Phase 1 - Personal backtest
Use the current Strategy Arena backtester now. The local WASM version remains the next technical layer, without server compute.
Strategy Arena Research Network
Citizen science quantitative trading: your GPU, our hypotheses, public papers.
The 3-phase roadmap
Phase 1 gives each researcher a local personal backtest. Phases 2 and 3 turn that base into a benchmark network, then cooperative research raids.
Phase 2 - GPU Benchmark Contest
Standardized hardware leaderboard. RTX 4090, Apple Silicon, Threadripper: speed, stability and reproducibility.
Phase 3 - Cooperative Themed Raids
Raids of 10 to 50 GPUs validate one pre-registered hypothesis. Each raid targets a publishable paper, with contributor credits.
Why it matters
Anti-2CV at scale
Every raid publicly pre-registers its hypothesis before spending compute. Negative results remain part of the record.
Reproducibility native — see desktop client →
Important slices are replicated across contributors; median consensus rejects outliers.
Citizen science
Passive paper trading becomes active contribution: benchmark, validation, audit, then open paper.
Useful precedents
Folding@home
Distributed compute for biology.
SETI@home
Distributed radio signal analysis.
Numerai Tournament
Market-focused data science tournament.
Kaggle
Standardized public ML benchmarks.
Strategy Arena Research Network combines these precedents with one central constraint: pre-registered, reproducible and falsifiable financial research.
Themed raids preview
These raids are public drafts. They are not active today.
| Raid | Hypothesis | Quorum | Status |
|---|---|---|---|
| Bear Market Survivor 2022 | Which strats survived Q2-Q3 2022 crash? | 10 GPUs | Reservable |
| Halving Cycle 3x | Same strat across 2016/2020/2024 halvings? | 25 GPUs | Reservable |
| Brier vs Sharpe correlation | Does ML calibration predict trading edge? | 50 GPUs | Reservable |
| Vol Regime Crossover | What works in VIX>30 vs VIX<15? | 15 GPUs | Reservable |
| Smart Money Stress Test | SME on 36 months walk-forward strict | 30 GPUs | Reservable |
V4 LIVE
Contribute compute now
Help Strategy Arena boost Monte Carlo validations, Atlas refreshes and Hospital scans. Your browser receives throttled micro-workloads lasting a few seconds, stoppable in one click.
Ready. Login required to contribute with your profile.
Session workloads0
Current-
Join the Research Network waitlist
We notify you when Phase 2 opens. We only collect email, GPU model and estimated availability.
Only Phase 1 is live today. Phases 2 and 3 are roadmap, not active implementation. We publish the intent publicly to invite feedback and contributors.
FAQ
Is my data private?
Yes. Computation runs locally; only aggregated metrics needed for consensus are shared.
Will I be paid?
Co-authorship credentials, contributor badges, Pro Researcher credits. No cash/crypto payments for now.
What hardware?
RTX 30/40 ideal, Apple Silicon supported, AMD ROCm coming, CPU fallback for small workloads.
Anti-cheat?
Each critical slice is replicated 3x. Median consensus, outliers rejected, post-raid audit.