rangeProb-Tracksmaps · odds · record
rangeProbDirection feedGet the appMembership · $100/mo
HOME
LIVE NOWFuturesStocksCrypto
BOARDS ▾
ROADMAPS ▾
↻ Past roadmaps
HIGHER TIMEFRAME
Weekly mapsMonthly maps
NEW YORK · 8:45am
NQESYMRTYGCCL
OVERNIGHT
ASIA · 8pm
NQESYMRTYGCCL
LONDON · 2am
NQESYMRTYGCCL
ALSO
Session statsStock roadmaps
TOOLS ▾
MEMBERSHIP ▾
START & LEARN ▾
The record

Twenty years of calls,
every one of them graded.

We only started publishing on 9 August. So here's the same engine pointed at twenty years of one-minute bars, with each year priced by a model that had never seen it. Download the whole thing and check us. No account, no email, no card.

Does it deliver what it claims

the only test that matters

Anyone can post a win rate. The question is whether a number we publish before the outcome exists turns out to be the number you get. Group every call by what we said it was worth, then count what happened.

Read the last column. That's the gap between the promise and the delivery.

By market

six futures

Year by year

including 2008, 2020 and 2022

No year is carrying the others. That matters more than the headline, because an edge that only worked in one stretch shows up here as one fat year and nineteen thin ones.

Read this before you trust it

Every grade is in here, not just the good ones

S, A and B together, nothing filtered out. Filter the grade column to compare them. S is the strongest read and B the weakest, and all three go out live exactly the same way, so a record that quietly showed you only the S calls would be worth nothing.

These are not calls we published in 2007

Our published record starts 9 August 2026 and it lives on the boards, graded in public as it happens. This file is what the engine would have called, run honestly against old bars. We'd rather say that plainly than let you assume otherwise and find out later.

Each year was priced by a model that had never seen it

The odds on a 2015 call were fitted only on 2006 through 2014. Then the window rolls forward and does it again, twenty times. Run an engine over the years you tuned it on and it comes back excellent and meaningless, which is what most published backtests are.

What we didn't roll: how a setup is defined in the first place. That structure was designed once, with the whole history visible. Stronger than a backtest, short of a perfect simulation, and we'd rather you heard the limit from us.

Every call carries two stamps, both ET

Fired at is the minute the engine made the call. Not the period close, not when someone got round to writing it down. Resolved at is the minute the setup settled and the call became a hit or a miss. The gap between them is how long it took, so you can see the holding time on all 208,553 of them rather than taking our word for it.

Line either one up against your own bars. That's the point of publishing them.

Longs run hotter than shorts

What a hit means here

The setup resolved the way the call pointed. That's a different question from whether price travelled far enough to pay you, and on our live ledger that second question grades around 63%. Both are real, both are ours, and we publish both rather than quoting the flattering one and hoping you read it as the other.

See today's calls Ask your own question →