Net Premium intraday call vs put premium — fade or continuation?

The read: price up with net call premium rising = real money behind the move (continuation). Price up while call premium bleeds or put premium builds = a rally nobody is paying for (fade candidate — today's TSLA 3:25 pop was exactly this). FUEL / FADE chips are computed from the last 15 minutes of tape. Solid lines = UW's real tape; dashed = Gammatron's own stream-derived replica (the post-trial version) — watching them track each other is the calibration.

UW net-prem ticks per minute (fast lane, 4-min refresh) + native per-minute net premium from liveflow prints. Local-only. Green = net call premium, red = net put premium, cumulative from the open.

Flow Feed

Direction Radar

Tickers where ≥75% of flagged premium leans one way — the strongest one-directional bets of the day — plus prints matching the "knew something" profile: far OTM, near-dated, opening, bought aggressively, sized.

One-Way Flow

Suspicious Prints

Key Levels

Support & resistance implied by options positioning: gamma walls, open-interest concentrations, gamma flip and max pain. Red = dealer resistance · blue = dealer support.

GEX Engine

Vol Lab surface integrity · variance risk premium · implied correlation

Three checks on the volatility surface, after Bloch’s A Practical Guide to Quantitative Volatility Trading. Integrity asks whether the quotes are even self-consistent before anything else trusts them. VRP replicates fair variance from the whole strike ladder and compares it to what the stock actually delivered. Implied correlation asks whether a basket is priced to move together or apart.

Surface integrity static no-arbitrage violations — broken quotes, not free money

A surface is arbitrage-free only if total variance never falls as maturity rises, call prices stay convex in strike, and the Durrleman condition holds. A violation means a quote is bad. Contracts failing convexity are excluded from GEX and the screener on this scan, so a stale print can’t invent a gamma wall.

Variance risk premium model-free fair vol vs realised

Fair vol comes from the log-contract replication over every OTM strike, not one ATM quote — the gap between them (skew prem) is what the wings are charging. VRP < 0 means the tape is moving more than the ladder is priced for; ratio > 1.35 means premium is expensive relative to delivered movement. H is the Hurst exponent: above 0.5 moves extend, below 0.5 they fade. Thin ladders are flagged — they truncate the integral and bias fair vol low. Realised vol is quoted jump-truncated: a gap cannot be captured by hedging, so leaving an earnings move in the trailing window would fake a cheap premium. A jump badge shows the raw number and how much of the variance that gap alone accounted for.

Implied correlation dispersion indicators per basket

Index vol against the weighted vol of its components. Low correlation = names are priced to move independently, which favours single-name convexity and lets sympathy pairs diverge. High = it is a beta tape and the index expresses it more cheaply. Mρ > 1 flags the index straddle pricing above the basket. Weights are approximate index weights from config.json.

Regime Radar Hurst exponent, tested rather than assumed

The popular version of this says: H above 0.5 means momentum is real and breakouts hold, H below 0.5 means every breakout is a trap. The maths is sound — it is the same long-memory framework Bloch works through — but three things decide whether the number means anything, and they are usually left out. (1) Raw R/S is biased up: a pure random walk reads ~0.56 on a 60-day window, so an uncorrected 0.61 is barely different from noise. We apply the Anis–Lloyd correction, which re-centres it on 0.50. (2) The estimate is noisy: standard error is ~0.09 on 60 bars, so a 0.46→0.59 flip is about one standard error. We bootstrap each name's own returns to get its noise floor and report a p-value. (3) One estimator is not enough: we cross-check against DFA, each measured against its own null.

Does it actually predict? measured on your universe, not asserted

Today’s tape what each session has done so far — descriptive, not a forecast

H needs about 24 bars, so on 5-minute data it only becomes computable around midday — it cannot be read off the first hour, which is when you actually trade. The efficiency ratio (net move ÷ distance walked) and the opening-range break are available from the open and are shown alongside it.

Daily regime 60-day window, bias-corrected, bootstrapped

HIRO Intraday hedging impact of options flow

powershell -NoProfile -ExecutionPolicy Bypass -File run_hiro.ps1

Lotto Tracker large OTM ≤3 DTE bought prints — did they hit?

Every big out-of-the-money print expiring within 3 days gets tracked from the day it fires: latest price, peak, and what it was worth at the open and close of expiration day — the data behind the "enter Wed/Thu, sell Friday morning" gamma play. HIT 4x+ marks options that went 300%+ over entry.

Dealer Levels+ GEX · vanna · charm — where hedging flows live

SpotGamma-style dealer positioning on the index complex. GEX pins or accelerates price; vanna is the $delta dealers must trade per 1-point IV move (positive net vanna = a vol crush forces them to buy); charm is the daily $delta drift as hedges decay. The hedging band spans the gamma flip and the dominant wall — inside it, hedging dampens moves; below it, hedging chases them.

Dollar greeks per strike, dealer convention (calls +, puts −), from EOD open interest and IV across all scanned expiries ≤31 DTE. Vanna matters most when IV is elevated — a big positive net vanna into falling vol is systematic buying pressure (the classic "vanna rally" after an event passes). Charm flows accelerate into Fridays as weekly deltas decay. A model of positioning, not a guarantee.

Screener risk/reward vs the market's own expected move

Two mechanical screens over the whole EOD chain, liquidity-filtered (spread ≤25%, OI ≥200, volume ≥20). Sell premium: puts at least one full expected move below spot, delta ≤0.30, ranked by annualized yield. Cheap convexity: near-ATM longs whose breakeven move is ≤80% of the expected move the market itself is pricing — convexity priced below its own forecast.

Sell premium cash-secured puts ≥1 expected move OTM

Cheap convexity breakeven inside the expected move

Expected move = spot × IV × √(DTE/365) per contract. Delta proxies probability of finishing ITM (EOD data, no real-time). Very low BE/EM ratios are usually deep-ITM stock-replacement calls — little extrinsic to lose, but little leverage either. EARNINGS = report lands inside the trade. A screen, not advice.

Live Flow

Prints inferred in real time from Schwab's streaming quotes: each row is a jump in a contract's traded volume, priced against the bid/ask at that instant. BOUGHT = lifted the ask, SOLD = hit the bid. Auto-refreshes every 10s while this tab is open.

Schwab doesn't sell a raw OPRA time-and-sales tape, so these prints are derived: volume deltas on the ~1,200 near-the-money contracts the streamer is subscribed to. Several trades inside one update window collapse into one row, and contracts outside the subscribed band won't appear. Start it with start-liveflow.bat; rebuild the universe midday if price moves a long way. Not investment advice.

Pre/Post Movers

Every US-listed symbol quoted through Schwab in real time, ranked by extended-hours move. % and $ are measured against the regular-session reference close — prior close during pre-market, today's 4pm close after hours. Click any column header to sort; use the filters to cut noise.

Catalysts & sympathy movers whose headlines name each other — one story, several tickers

Run run-premarket.bat in the options-flow folder to refresh this board (pre-market 4:00–9:30am ET, after-hours 4:00–8:00pm ET). Quotes are Schwab real-time; extended-hours prints are thin and spreads gap wide — the Spread and $ volume columns are there to show you what's actually tradeable versus a single odd-lot print. A screen, not advice.

Technicals EMA stack · compression · volume — crossed with today's flow

Daily bars, EOD scan. EMA stack = 8/21/50 exponential; 200-EMA available in data for regime checks. A screen, not advice.

ORB Pax-style opening range — first-minute range + extension ladder

Method (public Pax Group teachings): the true opening range is the very first seconds of the session — strength holds above it, weakness below. Fixed-step extension levels ladder out from the OR (new one added each time one breaks) and act as scale-out targets, trading in thirds/quarters. Our free data floors at 1-minute bars, so OR here = the 9:30–9:31 bar — the closest legal approximation of the 30-second range.

Pax shared levels paste levels he posts on X into pax_levels.json

Track record does the first break hold to the close? builds daily

OR breaks that HOLD to the close validate the "direction for the rest of the day" thesis; low hold-rates mean fade the break instead. Steps: SPY 1.5 / QQQ 1.6 (≈ ES 15 / NQ 65 scaled), 0.3% of price elsewhere — editable in config.json > orb. Methodology credit: The PAX Group (thepaxgroup.org); this is an independent free-data approximation, not their indicator. A study, not advice.

Momentum · Friday 0DTE Desk the basket · the tape · the catalysts

Every day split in two: the overnight gap (prev close → open) and the intraday drive (open → close). Strong regimes — the MU kind — gap the same way again and again and hold it into the close. Score blends trend strength vs the name's own volatility, up-day persistence, gap alignment and volume. Real/IV = realized daily move vs what IV30 charges per day — above 1.0 the swings out-pay theta. GREEN = strong momentum and overnight swings worth holding through; DAY ONLY = trade it, don't sleep in it.

Daily bars (Schwab real-time, Yahoo fallback) + CBOE IV30. Map: each column is one day — top dot the overnight gap, bottom dot the intraday drive (green up / red down, brighter = bigger). Hover a row for the reasoning. Regenerates after every scan. A screen, not advice.

IV Tracker front-month ATM implied-vol movers, day over day

Where did the market's expectation of movement change the most since the last scan? A jump in ATM IV means options got repriced for a bigger move (event risk, positioning, fear building); a crush means an event passed. Sorted by absolute change.

ATM IV = median implied vol of contracts within 3% of spot, 5–45 DTE, from the EOD snapshots. Implied move = spot × IV × √(DTE/365) — the one-standard-deviation move the market is pricing into the front expiries. Rising IV with a flat spot is the classic "market expects something" tell.

Tail Risk deep-OTM put buying — who's paying for crash protection?

A filter for Universa-style tail hedges: far out-of-the-money puts being bought, the trades that bleed a little every day and pay 10–100x in a crash. Universa's public profile (Spitznagel/Taleb lineage) runs ~0.5–1% of portfolio into S&P puts roughly two months out and ~30%+ OTM, rolled monthly. Prints matching that shape get the UNIVERSA PROFILE chip.

Filter: puts, not sold, ≥7% OTM, ≥7 DTE, from today's flagged flow. UNIVERSA PROFILE = ≥15% OTM, ≥15 DTE, ≤$2.50 — the closest match our 31-DTE scan window allows to the canonical ~2-month ~30% OTM structure (extend flow.max_dte in config.json to 60–90 to catch the full shape). Index/macro ETF prints matter most — single-name deep puts are usually hedges or bets on the name, not systemic tail positioning. A screen, not advice.

Trade Ideas

A mechanical screen of today's flow into the coming Friday expiration: heavy one-way premium, multi-day persistence, wall targets from the GEX engine, earnings warnings from the calendar. This is a data screen, not financial advice — every card shows its evidence so you can judge it yourself.

Event Calendar volatility catalysts ahead

Macro events: ForexFactory public feed (USD, high/medium impact). Earnings: Yahoo, universe tickers, refreshed every few days. Add investor days or other events to events_custom.json as [{"date":"2026-08-15","title":"NVDA investor day","ticker":"NVDA"}].