The Mana research engine

From observations
to strategy-level review.

Shared infrastructure. Market-specific models. A path from the research summary back to individual sessions.

01 / MANA

The research environment

The proprietary engine, knowledge base and shared engineering infrastructure.

02 / A103

A selected application

The selected ES and GC intraday strategies documented in the research report.

03 / R1

A portfolio calculation

The capitalisation and monthly sizing methodology applied to A103.

Multi-criteria selection

One engine.
Different selection priorities.

Mana compares strategies and combinations from several perspectives. The aim is to understand which trade-off fits the research objectives, rather than select on gain alone.

01 / PRIORITY

Performance

Start with the highest-gain combinations in the research universe, then examine their costs, net P&L and drawdowns.

  • Gain
  • Net P&L
  • Costs
  • Max drawdown

Key questionWhat setbacks accompanied the gain?

02 / PRIORITY

Drawdown & consistency

Prioritise combinations with smaller observed drawdowns and more consistent annual results. Examine the weakest year alongside the overall gain.

  • Max DD: points & %
  • Positive years
  • Weakest year

Key questionHow deep were the declines, and how consistent were the years?

03 / PRIORITY

Return / risk balance

Use Gain/DD and Sharpe alongside annual consistency and net P&L to compare the trade-off between gain, setbacks and day-to-day variability.

  • Gain / DD
  • Sharpe
  • Annual consistency
  • Net P&L

Key questionWhat trade-off does the combination offer?

These are selection priorities. Capital and position sizing are assessed separately, using consistent comparison assumptions.

How the criteria are interpreted

Inside the process

Follow a result back to its source.

Open each stage to explore its inputs and the questions considered during review.

Process overview. Four stages connect market observations to the examination of individual sessions.

01Market observations

Start with the instrument and its trading session.

Sessions, time zones and instrument-specific conventions frame the observations used in the research environment. Data integrity and the available observation window matter before a model is compared.

REVIEW INPUTS

  • Instrument and market
  • Session and time zone
  • Available observations

OutcomeA documented research scope

Review focus: available sessions, data integrity and coverage.

02Candidate strategies

Study configurations within each market.

Variables, interactions and parameter configurations are studied within a common engineering framework. Ongoing calibration of candidates is separate from changes to a selected reference strategy.

REVIEW INPUTS

  • Strategy identity
  • Market-specific parameters
  • Research constraints

OutcomeIdentifiable candidate configurations

Review focus: configuration identity, market-specific parameters and calibration choices.

03Results and combinations

Read the gain alongside the risk.

Inspect daily and period results, indicative costs and drawdown. For a combination, examine its components as well as the total, under explicit capital and observation conventions.

REVIEW INPUTS

  • Daily research results
  • Costs and capital assumptions
  • Components and periods

OutcomeA result with its calculation context

Review focus: gain, drawdown, costs and contributions across periods.

04Session replay

Return to the trading activity behind a result.

Replay a historical session or selected dates, examine entries and exits, and review the position details behind the daily result.

REVIEW INPUTS

  • Selected configuration
  • Historical session
  • Entries and exits

OutcomeA detailed session-level review

Review focus: the price path, timestamps and position outcomes.

Session-level examination

From the daily result
to the position details.

Replay individual historical sessions or selected dates and examine the trading activity behind each result.

What a focused demonstration can cover

  • The selected strategy and its research context.
  • Daily results and a chosen historical session.
  • Entries, exits and position details in the replay.
  • The relationship between individual operations and the daily result.

What stays private

Signal logic, calibrated parameters and internal knowledge structures remain proprietary. Public captures show the analysis tools with confidential settings masked.

Request a research-engine discussion →

Research discipline

Questions to ask of every candidate.

The research framework examines stability, traceability and risk across periods and operating assumptions.

Across periods

Regime-based evaluation and chronological out-of-sample or walk-forward methodologies.

Across assumptions

Parameter sensitivity, indicative transaction costs and adverse execution or liquidity conditions.

Across the process

Data-integrity checks, monitoring and post-session review within explicit operating constraints.

The A103 report sets out the scope and calculation assumptions of its published results.

Ongoing research · Reference configurations

Continuous research. Identified reference configurations.

The A103 research documentation records strategy selection in late 2025 and the same trading configuration for the January 2026–October 2026 study period.

New observations and candidate-model calibration enrich the knowledge base. Research on candidates runs alongside the monitoring of identified reference configurations.

Research scope

One infrastructure.
Separate market programmes.

Trading sessions, time zones and operating conventions are part of the common environment. Models are evaluated separately.

BROADER RESEARCH UNIVERSE

NQ · CL · FX · K200 · HSI · ZN

Additional research programmes include EUR/USD and GBP/USD. Each instrument has its own evaluation; the published A103 report covers ES and GC.

Professional discussions

Look beyond the performance table.

Discuss the engine, examine the methodology or explore a technology collaboration. Supplementary materials are shared case by case, subject to confidentiality.

Start a conversation