DATA & EDITORIAL STANDARD
Understand the evidence before using the analysis.
Methodology version 5 · Updated September 20, 2026. This is a reproducible description of the calculations and their limitations—not a claim of predictive performance.
Structure Lens ·
September 20 update: context, not a trading signal
We added a reproducible sentiment-history view: exact seven-day score change, the share of earlier observations at or below the last score, and consecutive daily observations in the same display band. The last observation is excluded from its rank comparison; missing days stop streaks and are never interpolated. Rank needs at least six earlier observations. These are descriptive statistics, not return forecasts or reversal probabilities.
Explore the explanation & evidence
We also corrected range descriptions that could imply broad participation, forced selling or leveraged crowding from sentiment alone. Those conclusions require separate evidence. News cards are links to original publishers, not our own reported articles.
The automated data brief and the dated editorial column are different publications. The brief is calculated on request. The column is AI-assisted research published with a frozen evidence capture, source references, counterarguments and review conditions. Its archive preserves the original view; publication and observation times are separate. Neither publication claims professional licensing or independently verified investment performance.
Material correction: September 6, 2026
A critical review found that earlier dashboard versions retained fixed reference values during API failures. Some derived analysis, live labels and scenario scores could still use those values. Those results were not reliable current-market evidence and should not be used to evaluate market conditions or forecast performance.
Explore the explanation & evidence
We removed the fixed market, funding, sentiment and trade fallbacks; retired composite scores and uncalibrated probability labels; corrected interval-based volatility; stopped comparing mismatched price/OI histories; and removed current snapshots from historical cutoff analysis. Missing values now remain missing. This correction replaces earlier methodology descriptions of the retired scoring models.
What each source actually measures
Complete price histories are preferred in this order: Binance BTCUSDT perpetuals, Coinbase BTC-USD spot, then Bybit BTCUSDT spot. An incomplete primary response does not prevent a healthy fallback. The chart explicitly changes its source label; we never join venues into one history or label a spot quote as a perpetual quote. Sampling can therefore differ: Coinbase supports 1-hour, 6-hour and daily intervals used here.
Explore the explanation & evidence
Funding and current OI come from Binance, or Bybit BTCUSDT when Binance funding history is unavailable. The displayed funding is a settled observation, not a prediction of the next payment. Bybit single-side OI value is used; if that field is absent we leave it blank rather than silently compare a both-side count with another venue.
Alternative.me supplies daily sentiment; CoinGecko supplies aggregate market capitalization and dominance; mempool.space supplies transaction backlog and fee estimates; Deribit supplies options open interest. These are venue/provider-specific observations, not a comprehensive tape of the entire crypto market. Source names do not imply endorsement.
Outages, timestamps and historical coverage
Non-price sources have a three-second deadline. Price providers are checked in parallel with a four-second deadline; if all price sources fail, transient connection/5xx failures are retried once, for at most eight seconds. HTTP access denials and rate limits are not retried. Provider status, completeness and latency are exposed in the price diagnostics. Cached market responses may be reused for up to one minute, while news and sentiment responses may be reused for five minutes. The fetch time is distinct from the source observation time. A successful HTTP response alone does not make every field available.
Explore the explanation & evidence
No valid response means null values and empty histories—not zero, synthetic candles or a dated sample quote. Only completed price candles before the requested end are shown. The chart uses actual timestamps, breaks across large gaps and exposes the first and last available observations. Public API record limits may shorten a selected window; a 1Y button is not a promise of a full year of data.
Daily price evidence is filtered to its cutoff. Global, network and options point-in-time archives do not currently exist on this site. Current snapshots are shown separately and never used as if they were known before an earlier cutoff.
Calculations you can reproduce
Window return uses the first and last available completed closes. Realized volatility is the sample standard deviation of log returns, annualized with the series' observed interval and a 365-day crypto year. Returns spanning missing intervals are excluded from volatility. This is a backward-looking statistic, not a prediction interval.
Explore the explanation & evidence
Maximum drawdown tracks the largest decline from an earlier sampled closing-price peak. It does not include intrabar extremes. Options put/call is total put OI divided by total call OI, shown as a percentage. Strike concentrations aggregate returned BTC options across expiries, then display the 12 largest combined concentrations. Aggregating expiries obscures term structure and says nothing about dealer gamma.
Window return = (last close / first close − 1) × 100 Log return = ln(close[t] / close[t−1]) Annualized realized vol = sample SD(log returns) × √(365 × 24 / interval hours) × 100 Put/call OI = total put OI / total call OI × 100
Funding cost and OI are not expected returns
Funding annualization is produced only when the current settlement interval is available. An historical interval is not inferred from today's setting. The funding calculator counts settlement boundaries and includes user-entered trading fees; its defaults are a clearly labeled hypothetical example, not live market quotes.
Explore the explanation & evidence
USD OI changes with both contract quantity and mark price. If 100 BTC of contracts remains open while price increases from $60,000 to $66,000, USD notional rises from $6.0M to $6.6M with no increase in contract quantity. That $600,000 is not evidence of new cash deposited. OI also does not reveal which side initiated the position.
The position-risk calculator estimates price-only loss at a chosen stop. Stops are not guaranteed execution prices. Gaps, slippage, liquidation and trading fees may increase realized losses beyond the stated budget.
Simple annualized funding = settled rate × (24 / verified interval hours) × 365 Funding payment ≈ position notional × settlement rate / 100 Illustration: 100 BTC × $60,000 = $6M; 100 BTC × $66,000 = $6.6M
How the daily note works
The price window ends at 00:00 UTC / 09:00 JST. After 00:05 UTC, requests use the next daily cycle. The note is generated on request; an open page checks every minute. The generation timestamp is not represented as a pre-existing publication time. Historical views reconstruct available provider records, not permanently stored original notes.
Explore the explanation & evidence
The note separates measured price facts, interpretation and conditional paths. Conditions reference sampled closing-price extremes. A close outside the range is a reason to reassess, not proof of a durable breakout. There are no probability scores, target prices, directional trade calls or claims of a backtested edge.
News links are restricted to the 24 hours before the selected cutoff. Feed coverage is not a full archive. Headline keywords are not scored as bullish/bearish evidence, and a price move is not attributed to an event without independent support. ETF flows, central-bank surprises and macro release histories are not connected data feeds here.
Editorial responsibility and commercial limits
Structure Lens Research Desk is the house publishing identity, not a claim that each page was reviewed by a licensed analyst. Automated systems assist calculations, drafting and translation. The site provides educational research, not individualized financial advice or guaranteed forecasts.
Explore the explanation & evidence
Corrections can be reported to hello@structurelens.com with the URL, visible timestamp and source evidence. Material errors are recorded here. Dates change when content or methodology materially changes, not to create artificial freshness. Source terms and redistribution permissions may impose limits on commercial use.
Optional analytics requires the visitor's consent. Advertising runtime scripts are disabled while the advertising consent setup is incomplete; the AdSense verification meta tag and ads.txt remain present. This consent control is not represented as a Google-certified advertising CMP. Approval and ad serving are separate decisions made by Google.