Every tool below survived the same process that buried the rest. Each one states exactly what it does — and in plain English, what we are not claiming it does.
Every tool ships with a published specification and a written statement of its limits.
A floor calculator for prop accounts. You enter your account state and it draws the exact chart price where your trailing max-loss limit breaches — with live distance, max size for a given cushion, and the 16:40 ET flatten countdown.
Not claimed: that it monitors your account. TradingView can't read your broker balance — every figure is one you type in, and it computes the floor from your chart. Re-enter your peak balance daily.
A Multi Time Frame Panel — a table of trend state, price vs EMA, across five timeframes, using only closed bars. Identical on history and live: certified non-repainting.
Not claimed: that ALIGN is a signal. Combining timeframe states showed no predictive edge in our testing — it's a count, not a confluence score.
Draws the Initial Balance (09:30–10:30 ET) — high, low and the 50% midline — plus optional ±1× / ±2× range extensions and a shaded IB zone. Builds in the window, freezes at 10:30.
Not claimed: which way an IB break goes. A break is a volatility event with no directional content in our data — the tool marks the levels, nothing more.
Marks when a higher-timeframe pool — prior-day/-week high or low, optional swings — is swept: price grabs past the level a little and closes back on the origin side. Triangle + pool tag at each sweep.
Not claimed: the reversal. Sweep → reclaim → reversal tested to a selection artifact with negative expectancy — so there's deliberately no target and no long/short trigger.
Where volume actually traded, and who was leaning on it — buying and selling split out at every price level.
Not claimed: that any level will hold, reverse, or be reached. It shows you the auction — reading it is your job.
Order blocks that only print when real volume showed up, with major and internal structure tracked separately.
Not claimed: that price reverses at these blocks. We tested order-block fades and they don't beat a random level — this is a structure map, not a signal.
Base-and-departure zones confirmed by a real impulse leg, not just one big candle.
Not claimed: that these zones are a tradeable fade. Level-and-limit fades tested dead on our data. Use them for location and context.
The FWD4 engine reduced to one condition: an institutional-volume breakout of the Value Area. It marks the moment size steps in behind a break — volume ≥ a multiple of average, with supporting CVD — not a bare price break. Tags the chart BRK ↑ / BRK ↓.
Not claimed: a trade or a direction. BRK names the condition — a big-trade-backed break of value — it doesn't tell you to enter. The raw setup's direction tests as noise against a time-of-day baseline; confirm with your own tape and decide.
The fade that shouldn't work — until you test it. When momentum diverges into a swing level, fade it: the market rejects that level more than chance allows. It cleared our full gauntlet across a decade of NQ, and again on ES and 6E — the same pattern in all three, not one lucky dataset. You take the entry; the exit is your call.
Monte Carlo · Breakpoint resampled from its own trade resultsNot claimed: a hit rate or a hands-off system. The entry went through the gauntlet; the exit is yours, and your execution is where the outcome is made. You trade it — it does not trade you.
The one that cleared everything — as a live signal. ADX 39 fires BUY / SELL on our most-tested, most-robust direction edge, and marks a modeled exit as a reference. It's no knife-edge: it sits in the middle of a stable band of settings, not one lucky number, and fires on nearly every session. You take the entry; you manage the exit.
Monte Carlo · ADX 39 resampled from its own daily resultsNot claimed: a hit rate or a performance figure. “Most robust” is a statement about how hard we tried to break it — not a promise about your account. It has losing stretches, and they’re in the record.
The equity curves shown for the models are Monte Carlo simulations of hypothetical / backtested results — not the results of actual trading, and not a promise of future performance. See the full hypothetical-performance disclosure in the footer.
We're in soft launch. Paid checkout opens mid-August — until then, founding members get in early through the community and help steer what we build next, before anyone pays a cent.
The full toolkit — the read-the-market tools, on any TradingView plan. Best if you run your own process and just want cleaner inputs.
Just the two models we put a decade of tick-level research behind — for traders who already have their charts set the way they like and only want the reads. Ships with the evidence.
Every tool, plus the two models we put a decade of tick-level research behind. This is the layer that took the most work to stand up — and it ships with the evidence. The tools show you the chart; the models tell you what kind of session you're reading it in.
A separate one-time seat: every tool and both models, plus everything we ship next — with no recurring billing, ever. Independent of the subscription plans above.
No performance representation is made for any product or plan. Nothing here is investment advice. See full risk disclosure in the footer.
Not a tearsheet. The things that actually let you judge whether a tool belongs in your process.
Exactly what triggers what. Every input, every condition, every edge case. You should be able to predict what the tool will draw before it draws it — that's the standard.
What we tested, on what data, and what came back. Including the tests that failed and the versions that got cut.
A plain statement of what the tool does not do and where it goes wrong. Every product page on this site carries one — that's not a legal footnote, it's the product.
Install, configuration for your instrument and timeframe, and a direct line in the community when something doesn't behave the way the spec says it should.
Ask for the full specification on any tool. Nothing here is sold on a screenshot.
