About
What the market has done, what it is trying to do, how good a job it is doing, and what it is likely to do next — then checked against the session as it runs. Every figure computed from the trade record, and every read graded afterwards against the session that followed.
Each session is read at four zoom levels — monthly, weekly, daily, intraday — because a level only means something relative to the timeframe that built it. The read names who holds territory and who is offside, maps the levels that decide the next move, and frames the branches the session can take.
Then the session is checked against it. Intraday snapshots revisit the morning’s call while the market is still running — what has actually happened, which branch is live, and what would change the answer. Analogs set the day beside the historical sessions that most resemble it at the same clock time, as 5-minute charts you can sort by eye, with what each one went on to do.
All three are for subscribers on the day and public the next, graded. The subscription buys the timing, not different work.
The framework does not change between sessions. Same four questions, same order, whether the read is interesting or dull. That constancy is what makes a run of reads comparable rather than a collection of opinions.
Every figure is computed, not estimated. Value areas, points of control, order-flow delta, initial balance, range extension — all derived from the market’s own trade record. A level the read cites that is not in the computed data is rejected before publication rather than printed. Prices are back-adjusted continuous, so windows spanning a contract roll stay comparable.
The synthesis is machine-written. A language model composes the narrative from those computed inputs through a fixed framework. The framework is human-designed and each read is human-reviewed, but the prose is not hand-written and it would be dishonest to imply otherwise.
When a read is late, it says so. The job runs unattended before the open and occasionally breaks. The read is still written from data cut at 08:05 — nothing from the session already underway — and carries a notice saying when it was written. The alternative is a silent gap in the archive on exactly the days something went wrong.
A read is not a prediction. It is a set of scenarios — a lean, and the branches that would mean the lean is wrong — each with the price that makes it live. It is graded on whether the set captured the session and whether a reader could tell early which branch was running.
So a read can lean the wrong way and still be a good read: on one session the lean was short, never triggered, and the alternate — a reclaim leaving sellers trapped — is what happened, on the level the read had named that morning. Equally, coverage bought by vagueness is not coverage: branches have to be separable, and breadth is graded down rather than rewarded.
A level breaks when price closes beyond it on two consecutive one-minute bars — a wick through and back confirms a level rather than breaking it. A scenario is live from the moment its trigger is accepted and dies only if its invalidation is accepted afterwards, so a session reads as a path: a level is taken, that branch is live, it may be stopped, another may take over.
Grades come from a different model than the one that wrote the read, distances are judged in average true range, and every read keeps its grade — including the bad ones, which stay up unedited. The scoreboard shows the whole record, including what each read called against what the session actually did.
Changed 12 August 2026. Invalidations used to be checked from the opening bell rather than from the moment a branch went live. That marked a scenario as beaten whenever the market simply opened on the far side of its invalidation — which had happened to every third scenario ever published, none of which had ever traded to its trigger. The acceptance rule itself is unchanged. The whole archive was re-scored under the corrected rule; this note stays because a scoring rule that changes quietly is not a rule.
It is not advice, and it is not a signal service. There are no entries, no exits, no targets to act on, and no track record being advertised — because no trades are being published.
Mechanical entry triggers built on order flow do not survive honest out-of-sample testing. The value of a read like this is subtractive: which levels matter, which side is already committed, and when the honest answer is to stand down. It is climate, not a forecast. Where a read is weak it says so; a read that only ever sounds confident is not being written carefully.
Nobody’s name is on these, deliberately. If a read is only worth reading because of who wrote it, the method has already failed. Accountability lives in the framework and the archive instead — every read, dated, in order, right or wrong.
Published pre-open, one per session. The archive has everything, oldest reads included, and the RSS feed carries each new one.
Read the latestThe pre-market read is published pre-open. Same structure every time: what the auction did, what it is trying to do, how well it is doing it, and what it is likely to do next.