Verification Hub
Every auto-execution placed through the OKX connection logs the gap between its signal price and its actual fill price. We publish the aggregate. No per-user attribution — just evidence. Our own bot runs on a separate path and is not counted here.
We run it ourselves — on paper, in public.
The strongest proof isn't a claim — it's a record. Below is the git-timestamped paper track, published as measured, win or lose. Its opening stretch is what we required of ourselves before risking real money. We do not publish information about real-money operation until the feature is complete. The track below is the paper one. The record keeps accumulating past that point, and we publish it either way.
We Run It Ourselves. Here's the Real Record.
We don't just backtest. This git-timestamped paper record (Jun 26, 2026 – Oct 6, 2026) keeps accumulating, and it is the track we publish — we do not publish information about real-money operation until the feature is complete. The figures below are the published snapshot of that window, preserved as measured, win or lose.
Lower exposure = the strategy de-risked into cash (crash avoidance, as designed).
Every Paper Track We Run — Win or Lose
Not just the winner. Every paper track we run, including the ones losing money right now. That's the whole point of "Don't believe — verify."
Trades (24h)
—
Trades (7d)
—
Slippage median
—
|fill − signal| / signal
Slippage p90
—
worst 10% bound
SL / TP set-rate
—
% of logged trades where OKX accepted our stop-loss + take-profit algo order. Misses trigger the emergency-close path (autotrader lesson L1+).
loading…
PLATFORM HEALTH
Live system status
If the API is down or data goes stale, we say so here — no PR spin.
API status
—
Version
—
Coins loaded
—
API uptime
—
Data age
—
Source: live api.pruviq.com/health
THE HONEST SCOREBOARD
We tested 22 strategies. Most failed validation.
We don't only show the winners. Every strategy we've researched — including the ones we killed — is counted here. Failure is published, not hidden.
1
Live
The published track is the paper one (trend ensemble) — verified by a record that accumulates daily, not by the backtest battery. We do not publish information about real-money operation until the feature is complete.
1
Conditional
edge only in specific regimes — the condition is published on the card
6
Still testing
live paper, no verdict yet
12
☠ Retired
failed validation — retired in public
2
Shelved
set aside — the reason is published on each card
“Verified” means reproducible out-of-sample, not a profit guarantee. The honest answer to most markets is disciplined risk management — not a magic edge.
Research Transparency
In everything we tested, no easy edge survived.
Most crypto sites promise “this strategy = profit.” We ran the opposite experiment — we tried to find a tradeable edge ourselves, and we publish what we found.
3.6M+
data points tested — all as of the 2026-06 sweep (238 OKX coins, 2.5-year window), BTC/ETH 9-year multi-cycle & 5-year options
17
strategy families — price, volume, basis, volatility & options, AI/ML, on-chain
0
market-beating edges survived realistic cost + regimes + overfitting (risk premia like trend & volatility persist — that's paid risk, not free alpha)
Statistical correlations exist, but once you apply realistic trading costs, test across bull/bear regimes, and guard against overfitting and look-ahead bias, the edge disappears. Peer-reviewed research agrees. So we don’t promise profits — we give you the tools and our honest findings, and you verify for yourself. Don’t believe. Verify.
Open the full research ledger — every verdict, published →
SO — WHAT SHOULD YOU DO?
No magic strategy exists. So here are two honest moves.
1 · Verify, don’t believe
Any “guaranteed” strategy — run it yourself first on real data. Our simulator is free, no signup. If an edge can’t survive your own backtest, it won’t survive your money.
Test a strategy →2 · Auto-trading · Coming soon
Auto-trading is coming soon — real-money execution is on hold. For now, you can test strategies in the simulator.
Check readiness →
DISCLOSURE & LEDGERS
Who pays us, and where to audit every claim
One revenue source, publicly disclosed — and every ledger we keep, linked from one place.
Methodology
Formulas, execution assumptions, what we don't model, and how to reproduce any number.
Open →Failed strategies
Every retired preset and the full out-of-sample registry sweep — losses published, CSV included.
Open →Forward pre-registration ledger (raw JSON)
Claims registered before outcomes, append-only, gate-enforced in CI.Last observation 2026-10-06 · 17 of 52 days since 2026-08-16 have no entry (missed days are not back-filled).
Open →Pre-registered forward tests
Rules fixed and timestamped (OpenTimestamps) before any forward data, with each file's hash and how to check it.
Open →For AI agents and developers
Every machine-readable file, the fields to cite, and how to re-run a verdict without trusting us.
Open →Changelog
What changed and when — including corrections.
Open →The conflict, stated plainly. Referral commissions come from users who trade — so we have a structural incentive to encourage trading. That is exactly why the verdicts on this site are enforced in code, not editorial: failed strategies stay published, scoreboard counts are computed from source files at build time, and forward claims are pre-registered before their outcomes exist. Pages that show verdicts carry no referral link, and neither do the machine-readable data files.
Affiliate disclosure: PRUVIQ earns a commission from OKX when you sign up through our referral link — OKX pays out 50% of your trading fees, of which 20% comes back to you as a fee discount and 30% is ours. Your discount is not affected by this. Not financial advice. Crypto trading involves substantial risk of loss.