Section II · Conviction view
Refreshed 1d ago

Signals.

The raw alpha sources, before they become trades. Aggregated conviction across every active source; multi-signal hits on the same ticker get a 1.3× boost.

3 active insights · alpha-decay enforcedInspired by Lean Algorithm.Framework
How this connects to Plays

This page is the raw feed — every signal the engine detects, scored and ranked by conviction. A signal here is not a trade yet. To become a Play, a signal has to survive three more gates: the risk veto (concentration, budget, event blackouts), position sizing (per-leg + aggregate caps), and the fundamentals check. Only the survivors get an exact limit order and a Final Read verdict on the Plays page.

So: Signals = where ideas come from. Plays = the handful worth acting on today.If a high-conviction ticker here isn't on the Plays page, it got vetoed or sized out — that's the system working, not a bug.

Reading the clock: Age vs TTL

Age = how long ago we detected the signal. It tells you nothing about whether to act. TTL (time-to-live) = how long the edge is expected to last before it decays — this is what decides timing. Each signal type has its own shelf-life: a Triple-Confirm momentum signal expires in 24h (fast-fading), a GEX-Flip in ~36h, a GEX-Flip-Down in 48h. Act while a signal is Fresh; once it's Expiringthe mechanical edge is mostly gone and you're chasing.

Fresh (>24h ttl)
0
Cooling (6-24h)
3
Expiring (<6h)
0
Reflection — awaiting your review

Suggested threshold updates per signal

Walks closed shadows ≥7d aged, sweeps candidate emission cutoffs against realised win rate, picks the cutoff that maximises a Wilson 95% lower confidence bound (refuses to chase a 100% WR at N=2). Nothing is auto-applied — these are suggestions for you to approve.

124 aged closes
refreshed 22h 45m ago
SignalCurrentSuggestWin rateNWilson 95% LBAction
GEX_FLIP4684%1962%set initial
intraday_scaninsufficient_data
UW_FLOWinsufficient_data
MANUALinsufficient_data

To apply: edit state/signal-thresholds.json directly, or ask the agent to write the values for you. The reflection job runs weekly Fridays after market close.

What each signal means

Win rates are live from the de-duped shadow book (14attributed signals). Signals live in today's feed are marked.

GEX_FLIP_DOWNGEX Flip Down (bearish mirror, new 2026-05-28)live nowpaper / calibrating
The exact inverse of GEX_FLIP. Spot crosses the gamma-flip level going DOWN, so dealer hedging flips from vol-suppressing to vol-amplifying on the downside — they sell into weakness, accelerating the drop.
Fires when: Spot crossing the dealer gamma-flip strike to the downside with confirming bearish flow.
How we trade it: Slightly-ITM long puts. Starts at PAPER weight (0.7) — still calibrating vs an N=0 baseline, so treat as confirmation, not a standalone trigger yet.
Track record: no closed shadow trades yet — sample still building
GEX_FLIPGEX Flip / Negative-Gamma Breakout
Stock trading just below dealers' net-gamma flip level. Below the flip, dealers hedge by selling rallies + buying dips (this suppresses volatility). When price crosses the flip going up, that flow reverses — dealers buy rallies + sell dips → mechanical breakout fuel.
Fires when: Spot within ~1% below the dealer gamma-flip strike, then crossing it to the upside with confirming flow.
How we trade it: Slightly-ITM long calls (delta ~0.65), 30-45 DTE. Our strongest single signal — when it fires fresh, it's the highest-priority Play.
Track record: 68% win over 63 closed shadow trades · +$56,717 cumulative
TRIPLE_CONFIRM_LONGTriple-Confirm Long (三信号同步, new 2026-05-28)paper / calibrating
Eric's three-signal-sync momentum rule. Only fires when price action, trend, AND participation all agree — a classic momentum-confirmation filter. You guessed right: it IS the 5-day / 20-day rule, plus a volume gate.
Fires when: close > 5-day MA AND close > 20-day MA AND today's volume ≥ 1.3× the 20-day average volume AND today is an up day.
How we trade it: 8%-ITM long call (delta ~0.75), 35 DTE, target +40% / stop −20% (asymmetric — momentum trades cut losses fast). Score = 30 base + volume strength + how decisively price sits above both MAs.
Track record: no closed shadow trades yet — sample still building
CSP_INCOMECash-Secured Put incomeearly sample
Put sale at the dealer-defended put-wall strike. Collects premium with the wall acting as soft support. If assigned, you own the stock at an effective cost below where it trades now — yield without taking a directional bet.
Fires when: Elevated IV (rich premium) + a clear dealer put-wall below spot to lean on.
How we trade it: Sell the put at/near the wall strike. This is one of the income plays Eric wants more of in the shadow book.
Track record: 100% win over 1 closed shadow trade · +$65 cumulative
CC_INCOMECovered-Call incomeearly sample
Sell a call against shares you already hold. Collects premium; if the stock runs past the strike the shares get called away (capped upside, not a cash loss). Pure yield-on-a-hold.
Fires when: You hold ≥100 shares and the stock is rising into a strike where the premium is worth ≥$300.
How we trade it: Sell the call ~30-45 DTE. Income play — the CC book is tracked separately from the share book.
Track record: 100% win over 3 closed shadow trades · +$165 cumulative
UW_FLOWUnusual Whales Options Flowearly sample
Unusual institutional options volume — sweeps, blocks, big premium-weighted bets. The net 5-day dollar bias tells you which way smart money is positioning.
Fires when: Net premium-weighted flow skewed materially bullish or bearish over the trailing 5 days.
How we trade it: Direction confirmation for another setup — rarely traded standalone. Small but positive sample so far.
Track record: 71% win over 7 closed shadow trades · +$11,708 cumulative
IV_RANKIV Rankearly sample
Where current implied volatility sits vs its own 52-week range. High (>70) = premium is rich, favor selling; low (<30) = premium is cheap, favor buying directional options.
Fires when: IV rank crosses into the rich (>70) or cheap (<30) zone.
How we trade it: Sets whether we lean premium-selling or premium-buying — no directional bias on its own.
Track record: 83% win over 6 closed shadow trades · +$1,284 cumulative
MANUALManual override
A signal entered by hand outside the auto-engine. Tracked separately so the override has its own accountability.
Track record: 100% win over 3 closed shadow trades · +$1,083 cumulative

Conviction by ticker

Each ticker's blended score (0-100). 50 = neutral. The bullish/bearish columns show which signals are voting. Multi = 2+ signals agree → 1.3× conviction boost.

TickerScoreBullish signalsBearish signalsCountMulti
NDX42
GEX_FLIP_DOWN
1
MRVL37
GEX_FLIP_DOWN
1
SHOP37
GEX_FLIP_DOWN
1

Active insights (3)

Newest first. Age = since detection · TTL = edge-life remaining (act while it's not red).

TickerSignalDirectionConfAgeTTLSource
SHOPGEX_FLIP_DOWNDOWN37%28.9h19.1hsetup_hunter.breakdown
NDXGEX_FLIP_DOWNDOWN23%28.9h19.1hsetup_hunter.breakdown
MRVLGEX_FLIP_DOWNDOWN36%28.9h19.1hsetup_hunter.breakdown