THE PLATFORM
Show your work.
Financial markets move fast.
Conviction should come with evidence.
Meet Meridian: market monitoring, specialist research and portfolio analysis connected to an auditable record.
A 16-slide presentation from Meridian Strata
THE PLATFORM
48 capabilities. One accountable workflow.
The case for Meridian rests on how much of the research process stays connected.
- ObserveCapture the evidence
- AnalyzeCalculate and challenge
- DecideRecord the rationale
- EvaluateMeasure what followed
The 48 capabilities share modules and infrastructure. They are not 48 independent services.
THE PLATFORM
Meet MTAME.
MTAME is Meridian’s Technical Alpha Mastery Engine: the shared framework that calculates technical indicators and helps research agents interpret them consistently.
- Calculate
- Turn timestamped price and volume observations into 50 defined indicators, including RSI, MACD and ATR.
- Check
- Flag missing data and insufficient history. Use completed observations available at the decision time.
- Interpret
- Group overlapping signals, account for market conditions and preserve disagreement across timeframes.
50 indicators, 7 analytical families
Its output is traceable technical evidence for the committee. Alpha Proof tests whether that evidence improves decisions.
Market monitoring connects this evidence with quotes, attributed news and portfolio context. Source coverage and freshness remain explicit.
THE PLATFORM
Research beyond the ticker.
Business economics
Financial forensics examines cash conversion, accruals, working capital and dilution. Catalyst research connects events to an economic mechanism.
Seven valuation methods
Earnings multiple · Book multiple · Enterprise multiple · Sum of parts · NAV scenario · Dividend discount · Discounted cash flow
Portfolio consequences
Portfolio Forge evaluates allocations, capital use, concentration and risk under explicit constraints.
Derivatives depth
Options chains, volatility structure, multi-leg payoffs and scenario analysis connect a thesis to instrument economics.
THE PLATFORM
A committee, with receipts.
Independent perspectives become a decision record the system can revisit.
Discovery & sentiment
Narratives, catalysts and contrary evidence.
Technical research
Market structure, setups and invalidation.
Fundamental research
Business quality, valuation and financial risk.
Securities strategy
Instrument choice and derivatives implications.
CIO synthesis
Final judgment with material dissent retained.
These are AI research roles. The completed historical study used five roles. A sixth contrarian role remains subject to certification and release gates. Recommendations do not bypass user authority over financial actions.
THE PLATFORM
Serious research needs a serious backend.
Research, portfolio controls and progress views.
Typed contracts, ownership checks and orchestration.
Research, ledger, decisions, approvals and audit events.
Separate research, live-paper and data workloads.
Versioned inputs, source lineage and usage receipts.
Persistent services and controlled maintenance.
Durable records outlive browser sessions and queue retries. Provider health and research readiness remain separately verified states. No uptime or scale benchmark is asserted.
ALPHA PROOF
A backtest is easy.
A fair test is harder.
Alpha Proof asks a narrower, tougher question: does richer analytical information improve the committee’s decisions after costs?
63 sessions per historical window
12 securities in a fixed universe
160 commitments retained for scoring
Completed retrospective development study. Each arm starts with $100,000 simulated capital, up to five holdings and 35% per security. It is not live-money performance.
ALPHA PROOF
Four arms. One question.
Alpha Proof holds the committee’s task fixed and adds layers of the Technical Alpha Mastery Engine (MTAME).
Ordinary indicators
A reconstructed baseline for comparison.
MTAME calculations
Shared indicators with defined formulas and data checks.
Regime selection
Choose relevant evidence for the prevailing market conditions.
Full MTAME
Account for overlapping signals and conflicts across timeframes.
Arm A reconstructs ordinary indicators; it is not an archived pre-MTAME production release. No automatic promotion follows a favorable backtest.
ALPHA PROOF
The result we refuse to hide.
Full MTAME did not beat the baseline on average net return in this completed study.
Full MTAME versus baseline
Arm D averaged 0.873%; arm A averaged 0.958% under base costs.
A useful platform should make this result impossible to bury.
40 repeat windows across 20 market periods. Arithmetic means, not annualized returns or one continuous portfolio. All three cost scenarios remain visible.
ALPHA PROOF
Lower exposure changed the risk profile.
The full system traded less and carried less market exposure. Its smaller drawdown comes with that context.
| Metric | A · Baseline | D · Full MTAME |
|---|
Drawdown is the mean of each window’s maximum drawdown. It is not the worst loss of a continuous strategy. These figures do not establish risk-adjusted outperformance.
ALPHA PROOF
The episode matters.
Basis points per 63-session window · base costs
A positive average in range-bound periods coexisted with negative averages elsewhere.
The full system changed decisions. This sample does not prove those changes improved investment outcomes.
Bullish: 8 periods. Range-bound: 8. Bearish: 4. Two repeats per period. These descriptive subgroups are too small to validate a regime-trading rule.
ALPHA PROOF
800 reports are still
only 20 market periods.
More model output does not create more independent market evidence.
The registered dependence supplement forms four chronological groups of five periods.
Its method requires at least twenty periods in each of four groups before inferential reporting.
Eighty periods is a sample-size floor, not proof of independence. Retrospective data, a fixed-symbol universe, corporate-action gaps and possible model hindsight still matter.
ALPHA PROOF
Evidence survives the verdict.
Reproducible accounting
Independent Decimal reconciliation checks saved decisions and simulated accounting. Scoring and archival verification required no new model runs.
Learning with constraints
Thirteen findings have recorded dispositions and completed CIO review. Economic changes remain held for fresh evidence.
“Independent” refers to a separate implementation inside the project, not external auditor assurance. Same-server retention is not an off-server recovery system.
ALPHA PROOF
The next claim needs the next test.
Monthly market periods
A separately registered 21-session study, with four arms, two repeats and up to 3,200 reports.
Registered follow-up · incompleteForward weekly cohorts
Fixed five-session cohorts scheduled from November 2026 through November 2027. Missed forecasts remain missed.
Prospective program · outcomes pendingFresh evidence must earn a production change. A new experiment cannot rewrite an old result.
These are distinct programs with different horizons and universes. This presentation reports no performance results from either follow-up and does not pool them with the completed study.
THE STANDARD
The ambition: the world’s most
accountable research platform.
Meridian’s case is the connection between analytical depth and the ability to inspect the work.
Broad enough to investigate
Technical, fundamental, derivatives and portfolio analysis in one research workflow.
Structured enough to challenge
Specialist perspectives, source lineage and recorded disagreements.
Honest enough to improve
Experiments that retain failures and withhold an edge claim when evidence falls short.
Conviction is easy. Accountability is the standard.
This is Meridian’s ambition and design thesis. No independent global comparison establishes “best,” and no investment edge is certified by the completed historical study.
THE STANDARD
The evidence behind the presentation.
Read the public aggregate
Download the exact tables, definitions, study limits and source fingerprints used in this presentation.
Download evidence JSON ↗What this edition establishes
Documented capabilities. A completed historical experiment. Reconciled simulated accounting. Retained adverse findings.
What remains open
Prospective investment edge, independent global ranking, universal provider availability and production activation of staged changes.
Selected source documents describe staged remediation. Their tests do not certify every live workflow. Private records and licensed raw data remain private.
Evidence snapshot: October 31, 2026. Public aggregate disclosure from Meridian’s retained internal records. Simulated historical results are development evidence.