THE WORK BEHIND MOMENTUM
Research & development.
A documented account of Momentum’s development, tested model variants and historical trade scenario evaluations.
RESEARCH, IN NUMBERS
The scale of the
research programme.
Archive review
19 September 2026
These figures describe the volume of development and testing. An evaluation is a recorded assessment of a historical trade under a particular test setting; the same trade can be assessed more than once.
Explore each development version.
Select a version to explore its recorded testing or development work.
The original research baseline.
The initial historical replay recorded 2,621 completed test trades. Candidate assessments are shown separately because a screened opportunity does not necessarily become a trade.
- Completed historical test trades
- 2,621
- Candidate assessments
- 36,290
Each completed test trade contributes one evaluation. Candidate assessments are excluded from the programme total.
Source: Original baseline report, preceding Version 2 and mapped to Version 1 in the development record.Each stage is shown in the units supported by its records. Model variants may recur across versions and are counted once in the programme’s minimum of 36. These figures describe research activity, not customer orders or investment returns.
Supplementary research included in the total
A separate historical research run, archived as Stage 8, contains 2,494 completed test-trade records assessed under 4 recorded settings. This contributes 9,976 scenario evaluations. It is distinct from the Version 8 tournament and from Builds 10 and 11.
The programme total is 1,138,102 evaluations from the numbered development record plus 9,976 from this supplementary run: 1,148,078 in total. The supplementary figures were reconciled against the retained trade ledger and summary, with their file hashes checked against the original manifest.
How the figures are counted
The version count follows Momentum’s development sequence from 1 to 11. Tested model variants are recorded implementations, including different implementations within the same broader strategy family. Proposed models that did not reach a test are excluded from the tested-model figure.
The evaluation count is a minimum supported by retained research records. It includes assessments under different market, allocation, leverage and cost settings. Repeated archive copies and overlapping all-period summary rows are excluded. It counts testing activity, rather than unique trades, completed customer orders or investment returns.
The market-data figure is 2,710,603 provider data rows recorded in the retained Version 2 evidence report. It describes the data volume of that research run and is separate from the trade scenario evaluation total. It is not a deduplicated lifetime count across all versions.
Some development stages contain research or data-readiness work without a completed trade test. The version count does not imply eleven completed backtests. These figures make no claim about profitability or future outcomes.
YOUR OWN TRADE HISTORY
Explore a historical scenario.
Choose a date range and an amount per trade using your own records.