Amagedon reads live market data from Schwab and turns it into three things: where price is against the levels that matter, what that structure usually leads to, and which side has a tested edge at this time of day. It works for any futures market or stock — pick yours once and every page follows.

• Home — What to expect, the levels ladder, gamma and stress at a glance, and the AI analysis.
• Levels — value area, initial balance, volume and today's gap, with the prices.
• GEX — dealer gamma, the flip level, gamma walls, put/call and the expected-move cone.
• Rotation — which sectors lead, risk appetite, and spreads on your stocks.
• Track record (menu) — the timed bias reports, their backtest and the live record.
• Settings — your markets, alerts and theme.

New to a word? Tap any dotted word on the site, or open the Glossary — gamma, POC, value area, base rate and the rest, in plain words.

Nothing here is financial advice. Every read says whether it has been tested and how often it was wrong.

Tap 🔍 (or the search box on Home) and type what you trade — a name like “gold”, “natural gas”, “10-year” or “tesla”, a futures root like “/NG”, or a ticker like “AAPL”. Futures come from a built-in list of common contracts; stocks and ETFs from Schwab's own search.

• The chips beside the search are your markets. Tap ☆ on a result to add it; manage the list in Settings.
• The market you pick follows you: choose natural gas on Levels and GEX, Home and Track record open on it too.
• Each market is measured on its own session hours. Using the stock market's 9:30–4:00 for natural gas would put its open and its value area in the wrong place.

A plain reading of the live structure: a scenario, a plan with the levels in it, and one line on each input. It is the textbook reading of these conditions, not a backtested signal — the card says so.

Volume uses the same cut-offs everywhere on the site, alerts included: 1.5× or more is heavy, 1.1× above normal, under 0.8× light.

Every plan ends with what would make it wrong. Markets with no options data (natural gas, crude, Treasuries, currencies, grains) are read without gamma, and the card says so.

Every level the site computes for your market in one column, sorted by price, with the live price in its place and how far each level is: value area (VAH, POC, VAL), the multi-day POC, overnight high and low, first-hour high and low, naked POCs, the gap-fill level, the gamma flip and largest gamma walls, open-interest walls, and the options-implied move for the nearest expiry. The six nearest on each side are shown; levels a few points apart are merged into one row. Tap a name for what it means.

What to expect adds a Tested edge line only during the main session, only when today has a side (the move from yesterday's close to 30 minutes after the open was big enough to call), and only when this market's backtest shows an edge for that rule at the current report time or a later one — in which case it names the time, e.g. “From 2:00 PM”. It gives the hit rate, the base rate and how often it was wrong, and disappears as soon as the record stops showing an edge.

The timed bias reports moved off Home to their own page. They are still captured on schedule, so the live record keeps growing; this page is where to check whether a tested edge holds on calls made before the outcome was known. Everything below describes that page.

The Now box

Which side to look for trades on at this minute, using the same rules as the timed reports below it:

• Look for longs / shorts — the rule for this time of day has shown an edge in testing.
• Wait for (time) — today has a side, but this market's tested edge only starts at a later report time, e.g. 2:00 PM. Nothing before then.
• No side yet — overnight with no usable evening read; it says when the session read comes.
• Lean long / short — a side, but the test is unclear. Not a proven edge.
• No clear side — the opening move was smaller than usual, or the rule has no edge.

Whenever a side shows, a line says how often it was wrong in testing and what the live record is.

The report times

Horizons end at this market's own close, and a report that would start after the close is dropped. Gold (close 1:30) has the 8 PM and 10:30 reports, the 10:30 one running into its 1:30 close. Energy (close 2:30) has 10:30 into 2 PM and 2 PM into its 2:30 close, and no 3:30 report. Treasuries, currencies, grains and livestock work the same way against their own closes. The 8 PM report runs to this market's open.

A move smaller than this market's median of the last 20 sessions is no call — small moves are noise. Reports are captured on time by the server and frozen, then graded when their horizon ends.

Reading the numbers

• Test — the same rule replayed over up to a year of 5-minute history: how often it was right.
• Base — what always saying the more common direction would have scored. In a rising market “always long” looks clever; a rule must beat this.
• edge — at least 20 calls, and ahead of the base by at least 100 × √(0.25 / n) points, roughly one standard error (about 7 points at 50 calls, 5 at 100). no edge — at or below the base. unclear — ahead, but not by enough yet, or fewer than 20 calls. Each market is tested on its own history.
• Live — only calls made before the outcome was known. This is the record that decides whether a rule works; −Γ / +Γ split it by dealer gamma at the time.
• 8 PM rules tested — alternative evening rules scored on the same nights, so a better one can replace the live rule on evidence.
• Simulation — every past call as a trade, in points and dollars. No commissions or slippage.

A short written read of the live data — at most about 110 words: a headline, Trade (entry, targets, stop), Why in plain words, and Changes if. Listen reads it aloud. It only runs when at least three of the five data sources are live; otherwise it says why instead of guessing. It is an opinion from the data on screen, not a tested signal.

• Value area — the price range holding 70% of the session's volume. POC is the busiest price, VAH / VAL the edges. Inside value is balance; outside it is either acceptance at new prices or an excursion the volume is not backing.
• Relative volume — today's volume against recent sessions at the same time of day. It decides which of the two a move outside value is.
• Initial balance — the first hour's range. A strong break of it is the classic trend-day sign; staying inside it is a balance day.
• Overnight range, naked POCs (old busiest prices no session has revisited) and low-volume nodes (thin pockets price moves through quickly).
• Range budget — how much of a normal day's range is already used.

Levels come from 5-minute bars, so they are accurate to the profile's bucket width, not to the tick.

• Fill level — the contract's own close of the previous main session: the price a gap fill trades back to.
• Open, the gap in points, Last and its distance from the fill level, and Filled so far with the time it filled.
• Fill rate — how often past gaps of this size and direction filled. Continuation — how often they closed beyond the open in the gap's direction. They are different numbers.

For NQ, ES (and their micros), RTY and gold the historical rates are measured on a cash ETF (QQQ, SPY, IWM, GLD), because a futures daily bar opens at 6 PM and does not show the day-session gap. Other futures have no such proxy, so they get no fill or continuation rates — only the fill level, the open and whether it has filled. Stocks and ETFs use their own history. The prices above are always the contract's own.

• Long gamma — dealers buy dips and sell rips; moves tend to be absorbed and breakouts fail. Short gamma — dealers chase price; moves tend to run.
• Regime — the sign of total dealer gamma at the current price, for the nearest expiry. Positive is long gamma, negative short.
• Flip — where modelled dealer gamma (Black-Scholes, nearest expiry, each option's own volatility) crosses zero, scanning 10% either side of price; the crossing nearest to price is shown. None is shown if there is no crossing in that range. Within 1% of it the regime can change during the day.
• Gamma walls — strikes with large dealer gamma, where price tends to slow or pin.
• Put/call (open interest and today's volume), OI walls, and the expected-move cone by expiry.

For futures, ETF strikes are converted to futures points with the live ratio of the future to the ETF, taken at the ETF's last trade. If the futures price cannot be fetched, a fixed approximate ratio is used and the page says so in a warning — levels can then be off by tens of points.

Sector strength against SPY or QQQ over 1, 5 and 20 sessions; risk appetite (cyclicals against defensives); how many megacaps are beating the index; and a tape of spreads and liquidity for the stocks and ETFs in your markets. Futures are left out of the tape because Schwab's quotes for them are unreliable.

Phone notifications that arrive with the screen off, for the markets you choose in Settings (up to 12; NQ if you choose none). The server checks every 5 minutes.

The last four fire once when their condition starts, re-arm after it clears, and repeat at most once an hour. Turn alerts on in Settings, pick your markets, then tap Save Markets.

• My markets — up to 12 markets of any kind. They are the chips in every market picker. Reset restores the defaults.
• Alerts — turn phone notifications on and choose which of your markets to watch.
• Theme — the colour scheme.

• Tested vs untested. The timed reports (Track record) have a backtest and a live record. What to expect and the AI analysis do not — they describe what conditions usually lead to.
• Base rates matter. A 60% hit rate is nothing in a market that went up 60% of the time.
• Samples are small. A year of 5-minute history is roughly 250 sessions, and each report time sees fewer calls than that after weak days are skipped.
• An edge is a tendency. A 66% rule is wrong one time in three. Size accordingly and use a stop.
• Data limits. Schwab gives 5-minute bars, not ticks; futures options chains are not available; some rates use a cash ETF as a proxy. Each page states where that applies.

Score, Macro and Amatout run on an older engine built around NQ and ES. Their detailed reference is in the older docs, which also describe some features that have since been removed.