Our data & methodology
What is measured, where it comes from and what remains unknown.
What MarketOwl records
MarketOwl records technical observations and subsequent price measurements. A divergence or gold pattern is a monitoring event, not a trade instruction. The platform does not place orders.
Signals and confirmation
TradingView is the source of divergence and gold-pattern calculations. Each record identifies its exchange instrument, timeframe, confirmation time and reference price. Pivot patterns can become visible only after later candles confirm them. A historical pivot timestamp is not the moment the alert became actionable.
The saved monitoring universe excludes stablecoins as signal assets. Quote currencies such as USDT remain permitted. The universe is a dated snapshot, not a claim of automatically refreshed market-cap rankings.
Evidence and historical records
Captured proof preserves recorded price and momentum pivots where available. Older observations without this evidence are explicitly identified; no missing divergence line is fabricated. Live alerts and later historical reconstructions remain distinguishable.
Outcomes, not trading returns
Measurements follow one, three and five closed native candles. Directional rates describe the closing price relative to the reference. They exclude entries, stops, commissions, funding and slippage. Maximum favorable movement is not an executable profit. Missing outcomes are reported, not treated as wins or losses.
Market pulse
The market pulse describes breadth among eligible recorded signals. Short-term context uses 4H and 12H; broader context uses 1D and 1W. Gold patterns are reported separately. Coverage limits, decay rules and experimental thresholds are disclosed on the market page. It is not a calibrated probability of a future move.
Free supplementary market data
The radar uses Bybit public futures endpoints and a BTC/ETH/SOL liquidation stream. Ethereum USDT, USDC, WBTC and WETH transfers are sampled from Blockscout. OLI community attribution may be incomplete or incorrect. Large transfers are not proof of purchases or sales, and the sample is not total exchange netflow.
Futures open interest is compared with a genuine recorded baseline. Funding is a venue estimate per settlement event. Liquidation notional uses reported bankruptcy prices and is not a trader-loss estimate. API interruptions and unobserved periods can create gaps.
Freshness and privacy
The signals page refreshes the shared Telegram ledger every ten seconds. The radar page refreshes every thirty seconds; upstream collection intervals differ by source. Stale states are visible. History remains preserved during interruptions.
This research trial is public. Watchlist choices and journal notes stay in your browser; they do not synchronize between devices. TradingView charts and external verification links connect to their respective providers. No member payment collection is enabled.
Research limitations
Observations can overlap and are not independent samples. Neither past directional rates nor a combination of whale activity and technical patterns establishes a predictive edge. Evaluate any approach using the actual information receipt time and realistic trading costs.
Methodology updated 17 October 2026.