Methods
What Pulse measures, what it never publishes, and how figures, limitations and revisions work.
Pulse describes aggregated instructions received by the PineConnector bridge. A tracked instruction is one instruction received by the bridge; it is not a broker fill, an executed order, an open position or a trading decision. The network methodology sets out the population, windows and exclusions.
Every recurring series, generated from the series registry. Each entry lists what the series does not measure.
Pulse publishes only shares of a published whole and their percentage-point changes, within-window indices, ratios of published shares or indices, percent changes of instruction activity between windows, and ranks or medians of those. It never publishes absolute network quantities, commercial sizing, broker or prop-firm attribution, or performance and profitability claims. Figures are instruction-weighted: each instruction counts once.
The indices and instruction measures present source-linked views of these registered measures.
The headline activity measure keeps the registered percentage-change form. It shows the raw calendar-month comparison alongside the standardised-week comparison, with an explicit label for each basis.
The share measures and retrospective Notes reuse approved fields from the August bulletin. No new extraction, population filter or audit attestation is implied. Where a prior share is derived, it uses the released current share and its percentage-point change. Downloads include the exact source-field address and transformation.
The August release recomputed July using consistent eligibility rules. Its July comparison can differ from the independently published July bulletin; these views do not splice the two releases into a continuous history. The source time zone remains unconfirmed.
The stop-loss-field measure covers recognised fields in alert text. It does not measure whether a stop is defined in the receiving EA or accepted at the broker. Weekend crypto share is a within-month comparison of day-type composition, not a historical trend.
Period reports end with their report content and reference data. General definitions and interpretation limits are available in this hub and the network methodology. Working Papers and Notes retain their numbered limitation notes and backlinks.
Working Papers open with a question, hypothesis and test written before the data is pulled, and publish them as written even when the hypothesis fails. A result that fails any robustness check is not published as a finding.
Figures and tables are numbered within each publication. Each has a caption, a source line and a CSV of the exact published aggregate behind it, at /data/<publication>/<figure>.csv. The CSV is generated from the same data as the chart, so the two cannot drift. External market context is attributed to its source and is not republished as Pulse data.
Period-report revision and correction records are preserved below from their source records. Working Papers and Notes retain their own revision history. Earlier values and reasons remain recorded; a presentation change does not restate data.
Partial-month coverage → complete-month coverage Complete source-calendar months replace the earlier matched partial windows. Both periods now exclude parser-rejected records, malformed command-symbol echoes and obvious placeholder symbols; these eligibility refinements also change the comparison population.
2026-08-27 → 2026-08-31 August coverage now includes the complete month.
2026-07-27 → 2026-07-31 The July comparison is recomputed for the complete month under this edition’s eligibility rules. These figures may differ from the separately published July report.
-0.02 → -4.8 Total instruction-activity change (%): recomputed on full August and July source-calendar months with consistent eligibility.
19.1 → 8.1 Entry instruction-activity change (%): recomputed on full August and July source-calendar months with consistent eligibility.
-23 → -22 Management and other instruction-activity change (%): recomputed on full August and July source-calendar months with consistent eligibility.
65.1 → 64.8 Entry share of all tracked instructions (%): recomputed on full August and July source-calendar months with consistent eligibility.
31.5 → 31.8 Gold share of entry instructions (%): complete-month replacement.
The Markets section uses public data with its own sources and terms. See Sources & Methods for each source, its reuse terms and its limits.