Skip to content
QEV

QEV examines whether a strategy's evidence survives. Below is a completed sample examination. Open the tabs, read the evidence, see the verdict.

Start your examination
QEV WorkspaceSample examination

Sample examination · Historical data · Not investment advice

BTC Funding Mean Reversion

BTC-USDT perpetual futures · funding-rate reversion

Frozen before any run. Rules, universe, window, costs, and the failure criterion are written down first.

  • UniverseBTC-USDT perpetual, single venue, no leverage stacking
  • Window2020-09 → 2025-06 · out-of-sample boundary frozen at 2024-03
  • ExecutionNext funding timestamp after signal · taker fees 4 bps per side
  • Failure criterionNet-of-cost Sharpe must exceed 0.8 OOS with max drawdown under 25% to count as reproduced

Data quality

2 conflicts

Out-of-sample

Passed

Parameter stability

Weak

Execution costs

Material

Verdict

Keep researching

The same examination, as a document

Read the full report →

EX-001 · EXAMINATION REPORT · SAMPLE EXAMINATION · HISTORICAL DATA · NOT INVESTMENT ADVICE

BTC Funding Mean Reversion: does the funding-reversion effect survive independent reconstruction?

Specification

Enter short perp exposure when 8h funding ≥ +0.10% (annualized ≥ 109.5%); exit at funding ≤ +0.03%. Mirror for negative extremes. 2020-09 → 2025-06 · out-of-sample boundary frozen at 2024-03

Data conflicts

2 conflicts — missing funding stamps (Mar 2021) and a divergent open-interest snapshot during a venue outage (Nov 2022). Resolutions recorded in the snapshot manifest.

Costs & assumptions

Taker fees 4 bps per side plus slippage; funding payments are the strategy's raw PnL and are never netted away. Costs consume 38% of gross capture in the base case; break-even at 9 bps per side.

Out-of-sample & robustness

PASS — effect holds beyond the frozen boundary; block bootstrap shows no tail blowup.

WEAK — parameter stability fails the comfort bar: the edge concentrates in a narrow entry-threshold band.

MATERIAL — costs are large relative to capture; capacity decays under crowding.

out-of-sample →20212022202320242025100
Examined strategyBuy-and-holdRandom-signal bandFigure 1 · Illustrative

Final status

KEEP RESEARCHING

What QEV tests

Full validation method →
Data integrity
Vendor reconciliation, adjustment audits, survivorship and point-in-time checks. The input must represent the real market before any result counts.
Look-ahead & leakage
Every rule is probed for information it could not have known at decision time — shifted signals, spec drift, boundary violations.
Costs & execution
Fees, spread, slippage, funding, turnover. Statistical edges must survive realistic trading costs and report their break-even level.
Parameter stability
The neighborhood around published settings, not one tuned point. Fragile thresholds are named as open risks.
Market regimes
Subperiods, conditions, structural shifts. Single-era luck is separated from durable behavior.
Monte Carlo & stress
Block bootstrap, scenario paths, adverse excursions — with the randomized assumptions stated.

Figure 2 · Parameter neighborhood

drag to rotate

Figure 2. Net Sharpe across the entry-threshold × holding-window neighborhood for the sample examination. The narrow cobalt ridge is the weak-stability finding: outside the 0.08–0.13% band, the edge decays fast.

Passing one test cannot pay for failing another. Gates stop at the first failure.

The research workflow

One continuous path from claim to monitored paper strategy. Every stage has an owner, a record, and a gate. Nothing moves forward on assertion.

Paper-only · no live orders · human gate

  1. Idea

    A paper, a rule set, a backtest, a link. Submitted through a standardized form — no code uploads, no credentials.

  2. Specification

    Frozen in writing before any code runs: rules, universe, window, costs, failure criterion.

  3. Evidence

    Checksummed snapshots, versioned deterministic runners, retained artifacts. Every number traceable to its bytes.

  4. Validation

    Replication, controls, costs, out-of-sample, stability, regimes, stress. Gates run cheapest-first and stop at the first failure.

  5. Paper monitoring

    Surviving strategies enter monitored paper study with decay and regime alerts. No live orders, ever.

Four surfaces, one record

QEV is one system with four working areas. qev.io demonstrates the desk; lab.qev.io will carry the authenticated environment behind login.

Live demo above

Research Desk

Submit and follow examinations as they run — the working surface shown at the top of this page.

Demonstrated on this page. Authenticated desk arrives with lab.qev.io.

  • Standardized claim intake: paper, link, or rule set
  • Live pipeline: specification, data check, validation, verdict
  • Evidence log with every finding as it lands

Methodology and trust

11

validation gates per examination, run cheapest-first

4

bounded verdicts — nothing in between, no alpha score

0

live orders, ever — paper-only with a human gate

Reproducible runs
Same specification, same snapshot, same code version — same numbers. Deterministic by construction, not by promise.
Versioned evidence
Every report keeps its history. Corrections are visible, never silent.
Independent validation
The maker never approves their own result. A checker with a falsification mandate signs every finding.
No automatic live trading
Paper-only. No broker authority, no execution path, no exceptions — with human confirmation before anything consequential.
Clear limitations
Every report states what it does not establish. Disclosures are part of the document, not a footnote.

Disclosure. QEV is a research and examination service. It is not investment advice, a broker, or an adviser, and it does not recommend or execute trades. The sample examination on this page uses historical data and is not investment advice. Examination reports describe what evidence survived stated tests at a stated time; they do not predict future results.

Pricing and access

Final prices publish at public launch. Founding-queue requests lock launch pricing.

Standard examination

Individual traders and evaluators

Full gate battery to a bounded verdict. Versioned, downloadable, permanent report.

Per claimRequest

Priority examination

Professionals on a deadline

Front-of-queue scheduling and direct clarification during the specification freeze.

Per claim · premium laneRequest

Research workspace

Research teams, family offices

Parallel examinations, team workspace on lab.qev.io, shared version history and read API.

Annual · scoped APIRequest

Start your examination

Requests are reviewed by hand. Testable claims get a scope reply with data needs and queue position.

Request access