BTA / Delivery desk
Alert Latency Is Not the Same as Slippage
A time interval and a change in attainable price describe different parts of a trading workflow.
Name the measured endpoints
Latency needs two observable events. A server dispatch and an endpoint receipt can describe transport. A device display and a person's review can describe attention. Publication-to-fill combines several systems and decisions. Do not shorten that chain to a delivery claim unless the record contains every endpoint required by the definition.
Save clock provenance and uncertainty. If two clocks disagree, subtraction can produce a negative apparent interval. That does not show that the message travelled backwards. It shows that the measurement, clock alignment or event labels need inspection. BTA's worksheet reports intervals under the reader's explicit clock assumptions and will not assign a vendor speed score from those inputs.
Price movement requires another observation
A later price can improve, worsen or cease to be actionable under the original rule. Elapsed seconds do not determine slippage. A fill comparison needs the instrument, side, size, quote convention, reference time, actual order and fees. A chart's best later or earlier price is not automatically an attainable account price.
A limit order controls its permitted price, not whether it executes. A market order has a different price boundary. Those distinctions matter when evaluating a delayed message. Do not transform a hypothetical timestamp exercise into a return calculation, and do not assign unfilled messages a favourable fill merely to complete a table.
Report two bounded findings
The operational finding might say that a particular route displayed a labelled test message within a measured interval under a defined device configuration. The trading finding would require separate account evidence. If that evidence does not exist, report it as absent. A route working once is useful evidence of that one test, not continuous availability or profitable execution.
Comparisons need identical endpoint definitions and compatible clocks. A vendor's server timestamp and another vendor's screen timestamp are not interchangeable. Publish sample scope, missing-event counts and the tested route alongside any future timing summary. Without those records BTA compares documented workflow, not a fastest-alert leaderboard.
Original Evidence Packet
| Artifact | Record to retain |
|---|---|
| Interval definition | Exact start/end event and clock bounds |
| Price evidence | Reference convention and actual order or explicit absence |
| Sample scope | Route, environment and failed or missing trials |
Continue the Investigation
Source Boundaries
Official references describe published mechanisms. BTA's evidence packets and proposed test procedures do not claim a completed paid-feed or account trial.