key MA, ranked by relative strength. as of

Discretionary surfacing screen, not a backtested edge. A name shows only if it is in an overall uptrend (rising 200-day SMA) and is currently pulling back to a key MA — 50/100/200 (whichever of EMA/SMA price is nearer) — either within ~2 ATR of the MA or having crossed it in the last few sessions. Rows are ranked by relative strength (the one durable factor), not a return-predicting score. The MA-pullback-reclaim was tested as a mechanical edge and falsified (Keystone Phase 0b: ≈88% standalone stop-out, a variance trade). Use this to find names to look at, then read the chart yourself.
shown

Trigger Watch

Leading-theme RS leaders on a key MA that printed an entry-trigger bar, ranked by relative strength. as of

Discretionary surfacing screen, not a backtested edge. A name shows only if it is a top-decile RS market leader (≥90th pct) in a leading-theme sector (RRG Leading/Improving), a real mover (ADR > 2%, >$20M avg $-vol), sitting within ~1.5×ADR of a key MA (10-EMA / 21-EMA / 50-SMA), that just printed a low-risk entry-trigger bar — a tight/NR coil on dry-up volume (TB), a shakeout tail (ST), or a gap-down reversal (GD). Rows are ranked by relative strength (the one durable factor), not a return-predicting score. Makes no edge claim. Use it to find names to look at, then read the chart yourself.
shown

Up/Down Volume

Ranked by money-flow-weighted accumulation (50-day). as of

Discretionary surfacing screen, not a backtested edge. Ranks each liquid name by its money-flow-weighted up/down ratio — every day's volume is weighted by the size of that day's % move, then up-day flow ÷ down-day flow. >1 = accumulation (volume concentrated on up moves), <1 = distribution. Shown over both a 21- and 50-day window and ranked by the 50-day; the accel column (21d − 50d) flags when recent accumulation is picking up or fading. No return is predicted — use this to find names to read.
shown

Universe Scanner

Snapshot ·
Snapshot: Jul 13, 2026 05:01 PM Updating prices — ✓ Updated tickers Update error · Ready Scanning — ✓ Scan complete Scan cancelled Scan error
HOW THIS SYSTEM WORKS

Scan candidates are ranked by a conviction score from a gradient-boosted model trained on 3.2M simulated trades. Each trade is sized with a concrete stop and managed with a tested exit rule — a 21-EMA close trail with a 6R breakeven, no partial — selected from a leak-free exit-parameter sweep.

Read the methodology →
✓ Scan Complete
Scan cancelled. Partial results may be shown below if any stocks were scored before cancellation.
Scan Error

Position Sizer

Account Size ($)
Setup
Trade Grade
Risk % (R)
Entry Price ($)
Stop Price ($)

Log a Trade

Leave Exit blank to log an open trade. Close from the Open Positions panel below.
Open Positions
Trade Details
Date
Ticker
Setup
Direction
Entry ($)
Stop ($)
Exit ($) — blank = open
Shares
R% Risked
Grade
Context Snapshot
Looking up... ✓ From scan Not in latest scan Enter ticker to auto-fill
Regime at Entry
EMA Setup
Composite /100
Tech /25
RS Percentile
Conviction
Pred_R
Size Tier
V9 ∧ Timing
ML_Pred is computed server-side at log time (requires ≥1 trading day after entry).
Notes
Part XII Scorecard
Total Score

Weekly Review

Weeks back:

ML Backtest — Full History

Step 1 — Build Features & Train
Computes point-in-time features across all history, trains LightGBM with walk-forward CV.
Model — Out-of-Sample Results