This is the operating manual for the FlowState desk — the session windows, the markets, the contract math, and the trade lifecycle, written plain. We don't publish the engine's full internals. We publish everything you need to understand every trade it calls — and to learn the discipline yourself.
Every FlowState day is the same shape, every market, no exceptions. The bar below is live — it tracks New York time and shows exactly where the desk is in its cycle while you read.
Same engine everywhere — but the engine respects each market's character. Indices lose their breakout edge fast, so they get a curfew. Gold trends all morning, so it doesn't. Validation numbers are from the public backtest window (May 17 – Jun 10, 2026, 1-minute data) — backtest, not live.
| MARKET | CHARACTER | ENTRY WINDOW | $ / POINT | MICRO VERSION | VALIDATION | VERDICT |
|---|---|---|---|---|---|---|
| NQ · Nasdaq 100 | fast, violent, trend-then-chop | first 90 min only | $20 | MNQ · $2 | +$9,370 · 100% (10/10) | the flagship · earns the curfew |
| ES · S&P 500 | steadier, deeper liquidity | first 90 min only | $50 | MES · $5 | validation queued | trades when the loop signs off |
| YM · Dow | slow grinder, cleaner levels | first 90 min only | $5 | MYM · $0.50 | validation queued | in line behind ES |
| RTY · Russell 2000 | small caps, honest ranges | first 90 min only | $50 | M2K · $5 | +$12,850 · 91.9% · PF 4.68 | quiet overachiever |
| GC · Gold | trends for hours when it goes | all morning · no cutoff | $100 | MGC · $10 | +$88,780 · 95.6% · PF 10.6 | the morning marathon |
Point values are CME contract specs. A 100-point NQ move = $2,000 on one NQ, $200 on one MNQ. Same chart, very different heartbeat — that's why sizing is half this manual.
18:00–04:00 ET builds the overnight range: high, low, and the golden midpoint. The range IS the map — every level the engine trades comes from it.
At 04:00 the box locks and gets graded against the daily ATR. Too wide (indices: above ~0.40×) and the whole day is RANGE REJECTED — announced out loud, no trades, no exceptions.
Two entries exist. Breakout: candle-close through a rail. Midpoint launch: a clean reaction off the golden mid. Every setup is scored 0–100 (HTF bias, range quality, time, prior control zones).
Indices only fire in the first 90 minutes — late index entries went 0-for-16 in validation, so the engine wrote itself a curfew. Gold keeps hunting all morning. Re-entry only on a full reclaim with HTF agreement.
Targets ladder off the range (TP1→TP4). At +0.45R of open profit the stop jumps to break-even — the seatbelt. Each TP fill drags the stop up behind it. Risk only ever shrinks.
15:55 ET: everything goes flat. No overnight bags, no "it'll come back," no funding-rate roulette. Tomorrow gets a fresh box and a fresh mind.
The desk trades a ladder: 3 contracts on midpoint launches, 2 on breakouts, +1 bonus runner — scaled out at TP1→TP4. But the count only makes sense relative to the stop distance. Wide box = smaller size. Tight box = full playbook. The rule that matters: decide the dollars you'll lose if the damage line hits, before you care about how many contracts that is.
Educational math, not advice. Read the stop distance off the actual chart — the damage line is drawn on every setup the desk calls.
Micros are the same chart, same fills, same lesson — at 1/10th the heartbeat. One MNQ through a full FlowState cycle teaches you more than a hundred hours of watching. Graduate to minis when your journal says so, not your ego.
We didn't design FlowState for prop evals — discipline just happens to look the same everywhere. Defined risk before entry. One setup family. Flat by 15:55. No news gambling, no revenge sequence, no overnight exposure. That's a checklist most funded accounts die failing.
Our own prop-futures field guide: 14 modules on eval rules, drawdown math, firm-by-firm fine print, and the calculators to survive them. Built by this desk, for the people grinding evals.
◈ GET THE FUNDED EDGEPartner slots: one per category, disclosed always · partners never touch the engine.
Overnight box still drawing. Nothing to do but watch the map form.
Box locked, quality passed, window open. The engine is hunting both rails and the midpoint.
Price is closing in on a trigger. This is the chip that should raise your pulse.
In a position. Stop, targets, and protection step all on the tile.
Box too wide vs ATR. The day is declined. Discipline you can watch.
Index curfew passed. Manage what's on, take nothing new.
A stop-out left a reclaim level. If price reclaims it WITH HTF agreement, one second chance.
Flat. Journaled. The desk is studying itself until 18:00.
The overnight session range, 18:00–04:00 ET. High, low, and midpoint of everything that traded while the US slept. The entire day is traded off this structure.
The exact middle of the box. Price respects it like a magnet — clean reactions there are one of the two entry triggers.
The invalidation level — where the trade idea is proven wrong. Stop lives there from the first second. Capped against ATR so one bad box can't nuke a week.
Box size divided by the daily ATR. Compact box = energy stored = tradeable. Bloated box = energy spent = RANGE REJECTED.
0–100 grade on every trigger: higher-timeframe bias, range quality, time of day, prior control zones, VWAP side. The score is shown live — you'll learn what an 83 feels like vs a 55.
Projection targets off the box (TP1→TP4). The position scales out piece by piece instead of begging one exit to be perfect.
At +0.45R open profit the stop jumps to entry. Born from a trade that went +$3k green and died −$26k red. Never again — literally a rule now.
Two of them. Indices stop taking entries 90 minutes after the lock; everything goes flat 15:55 ET. Both written in backtest blood.
After a stop-out: if price comes back through the level that killed the trade — and the higher timeframe agrees — the engine takes one disciplined second shot. One.
The higher-timeframe read (sweeps + candle structure). It scores every setup, and it's the gatekeeper for any re-entry.
The line we never cross: this manual explains how the engine trades a simulated book, for education and entertainment. It is not financial advice and not a recommendation that you place any trade. If you trade, that's your decision, your account, your risk — the manual just makes sure it's an informed one.