Independent validation desk

You have a tradinghypothesis.We validate it.

A backtest that looks good isn’t one you can trust yet. Send us your hypothesis — we run it through an institutional, multi-gate validation gauntlet built to expose overfitting, then hand back an independent report. Typically within 48 hours for Tier A scope.

Submit a brief See a sample report

Validation report

False-breakout 20-day reversal · US equities · Swing

Live sample

Page 1 / …

Open full sample report →

Real PDF from the validation pipeline · sanitized for public release

What you can validate

Backtest almost any trading strategy

Backtesting is how a trading idea earns trust: replaying a strategy over decades of historical market data to measure how it would have behaved — before any capital is at risk. Validraft is built for rigorous, reproducible quantitative research. You don’t need a fully coded algo: discretionary, systematic, and hybrid hypotheses all start the same way — a brief in plain English. We translate your rules into a testable specification, then run it through the same institutional engine with walk-forward and out-of-sample validation by default.

The strategy families we routinely backtest span technical, fundamental, event-driven, and alternative-data approaches — on futures, US equities, FX, and crypto. Pick a family to see concrete example backtests, or describe your own idea and we’ll scope it.

Futures
US equities
FX
Crypto
Cross-asset

Browse all strategy families

Don’t see your exact approach? Discretionary playbooks and partial rules are exactly what the brief is for — we scope what can be tested honestly before any compute runs.

How it works

Inside the validation engine

Every brief travels the same eight-stage pipeline — from immutable, point-in-time data to a signed-off report with full lineage. Step through it, or let it walk you through itself.

Validation pipeline

STAGE 01 / 08

Brief & scope

You describe the asset, timeframe, rules, and what success looks like — in plain English. We translate it into a precise, testable specification and confirm data coverage and feasibility, so the engagement is scoped to what the data lake can genuinely support.
- Hypothesis intake
- Feasibility review
- Tier scoping

Validation framework

Evidence, limitations, and lineage in one audit-ready PDF

Each engagement ends with a structured validation report: what was tested, which controls ran, where the edge held up, where it failed, and what the result does not prove. No black-box score. No recommendation language.

Behind every report sits a scoped validation profile drawn from 66+ controls: data integrity, statistical validation, stress testing, risk measurement, benchmarking, and reproducibility.

See every control we run See a sample report

Control stack

Ordered by what can break a backtest

Hover to scan

Data

Point-in-time inputs

Survivorship, leakage, joins, gaps, and source lineage checked before the run.

Stats

Overfitting controls

OOS, walk-forward, DSR, PBO, haircut Sharpe, null tests, and purging.

Stress

Regime behavior

Crisis windows, liquidity stress, Monte Carlo path risk, and regime splits.

Risk

Performance evidence

Drawdown family, tail risk, attribution, benchmarks, sizing, and limits.

Audit

Reproducible report

Run manifest, versioned inputs, model card, methodology, and limitations.

Phase 0Formalized economic hypothesisEconomic rationale & pre-registration
Phase 0Edge classificationEconomic rationale & pre-registration
Phase 0Ex-ante expected valueEconomic rationale & pre-registration
Phase 0Publication bias & factor decay reviewEconomic rationale & pre-registration
Phase 0Causal robustness checklistEconomic rationale & pre-registration
Phase 1Survivorship-bias correctionData integrity & point-in-time
Phase 1Point-in-time validationData integrity & point-in-time
Phase 1Corporate-actions adjustmentData integrity & point-in-time
Phase 1Data-quality auditData integrity & point-in-time
Phase 1Timezone & holiday alignmentData integrity & point-in-time
Phase 1Pipeline isolationData integrity & point-in-time
Phase 1Investable-universe constructionData integrity & point-in-time
Phase 2 & 3Look-ahead bias detectionSignal & feature integrity
Phase 2 & 3Feature leakage detectionSignal & feature integrity
Phase 2 & 3Overfitting diagnosisSignal & feature integrity
Phase 2 & 3Stationarity tests (ADF / KPSS)Signal & feature integrity
Phase 2 & 3Information Coefficient (IC / ICIR)Signal & feature integrity
Phase 2 & 3Feature stability & redundancySignal & feature integrity
Phase 2 & 3Parameter sensitivity analysisSignal & feature integrity
Phase 3.5Bid-ask spread simulationExecution & cost realism
Phase 3.5Commission & financing costsExecution & cost realism
Phase 3.5Market-impact modelingExecution & cost realism
Phase 3.5Turnover analysisExecution & cost realism
Phase 3.5Capacity analysisExecution & cost realism
Phase 3.5Short-borrow costExecution & cost realism
Phase 4Out-of-sample testingStatistical validation
Phase 4Walk-forward analysisStatistical validation
Phase 4Combinatorial Purged CV (CPCV)Statistical validation
Phase 4Deflated Sharpe Ratio (DSR)Statistical validation
Phase 4Probability of Backtest Overfitting (PBO)Statistical validation
Phase 4Haircut Sharpe (Harvey–Liu)Statistical validation
Phase 4Permutation / null-edge testsStatistical validation
Phase 4Walk-forward permutation testStatistical validation
Phase 4Multiple-testing correctionStatistical validation
Phase 4Autocorrelation (Ljung–Box)Statistical validation
Phase 4Purging & embargoStatistical validation
Phase 4Minimum backtest length (MinBTL)Statistical validation
Phase 4.5Historical crisis scenariosStress testing & regime analysis
Phase 4.5Regime-conditional analysisStress testing & regime analysis
Phase 4.5Regime contract (ex-ante vs ex-post)Stress testing & regime analysis
Phase 4.5Hidden Markov & change-pointStress testing & regime analysis
Phase 4.5Monte Carlo path riskStress testing & regime analysis
Phase 4.5Perturbation & synthetic scenariosStress testing & regime analysis
Phase 4.5Liquidity stressStress testing & regime analysis
Phase 5Risk-adjusted returnsPerformance metrics
Phase 5Drawdown familyPerformance metrics
Phase 5Tail risk (CVaR / CDaR)Performance metrics
Phase 5Distribution shapePerformance metrics
Phase 5Equity-curve stabilityPerformance metrics
Phase 5.5Buy-and-hold benchmarkBenchmarking & attribution
Phase 5.5Alpha / beta decompositionBenchmarking & attribution
Phase 5.5Multi-factor attribution (FF5 + UMD)Benchmarking & attribution
Phase 5.5Information ratioBenchmarking & attribution
Phase 5.5Tail-risk correlationBenchmarking & attribution
Phase 5.5Market-neutral residual testBenchmarking & attribution
Phase 6Ruin & drawdown analysisRisk, sizing & portfolio
Phase 6Position sizing (Kelly / vol target)Risk, sizing & portfolio
Phase 6Risk-parity & constrained allocationRisk, sizing & portfolio
Phase 6Correlation & crowding (internal)Risk, sizing & portfolio
Phase 6Drawdown limits & kill-switchRisk, sizing & portfolio
Phase 8–10Pre-registration & sign-offGovernance, lineage & reproducibility
Phase 8–10Versioned data & codeGovernance, lineage & reproducibility
Phase 8–10End-to-end data lineageGovernance, lineage & reproducibility
Phase 8–10Audit trail of decisionsGovernance, lineage & reproducibility
Phase 8–10Model cardGovernance, lineage & reproducibility
Phase 8–10ReproducibilityGovernance, lineage & reproducibility
Phase 0Formalized economic hypothesisEconomic rationale & pre-registration
Phase 0Edge classificationEconomic rationale & pre-registration
Phase 0Ex-ante expected valueEconomic rationale & pre-registration
Phase 0Publication bias & factor decay reviewEconomic rationale & pre-registration
Phase 0Causal robustness checklistEconomic rationale & pre-registration
Phase 1Survivorship-bias correctionData integrity & point-in-time
Phase 1Point-in-time validationData integrity & point-in-time
Phase 1Corporate-actions adjustmentData integrity & point-in-time
Phase 1Data-quality auditData integrity & point-in-time
Phase 1Timezone & holiday alignmentData integrity & point-in-time
Phase 1Pipeline isolationData integrity & point-in-time
Phase 1Investable-universe constructionData integrity & point-in-time
Phase 2 & 3Look-ahead bias detectionSignal & feature integrity
Phase 2 & 3Feature leakage detectionSignal & feature integrity
Phase 2 & 3Overfitting diagnosisSignal & feature integrity
Phase 2 & 3Stationarity tests (ADF / KPSS)Signal & feature integrity
Phase 2 & 3Information Coefficient (IC / ICIR)Signal & feature integrity
Phase 2 & 3Feature stability & redundancySignal & feature integrity
Phase 2 & 3Parameter sensitivity analysisSignal & feature integrity
Phase 3.5Bid-ask spread simulationExecution & cost realism
Phase 3.5Commission & financing costsExecution & cost realism
Phase 3.5Market-impact modelingExecution & cost realism
Phase 3.5Turnover analysisExecution & cost realism
Phase 3.5Capacity analysisExecution & cost realism
Phase 3.5Short-borrow costExecution & cost realism
Phase 4Out-of-sample testingStatistical validation
Phase 4Walk-forward analysisStatistical validation
Phase 4Combinatorial Purged CV (CPCV)Statistical validation
Phase 4Deflated Sharpe Ratio (DSR)Statistical validation
Phase 4Probability of Backtest Overfitting (PBO)Statistical validation
Phase 4Haircut Sharpe (Harvey–Liu)Statistical validation
Phase 4Permutation / null-edge testsStatistical validation
Phase 4Walk-forward permutation testStatistical validation
Phase 4Multiple-testing correctionStatistical validation
Phase 4Autocorrelation (Ljung–Box)Statistical validation
Phase 4Purging & embargoStatistical validation
Phase 4Minimum backtest length (MinBTL)Statistical validation
Phase 4.5Historical crisis scenariosStress testing & regime analysis
Phase 4.5Regime-conditional analysisStress testing & regime analysis
Phase 4.5Regime contract (ex-ante vs ex-post)Stress testing & regime analysis
Phase 4.5Hidden Markov & change-pointStress testing & regime analysis
Phase 4.5Monte Carlo path riskStress testing & regime analysis
Phase 4.5Perturbation & synthetic scenariosStress testing & regime analysis
Phase 4.5Liquidity stressStress testing & regime analysis
Phase 5Risk-adjusted returnsPerformance metrics
Phase 5Drawdown familyPerformance metrics
Phase 5Tail risk (CVaR / CDaR)Performance metrics
Phase 5Distribution shapePerformance metrics
Phase 5Equity-curve stabilityPerformance metrics
Phase 5.5Buy-and-hold benchmarkBenchmarking & attribution
Phase 5.5Alpha / beta decompositionBenchmarking & attribution
Phase 5.5Multi-factor attribution (FF5 + UMD)Benchmarking & attribution
Phase 5.5Information ratioBenchmarking & attribution
Phase 5.5Tail-risk correlationBenchmarking & attribution
Phase 5.5Market-neutral residual testBenchmarking & attribution
Phase 6Ruin & drawdown analysisRisk, sizing & portfolio
Phase 6Position sizing (Kelly / vol target)Risk, sizing & portfolio
Phase 6Risk-parity & constrained allocationRisk, sizing & portfolio
Phase 6Correlation & crowding (internal)Risk, sizing & portfolio
Phase 6Drawdown limits & kill-switchRisk, sizing & portfolio
Phase 8–10Pre-registration & sign-offGovernance, lineage & reproducibility
Phase 8–10Versioned data & codeGovernance, lineage & reproducibility
Phase 8–10End-to-end data lineageGovernance, lineage & reproducibility
Phase 8–10Audit trail of decisionsGovernance, lineage & reproducibility
Phase 8–10Model cardGovernance, lineage & reproducibility
Phase 8–10ReproducibilityGovernance, lineage & reproducibility

Post-validation monitoring

Validation continues after delivery

A strategy that passed validation yesterday can drift tomorrow. For ongoing engagements, we keep watching rolling Sharpe, drawdown, signal decay, and regime behaviour — benchmarked against the same gates that produced your report.

Alerts and history live in your client workspace, with a direct line back to the desk when something needs a human read.

Open workspace See pricing

Client workspace · preview

Opening-range breakout — NQ

Healthy

Sharpe (full): 1.42
Sharpe (30d): 1.31
Max DD: −8.4%

Illustrative metrics — your live workspace reflects each delivered engagement.

Rolling Sharpe

30-day Sharpe tracked against the validation baseline. Watch and critical bands when performance drifts.

Drawdown tracking

Graded alerts as peak-to-trough loss approaches stress-test envelopes from your report.

Signal decay

Second-half vs first-half performance comparison — early warning when the edge starts to fade.

Regime shifts

Conditional correlation and behaviour across market states — flagged when regime context changes.

Data

Three decades of institutional market data

Every brief is tested against the datasets your hypothesis actually needs — price, fundamentals, macro, sentiment and alternative signals, on the same immutable, point-in-time lake our research desk runs. Coverage windows reflect what’s on disk today.

1990200020102020Now

Databento
2012
Massive
2003
FMP
1990
FRED & ECB
2000
News + FinBERT
2015
Quiver
2023

Price
Fundamentals
Macro
Sentiment
Alternative

Price

CME futures

since 2012

Continuous CME futures, point-in-time across calendar and volume rolls — from minute bars to full L2/L3 order book.

ESNQCLGCSINGRTYYMZBZNZTMESMNQ

OHLCV 1-minute bars · 13 continuous contracts2012
Trades — tick · full trade history2025
Order book — MBP-10 & MBO · ES, NQ2026
Definitions, statistics & status · contract metadata2026

Price

US equities

since 2003

The full US equity cross-section — trades and minute bars back to 2003, with fundamentals and corporate actions.

Trades — daily cross-section · ~10K US tickers2003
Minute aggregates · raw + corporate-action adjusted2003
Corporate actions · splits, dividends, IPOs
Fundamentals · balance sheet, cash flow, income, ratios, float, short interest
Reference universe · tickers & classifications

Fundamentals

Fundamentals & multi-asset markets

since 1990

The widest feed in the lake: company fundamentals, estimates, ownership, index membership, FX, crypto and macro.

Company fundamentals · 19 statement types · ~8,900 companies2003
Earnings · calendar & surprises2003
Corporate actions · dividend & split calendars2003
Analyst coverage · estimates, grades & ratings · 5–8K symbols
Insider & institutional · Form 4 + 13F · ~8K / ~7.5K symbols
Index constituents · S&P 500, Nasdaq, Dow + historical
ETF holdings & weightings · 33+ ETFs
FX spot — EOD · 15 major pairs
Crypto — EOD · 10 assets
COT positioning · 64 futures markets2003
Treasury rates · US yield curve1990
Economic calendar · global releases2015

Macro

Macroeconomic indicators

since 2000

Point-in-time macro: 230+ vintage series across the US, Europe, UK and Japan — no look-ahead on revisions.

Macro series — vintage · 230+ point-in-time series2000
Regions · US, Eurozone, UK, Japan, Global
Treasury & policy rates · via FRED / ALFRED

Sentiment

News & sentiment

since 2015

News and press releases scored for sentiment with FinBERT — five streams across equities, FX, crypto and macro.

Stock news2015
Press releases2019
Forex news2018
General macro news2020
Crypto news2021
FinBERT sentiment scoring · transformer NLP, full universe

Alternative

Congressional trading

since 2023

US congressional trading disclosures as an event ledger plus point-in-time holdings snapshots.

Congressional trades — event ledger2023
Holdings snapshots · point-in-time2026

Don’t see your market?

Coverage reflects the current research lake. If your hypothesis needs a feed we don’t hold yet, we scope it as part of the brief before any work begins.

Submit a brief

Provider names and marks are shown for identification only. Validraft is independent and not affiliated with or endorsed by the listed providers. Coverage shown reflects the current research lake; some feeds refresh on demand for scoped engagements.

Positioning

Rigorous without the grind. Independent without the boutique markup.

Validraft

Human-scoped validation with feasibility review upfront
48-hour turnaround for standard Tier A hypotheses
Confidential — your idea never enters a public template library
Same infrastructure and gates as our internal research desk

DIY backtest tools

You own the full pipeline — and every lookahead bug
No deflated Sharpe, PBO, or crisis stress unless you build them
Easy to overfit; hard to know when you have

Generic quant consultants

Opaque methodology and inconsistent gate coverage
Prescriptive language that blurs research vs advice
Weeks of back-and-forth before a deliverable

Security & confidentiality

Your edge is yours. We treat it that way.

You are handing us a trading hypothesis — often the most valuable thing you own. Validraft is built around a simple rule: your strategies and client data are never shared, never pooled, and never repurposed.

From the first brief to the final PDF and any ongoing monitoring, every touchpoint is scoped to your engagement and accessible only to you and the desk team running the work.

Your hypothesis stays yours

Every brief, parameter set, and deliverable remains your intellectual property. We claim no rights to your idea — ever.

Per-engagement isolation

Each validation runs in a scoped workspace: your brief, runs, and reports are tied to your engagement — not mixed with other clients.

Never enters a public library

Client strategies are never published, templated, or reused for other engagements. What you send us stays between you and the desk.

Gated client workspace

Reports and monitoring live behind authenticated access in your private workspace — not on open links or shared folders.

Operational practices

Encrypted in transit (TLS) for brief intake, dashboard access, and report delivery
Least-privilege access — only the assigned desk team sees your engagement materials
No cross-client benchmarking, aggregation, or anonymised reuse of your logic
NDA available before material exchange on Tier B/C and custom scopes
Descriptive outputs only — research and simulation, not investment advice

Research notes

Learn how robust validation actually works.

Plain-English essays on overfitting, validation gates, data discipline, and the difference between an attractive simulation and a durable hypothesis.

View blog

First article

01 / Overfitting

The curve can be beautiful and still be wrong.

Backtest overfitting

How to Tell If Your Backtest Is Overfit

A practical checklist for spotting curve-fitting before a beautiful equity curve becomes an expensive mistake.

June 12, 20268 min read

Read article

Your hypothesis. Independently validated.

Send a brief. We confirm scope, run the validation, and deliver the report. Research and simulation only.

Submit a brief Client dashboard