The research platform is live and free right now. The signal service opens when strategies actually clear the bar — not before. Here is exactly where things stand.
Ten years of 1-minute bars across the S&P 500, Dow 30 and Nasdaq-100, 100 crypto pairs and 44 FX pairs — roughly 675 million rows on one spine, aggregated to eight timeframes.
A strategy needs a t-statistic of 3.0, not 2.0, on at least 30 trades. The threshold was fixed before any testing started, so it cannot be quietly relaxed to let a favourite through.
Nineteen hypotheses tested, nineteen killed. Cross-sectional momentum inverted. Carry failed at t = 0.98. Six classic technical entries came back significantly negative. That is the honest state of the research.
Not a summary of our work — the actual workstation. Write strategies as formulas, screen the whole market with them, and run the same backtester against the same data we use.
Fills at the next bar's open, spread and fees on both sides, and stops assumed to hit before targets when a bar could have done either. Optimistic assumptions are how backtests lie.
No book, no track record, no frozen portfolio — yet. When a strategy clears the bar it gets published with its full test history, benchmark included, and the losers stay on the page.
A hypothesis is written down before it is tested, with the threshold fixed in advance. It runs on data it was never fitted to, gets charged real costs both ways, and is checked for whether one lucky instrument is carrying the whole result.
A t-statistic of 3.0 on at least 30 trades, or it does not ship. Nineteen hypotheses have been through this so far and nineteen were killed — several of which looked excellent until the costs went in.
Whatever clears goes up with its full test history and its benchmark beside it. Until something does, there is nothing to sell — and the platform stays free either way.
hypotheses in,
survivors out
t ≥ 3.0 required
Entry or exit
Real venue price
Per-strategy rules
Telegram
Email
Dashboard
Manual or
Automated
Custom allocation
Exit rule fires
signal or time based
capital recycles
Ten years of 1-minute data across the S&P 500, Dow 30, Nasdaq-100, 100 crypto pairs and 44 FX pairs. Build a strategy, scan the whole market with it, and get an honest verdict on whether it actually works.
platform-shot.png in the repo to show the real thingCharts and analysis need no account. Screening and backtesting need only an email. There is no paid tier.
There is no subscription because there is no validated book. Nineteen hypotheses have been tested and nineteen were killed. When one clears a t-statistic of 3.0 it gets published — with its full test history, its benchmark, and its drawdowns — and only then is there anything worth charging for.
Ten years of 1-minute data, 36 indicators, eight timeframes, a formula language, a market screener and a backtester that reports a t-statistic instead of a highlight reel.
Charts and automatic analysis need no account. Screening and backtesting need only an email.
If and when strategies clear the bar, the signal service opens. You will see the evidence before you see a price — the full test history, the benchmark it was measured against, and the losers alongside the winners.
No charge, no commitment. We will tell you what cleared and what did not.
Automated execution on your own accounts across stocks, FX, and crypto — regulated venues, your API keys, your control.
Sign up now for free beta access. Subscriptions and promotions start August 15, 2026.
Yes — two ways. Every plan delivers algorithmic trading signals with exact entries and exits you can act on yourself. Add the Automation add-on and TradeQRX becomes hands-free auto trading on your own brokerage accounts: Alpaca for stocks & ETFs, OANDA for forex, Kraken for crypto. You set the allocation per trade; the engine handles entries, exits, and sizing.
The engine is machine-driven, not vibes-driven: strategy discovery searches hundreds of thousands of rule combinations per asset, and every candidate must survive double-holdout validation, statistical luck filters, and real-fee stress tests. Searching broadly is exactly what makes a backtest optimistic: test enough rules and some will look brilliant on luck alone. That is why the threshold is fixed before testing, and why nothing has been published yet.
US equities (S&P 500, Dow 30 and Nasdaq-100), 100 crypto pairs and 44 FX pairs, all on 1-minute data back to 2016. No futures — we only sell signals for markets our engine actually trades.
Yes — the Daily 5: up to five live BUY / HOLD / SELL stock & ETF signals every trading day at 10:00 ET, on the web and by email. The platform is free to use today: charting and automatic analysis need no account, and screening and backtesting need only an email. No card required.
Most trading bots run one untested strategy until it blows up. TradeQRX has no live book and no track record yet. The platform is a research tool; when a strategy does clear the bar it will be published with its full history rather than a highlight reel. Automation executes on your broker accounts with trade-only API keys; we never hold your funds.