SKU 04 / Quantitative research desk
We run an institutional-grade backtesting operation: strategy ideas coded, tested against 16 years of futures data, down to tick resolution, validated out of sample, and stress tested until they break or earn trust. You bring the idea. We prove it or kill it, and show you which one happened.
4,083
Backtests cataloged
269
Research projects
38
Registered algorithms
16 yrs
Futures data down to tick level
| Program | Mandate | Evaluation | Result | Status |
|---|---|---|---|---|
| MERIDIAN | Multi-asset risk-parity futures portfolio | 2020–2026 blind | Sharpe 2.29, max DD 6.4% | VALIDATED |
| HELIOS-3 | Dynamic-exposure index futures program | 2021–2026 + OOS | Sharpe 1.72, PSR 79%, OOS +10.0% | VALIDATED |
| TRIAD | Three-sleeve futures portfolio | 2020–2026 blind | Sharpe 1.61, +525% cumulative | VALIDATED |
| SENTINEL-4 | Adaptive index futures system, current generation | refreshed 2026-07 | Sharpe 1.45, PSR 81%, DD 9.7% | VALIDATED |
| AURORA | Equity session-momentum system | 2012–2015 frozen-parameter | Sharpe 2.11, PSR 98% | RETIRED |
The curves behind the table
These are the equity curves produced by the engine backtests behind the ledger above: indexed to a starting value of 100, plotted on a log scale. Simulated performance, same disclaimers as everything else on this page.
MERIDIAN and TRIAD are portfolio constructions that blend underlying program sleeves, so no single engine run produces their curve; the standalone program curves are shown here. All curves are backtested, hypothetical performance. See the disclaimer below.
Three ways to use the desk
Quant 01 / Backtest Build
Your strategy idea, implemented cleanly on QuantConnect and LEAN and run against institutional-quality data. You get the full tearsheet: returns, Sharpe, drawdown, trade log, exposure profile.
Quant 02 / Validation Suite
Everything we run on our own programs before trusting them: out-of-sample windows, walk-forward regime testing, parameter stability sweeps, slippage stress at multiple levels, and probabilistic Sharpe reporting.
Quant 03 / Backtest Audit
You have a backtest from a developer, a vendor, or your own research and the numbers look too good. We audit it adversarially: overfitting forensics, look-ahead and survivorship checks, cost realism, and reproduction attempts.
The research protocol
Every algorithm version is committed and tagged before its backtest fires, byte-exact. Results always trace to the exact code that produced them.
Runs execute on QuantConnect and LEAN against 16 years of data at the resolution the strategy needs, down to tick, and land in an append-only catalog. A test is never quietly re-run until it says something nicer.
Out-of-sample windows, blind multi-year evaluations, frozen-parameter pre-period tests, and cross-market checks separate signal from curve fitting.
Slippage stress at multiple cost levels, parameter stability sweeps, and reproducibility reruns. Fragile results get labeled fragile.
Validated results ship with their full context. Results that fail later interrogation are retracted on record, not buried. You get the same honesty as a client.
The fine print, stated plainly
All performance figures on this page are backtested or simulated results from internal research programs. Backtested performance is hypothetical: it does not represent actual trading, involves the benefit of hindsight, and does not guarantee future results. VerifiedDesk sells research and engineering services only. We do not provide investment advice, trading signals, portfolio management, or access to any trading system, and nothing on this page is an offer or solicitation to invest. Whether and how to trade any strategy is entirely your decision and your risk.
Start the interrogation
Describe the strategy in plain language, at whatever level of detail you are comfortable sharing. You get back a scoped plan with a fixed price, usually within 24 hours. Your idea stays yours: we sign NDAs and deliver the runnable project at handover.