This is the page a published measurement of the Tickmill feed will occupy. No measurement exists. No capture has been run against this feed, no artifact has been generated and no digest has been computed, so every readout below reads NOT YET PUBLISHED rather than showing you something plausible. What is here is the shape: the identity fields a portrait will carry, the six dimensions of the fidelity battery, the chart each one will plot, the words a finding will be written in, and — stated before any result exists to soften them — the claims each dimension will still refuse to support afterwards.
A portrait is a profile, never a rating. There is no composite score on this page and there will not be one: a single editorial number is exactly the judgement surface corruption needs. Appearing on the roster is not a recommendation, a warning, or a statement that this broker has been contacted, has agreed to anything, or holds any relationship with Hadal. The roster is alphabetical and unranked, and it is published in full.
Identity of the thing measured
A feed measurement is only meaningful against a named account, a named interface and a named machine. "Tickmill" on its own is a brand, and a brand can front several entities, several account classes and more than one price source. The fields below pin the portrait to one of each, so a reader knows precisely what the numbers are about — and, just as importantly, what they are not about.
Identity fields the Tickmill portrait will carry. None is populated: no capture has been run.
Identity field
Value
Legal entity measuredThe specific entity whose feed was captured. A broker brand can front several, and they do not have to behave alike.
NOT YET PUBLISHED
Regulatory registration in scopeThe registration under which the measured entity operates, recorded as context for the capture — never as an endorsement.
NOT YET PUBLISHED
Account type measuredThe account class the capture ran on. Feed behaviour can differ between account classes at the same broker.
NOT YET PUBLISHED
Instruments in scopeThe symbols captured, listed individually. A portrait speaks about the symbols it measured and about no others.
NOT YET PUBLISHED
Platform or API capturedThe exact interface the ticks were read from, including version. A different interface is a different measurement.
NOT YET PUBLISHED
Capture host and locationWhere the capturing machine sat, because distance to the venue is part of what any timing figure describes.
NOT YET PUBLISHED
Relationship disclosureWhether an introducing-broker relationship exists, stated on the portrait itself under the gag-free partner filter.
NOT YET PUBLISHED
Measurement window and dataset
Every figure a portrait publishes is a statement about one bounded stretch of time and one recorded dataset. The window is published with the result, not buried under it, and the dataset is published with a digest computed from its own bytes — so a reader can check that the file they downloaded is the file the numbers came from, without asking me for anything.
Observation window opens
NOT YET PUBLISHED
The first timestamp in the capture, stated in UTC alongside the venue's own session clock.
Observation window closes
NOT YET PUBLISHED
The last timestamp in the capture. A portrait describes this window and makes no claim about any other.
Quote records captured
NOT YET PUBLISHED
The count of records the battery ran on, published so a reader can judge whether the sample carries the findings.
Dataset digestSHA-256NOT YET PUBLISHED
The digest is computed from the artifact's bytes at publication. It is never written by hand and never reserved in advance, which is why this slot is empty rather than holding a plausible-looking string of hex.
The battery, dimension by dimension
Every broker on the roster receives the identical six-dimension battery. Identical is the point: a battery that is adjusted per subject is an opinion with a method attached. Each panel below carries the same four parts — what the dimension measures, the chart it will plot, the readouts it will publish, and the claim it will not support. The fourth part is written now, while nothing depends on it, rather than conceded later under questioning.
01
Publication cadence
NOT YET PUBLISHED
The interval between consecutive quote updates on a symbol, taken as a distribution rather than an average: the floor, the tail, and how both move across a session. A mean cadence hides exactly the behaviour worth seeing.
FIG. 01NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot the inter-quote interval distribution across the observation window, per symbol, with session boundaries marked and any interval where no quote arrived drawn as absent rather than interpolated.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Median inter-quote interval
NOT YET PUBLISHED
Interval floor observed
NOT YET PUBLISHED
Upper-tail interval
NOT YET PUBLISHED
Quote records in window
NOT YET PUBLISHED
Symbols in scope
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will describe the shape of the interval distribution in words, name the symbols and sessions it covers, and state plainly where the sample was too thin to support a claim.
Does not establish. Cadence does not establish execution quality, fill price or slippage. A fast feed can still fill badly and a slow one can still fill honestly — cadence describes the publication of prices, not what happened to an order.
02
Staleness distribution
NOT YET PUBLISHED
How long each side of the quote stands unrevised before it moves, measured per side rather than pooled — because an asymmetry between bid and ask disappears the moment the two are averaged together.
FIG. 02NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot per-side quote age at revision, bid and ask as separate distributions over the observation window, with the two drawn on one axis so the asymmetry is readable directly.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Median bid age at revision
NOT YET PUBLISHED
Median ask age at revision
NOT YET PUBLISHED
Per-side asymmetry
NOT YET PUBLISHED
Longest unrevised interval
NOT YET PUBLISHED
Sessions covered
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will state whether the two sides age at the same rate, where the asymmetry sits if they do not, which sessions carry it, and how much of the window supports the statement.
Does not establish. Staleness does not establish that a quote was untradeable or that a price was wrong. An unrevised quote can be a live price in a quiet market, and this dimension cannot tell those apart.
03
Spread regimes
NOT YET PUBLISHED
How many distinct spread states the feed occupies, where the boundaries between them sit, and what accompanies a transition. Most feeds are described as though they had one state, and most do not.
FIG. 03NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot spread-state occupancy across the observation window with transitions marked against the session grid, and every unclassified interval left visibly unclassified.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Distinct regimes identified
NOT YET PUBLISHED
Occupancy of the widest regime
NOT YET PUBLISHED
Occupancy of the tightest regime
NOT YET PUBLISHED
Transitions observed
NOT YET PUBLISHED
Segmentation rule version
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will name each regime, describe the conditions under which the feed sat in it, and state which transitions the observation window was too short to characterise at all.
Does not establish. A regime count does not establish a typical spread, a quoted cost, or what any individual order would have paid. Regimes describe the published stream; they do not describe a trade.
04
Bid-ask coordination
NOT YET PUBLISHED
Whether the two sides of the quote update together or independently — the sentence structure of the feed, read as the joint pattern of bid and ask revisions rather than as two separate series.
FIG. 04NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot the joint revision pattern across the observation window — both sides revised, one side only, neither — as occupancy over time against the session grid.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Two-sided revision share
NOT YET PUBLISHED
Bid-only revision share
NOT YET PUBLISHED
Ask-only revision share
NOT YET PUBLISHED
Longest one-sided run
NOT YET PUBLISHED
Records classified
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will state in words whether the feed revises both sides together, which side leads when it does not, how often that pattern changes, and where the classification was ambiguous.
Does not establish. Coordination does not establish intent, manipulation, or that a price was synthesised. It describes the update structure of a published stream and says nothing about the motive behind it.
05
Freeze and rollover behaviour
NOT YET PUBLISHED
What the feed does at the seams: contract rollover, session close and reopen, and the intervals where quoting stops or repeats without revision. The seams are where a feed is least like its own description.
FIG. 05NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot feed state through each rollover and session boundary in the observation window, with quoting gaps drawn as absent regions — never joined across with a line.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Rollover events observed
NOT YET PUBLISHED
Longest quoting gap
NOT YET PUBLISHED
Repeated-quote intervals
NOT YET PUBLISHED
Session boundaries covered
NOT YET PUBLISHED
Gaps classified unexplained
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will describe what happened at each seam and classify every gap as calendar, venue or unexplained — the three are reported separately and never pooled into one number.
Does not establish. Freeze and rollover behaviour does not establish that orders could or could not be placed during a gap. It records what the published stream did, which is a different question from what the platform accepted.
06
Non-crossing enforcement
NOT YET PUBLISHED
Whether the published bid ever meets or crosses the published ask, and if it does, for how long and under what conditions. A crossed quote is an infrastructure disclosure, not a feature.
FIG. 06NOT YET PUBLISHED
NOT YET PUBLISHED
No capture has been run against the Tickmill feed, so there is nothing to plot. The chart renders here when an artifact stands behind it, and not one moment before.
The published chart will plot crossed-quote occurrences across the observation window against the session grid. If none were observed, the plate stays empty and the caption says so in words.
SourceNOT YET PUBLISHEDSHA-256NOT YET PUBLISHEDDateNOT YET PUBLISHED
Readouts this panel will publish, with the value each will carry.
Readout
Value
Crossed-quote occurrences
NOT YET PUBLISHED
Longest crossed interval
NOT YET PUBLISHED
Records inspected
NOT YET PUBLISHED
Symbols inspected
NOT YET PUBLISHED
Detection rule version
NOT YET PUBLISHED
Plain-language finding
NOT YET PUBLISHED
A published finding will state whether crossing was observed at all. If it was not, that is the finding, and it will be published as plainly and as prominently as any other result.
Does not establish. Non-crossing does not establish feed accuracy or venue integrity. A feed that never crosses can still be stale, slow or wide, and every one of those is measured elsewhere in this battery.
---
Publishing measurements a stranger can re-derive is the whole position, and it costs nothing while there is nothing yet to hide. The manifest below is what will ship beside the Tickmill portrait when it publishes — the file itself, its format, its size, its digest, the version of the code that generated it, and the licence you may re-use it under. Every field is empty today for the same reason every readout above is: the capture has not been run.
Artifact manifest. Each field publishes with the file, computed from the file.
Manifest field
Value
Artifact fileThe raw capture as CSV or JSON, downloadable without an account or an email address.
NOT YET PUBLISHED
Format and schema versionThe column contract the file obeys, versioned so a downstream reader can pin it.
NOT YET PUBLISHED
Size on diskPublished so a download can be checked for truncation before it is trusted.
NOT YET PUBLISHED
SHA-256 of the bytesComputed from the published file. Recompute it yourself; a mismatch is my defect, not yours.
NOT YET PUBLISHED
Generating code versionThe tagged version of the battery that produced the file, so the run can be repeated.
NOT YET PUBLISHED
Re-use licencePermissive with attribution. The licence text awaits my ratification and is not yet posted.
NOT YET PUBLISHED
Methodology version, pinned
battery Feed & Executionmethod version v1.0portrait state NOT YET PUBLISHED
A measurement without a pinned method is an anecdote. This portrait is pinned to the Feed & Execution receipt page at version v1.0 — the document that defines each metric, its estimator, its known failure modes and its limits, and that carries the pre-registrations and the recomputation procedure. When a portrait publishes it names the version it was measured under, and that version is frozen: a definition that can move after a result is published is not a definition, it is a degree of freedom.
The method publishes before the first broker is measured, not after — which is the only ordering under which a pre-registration means anything at all. The wider framework sits on the methodology index, and the instrument that performs this battery is the Broker Feed Auditor.
The rest of the roster
Listed alphabetically and never ranked. Every portrait below is in the identical state as this one — unmeasured — so no order among them could carry a judgement even if one were intended.
Back to the Observatory index for the methodology banner, the anti-rating declaration and the gag-free partner filter that governs which relationships Hadal is willing to hold at all.