LIVE MARKET DATA · THE MEASURED REACTION
The measured reaction, beside the read
The same platform holds the live market data, so it measures what the market did after an event and publishes that beside its own read — including when the two disagree.
LIVE MARKET DATA · DELIVERY
50-100 msLIVE
Top of book is live across global markets. A separate measurement from the event push and the written interpretation, over a different stage.
WHY HOLDING THE DATA IS THE POINT
The pulse measures the market, not us: no accuracy or hit rate is published from it.
BEFORE THE EVENT · POSITIONING MOVES THE READ
For a scheduled event, positioning in the final day is treated as live evidence about the event.
From a day ahead to the final minute the pre-event view consumes two order-flow readings for the event's reference instrument. Crowded positioning can widen the stated ranges or push the card into abstention.
It depends on positioning data for that instrument. Outside the final-day window the trigger does not run.
AT THE EVENT · THE INSIDE MARKET ON THE CARD
When an event publishes, the card carries the live inside market for the instrument it points at.LIVE
ON THE CARD · THE QUOTE
Price, both sides, spread
The live price with both sides of the quote and the spread between them, so you can see whether the move you are being shown happened in a market that was trading normally or one that had gone wide.
ON THE CARD · THE BASELINE
The move, and what it is measured from
The percentage move is never published without naming its baseline: a quote captured before the event, a quote captured at detection, or the prior session close. Where none of those is available the card says the baseline is none rather than picking one.
The quote block states its own condition: live with a measured reaction, live with none measured yet, or a quote still outstanding.
AFTER THE EVENT · THE REACTION PULSE
After the event the pulse measures the move at four fixed horizons. Measurement horizons, not delivery times, and the set is fixed.
HORIZON 01
5 seconds
The first read. Fast enough to catch the initial reaction before it is reshaped by anything else, and the baseline the later horizons are compared against to tell a move that held from a spike that did not.
HORIZON 02
30 seconds
Whether the first move survived its first half-minute. If the early spike has already given most of itself back, the reading changes from a confirmed move to a faded one.
HORIZON 03
120 seconds
The settling read, two minutes after the event. Compared against the first read, it separates a move that stayed from one that has already started unwinding.
HORIZON 04
300 seconds
The last fixed horizon. Nothing is measured beyond five minutes, so we never describe a move hours later as part of the reaction to the event.
WHAT COUNTS AS A MOVE IS SET PER INSTRUMENT
A move is judged against what that contract normally does over a short horizon, not against a flat percentage.
45
BENCHMARK CONTRACTS · THE RULER
A fixed benchmark set checked after an event to see whether broad markets reacted. It is the measuring set, not a research universe and not a count of what the product covers: research and alerting are not confined to it.
+ THE EVENT'S OWN ASSETS
MEASURED ALONGSIDE THE BENCHMARKS
The instruments the event itself touches are measured as well as the benchmark set, so a reaction in the thing directly affected is not hidden behind a broad index that barely moved.
COMPANY EVENTS
THE BENCHMARKS ARE DROPPED
For a company event the benchmark set is removed entirely and the reaction is measured on the issuer's own ticker. A broad index move tells you nothing about one company's news.
THE PUBLISHED VERDICTS, IN PLAIN WORDS
CONFIRMED
A move large enough to matter for that instrument, and it was still there at the horizon being quoted. It says the market moved materially — it does not say we were right. For a smaller qualifying move, or for a card that published no direction, this verdict is set whichever way the market went.
DIVERGENT
A large move against the direction the card published. Where a card states a direction and the market moves strongly the other way, this is the verdict, and it is published on the card rather than quietly dropped.
FADED
There was an early move, and most of it was given back by the later horizon. The reaction happened and then unwound, which is a different fact from no reaction at all.
ABSORBED
Heavy one-sided aggressive buying or selling that did not move the price, because something passive on the other side kept taking it. One of the few readings only a depth and flow view can give you.
NO REACTION
Measured, and nothing material happened. An event that moved nothing is a finding, and it is published as a finding rather than left blank.
No verdict means “we were right”. The pulse reports the market, and DIVERGENT names a move against the published direction.
THE DEPTH AND FLOW YOU CAN INTERROGATE
Behind the verdict sits a ten-level book and an order-flow read, which you can ask for in chat where available.
DEPTH · TEN LEVELS
Live-anchored, and it tells you how
The top of book is live across global markets. The deeper structure refreshes on a delay in some markets, so the ten-level book is live-anchored: the structural shape is re-centred onto the live inside market, keeping the relative offsets between levels. Where no usable structural profile exists the ten levels are anchored instead on the live top-of-book trajectory, swept-volume depletion, passive replenishment and the session's traded volume at each price. Every pull is stamped with which route produced it and with its data mode, so you are never left guessing which you are looking at.
FLOW · THREE-REGIME READ
Adds, cancels and aggression
The order-flow read separates what was added to each side of the inside market, what was cancelled off it, and what was executed against it, across price-up, price-flat and price-down transitions. It publishes event pressure, cancel pressure, trade pressure and sweep direction, plus the size imbalance at the inside market and across the ten levels.
WHAT THE ORDER-FLOW READ ACTUALLY SAYS
- a stacked buying-imbalance support zone has formed, and price is now retesting it;
- sellers appear trapped at the low of the bar;
- selling dried up to a single print at the low, which is exhaustion;
- an outsized print at the low suggests a strong passive buyer;
- aggressive sellers were absorbed in the lower wick before buyers took control;
- price is breaking out of a two-sided absorption zone.
Each arrives with its cautions attached in the same voice.
WHERE IT SAYS IT DOES NOT KNOW
These are published states, not error pages, and each is a different fact.
STATE · UNMONITORED
No tape for this instrument
The instrument the event points at is not one we hold a tape reading for. The pulse reads unmonitored and the card says so, rather than publishing a tidy zero that would look like a measured absence of movement.
STATE · PENDING
Measurement outstanding
The event is published and the reaction has not been measured yet. The card carries the event with the reaction explicitly pending rather than holding the event back or filling the field in early. A pending reading that never resolves is moved off pending deterministically instead of sitting there indefinitely.
STATE · DEGRADED
Depth did not arrive
The reaction was reached without the depth evidence that would normally sit behind it. The pulse is labelled degraded so you can tell a reading made on thin evidence from one made on full evidence.
STATE · PRICE ONLY
Price measured, no depth behind it
A separate, narrower honesty on the card: the price reaction is real and the depth and flow evidence is either absent or still pending. The card distinguishes the two cases instead of implying depth it did not have.
STATE · NO QUOTE APPLICABLE
No market quote belongs here
Some events have no instrument a market reaction could sensibly be read on. Rather than attach a plausible-looking number to the nearest index, the card states that no market quote is applicable.
STATE · PRE-EVENT ABSTENTION
Unscheduled shocks publish no pre-event read
The positioning trigger refuses to produce a pre-event probability for an unscheduled exogenous shock, and it also refuses when its inputs are stale or an input sits outside the range it was calibrated on. It abstains with a reason rather than guessing.