Dude Report

Dudelytics now explains in one place what may be done with its own content and public data, where redistribution stops, and which external rights continue to apply.

• German and English now have a dedicated licence page with a short summary, usage matrix, third-party source overview, AI/agent guidance and an open-source section.
• Legal information, data sources and the AI/agent page link to the new rights overview without touching live market or index data.
• Public JSON and discovery endpoints are explicitly separated from any blanket licence to mirror, resell or white-label complete datasets.

Licences & rights → Data sources → AI & agents →

DVIX measures stress in the productive crypto market on a transparent 0–100 scale while remaining methodologically separate from DPMI Fear & Greed and DPMI Holder Quality.

• DVIX combines realized DPMI volatility, a negative 24h shock, co-movement across the seven DPMI sectors and shared drawdown pressure.
• The calculation uses observed DPMI and sector-index history only and makes no additional provider calls of its own.
• DPMI-HQ remains a separate index with its own index level; the DVIX view compares its 24h performance with the main DPMI instead of treating it as a 0–100 quality score.
• Rebalancing windows are marked diagnostically but do not artificially change the DVIX score.

DPMI Family · Dude-VIX → Public DPMI page →

Dudelytics now bundles product terminology, methodology boundaries and public data endpoints on a dedicated AI/agent page and in a machine-readable discovery manifest.

• /en/ai/ explains public Dudelytics concepts for humans, search systems and browser agents.
• /data/ai/ exposes stable ASCII IDs such as dpmi, dpmi_hq, fear_greed and dvix together with their public endpoints.
• llms.txt and llms-full.txt point to the same public product truth instead of creating a parallel data universe.
• Account, admin, candidate/shadow, secret and private working data remain excluded.

AI / agent page → AI Discovery JSON → llms.txt →

The scheduled monthly rebalancing for October 1 continues both index series while preventing basket changes from appearing as synthetic market moves.

• The main DPMI remains the broad productive-market index; DPMI Holder Quality remains a separate index with its own composition and history.
• Before the basket changes, the existing index level is valued at current prices and the new basket continues from that same level.
• The rebalancing path remains fail-closed: incomplete candidate or market data does not create a partially written production basket.

DPMI Family → DPMI methodology →

DPMI-HQ represents a dedicated holder-quality sub-universe as a full index with its own level, history and monthly rebalancing.

• DPMI-HQ is explicitly not a 0–100 quality score.
• The index is interpreted through its relative performance and especially its comparison with the main DPMI.
• Historical reconstructions and genuinely observed points remain distinguishable by data quality.

DPMI Holder Quality → DPMI Family →

Governance data quality now distinguishes cases that truly block a readiness classification from additional detail enrichment for projects that are already classified.

• Existing local evidence is re-evaluated during registry rebuilds but is not deleted.
• Unknown classifications remain genuine research cases and are not converted into zero or negative values.
• The separation reduces unnecessary manual closeout work without lowering evidence requirements.

Data sources & coverage →

The DPMI Family now has a cross-asset layer: DPMI, BTC and ETH are compared with the S&P 500, Russell 2000 and gold on a common base of 100; three pairwise Pearson cards below show current market co-movement.

• DPMI, BTC and ETH come from the daily DPMI benchmark with a completed 00:00 UTC reference point. The S&P 500 is represented by IVV, the Russell 2000 by IWM, and gold directly by XAU/USD; tokenized ETF prices are not used.
• The chart starts at 100 on the first complete shared real observation. DPMI, BTC, S&P 500 and gold are visible by default; ETH and Russell 2000 can be added.
• Missing exchange days do not create synthetic data. Lines are carried forward only for chart display until the next real point; those visual continuations never enter the correlation calculation.
• The three correlation cards show the DPMI, BTC and ETH perspectives against the other markets. Pearson is calculated from real shared return intervals: provisional from n=7 and regular from n=10.
• Source, proxy notes and the latest real observation stay visible next to the chart. Because ETH is itself a DPMI constituent, the DPMI/ETH correlation is explicitly treated as not fully independent.

DPMI Family · Cross-Asset → Dude Report →

A much larger part of the public website is now available in English. DPMI, methodology, features, data sources, FAQ, Dude Report, research, legal pages and support still use the same central Dudelytics data foundation.

• German remains the canonical content base; the English presentation layer uses the same central data, methodology versions and public research snapshots instead of creating a separate data world.
• The English DPMI layer covers the main index, sector family, benchmark, Fear & Greed, methodology and public JSON/research entry points.
• FAQ, data sources, features, Dude Report, Dudipedia, support, privacy and legal pages were aligned linguistically and structurally with the English public navigation.
• The change does not alter the DPMI engine, provider data, scores or schedulers; it expands public access to the same product truth.

English Home → English DPMI → English Dude Report →

Opening Benchmark Data no longer drops users into a technical source/JSON view. Daily values are shown directly inside the DPMI Family, and navigation remains in the correct section.

• Daily benchmark values are displayed in an accessible table inside the Web App.
• The technical public JSON file remains available as a data source, but it is no longer the primary user destination behind the “Benchmark Data” button.
• DPMI Family navigation stays consistent when opening and closing the benchmark-data view.

DPMI Family → Public benchmark data →

The daily briefings now move more clearly from the big picture to the details: market, macro/liquidity, derivatives and research are combined with the DPMI Family and then the relevant coins into a more compact frozen daily view.

• The DPMI Family is now a fixed part of the briefings: main DPMI, 24h/7d context, DPMI Fear & Greed, the DPMI-vs-BTC-vs-total-market benchmark and all seven productive sectors are assessed together.
• The new “forest-to-trees” structure starts with market breadth and overall-market context, adds macro/liquidity, futures/derivatives and research/news, and only then moves to individual coins.
• For the DPMI Top 30, the briefing now provides a compact breadth and extremes view with BTC reference, status picture and Top-3/Bottom-3 context instead of a long undifferentiated coin list.
• Morning and Evening remain separate frozen snapshots with their own data cutoff and comparison with the current state. Duplicate highlight blocks were removed from the presentation.
• Web App and email present the same centrally frozen briefing content; the proven mail transport and delivery deduplication remain unchanged.

Morning Briefing → Evening Briefing →

Watchlist-based price-alert rules are evaluated from the existing five-minute market stream and, with explicit email opt-in, delivered through a separate worker without mixing signal logic with mail delivery.

• Users can configure personal rules for price moves over 15, 30, 45 or 60 minutes with fixed thresholds and direction.
• The signal engine evaluates market data already collected by Dudelytics and makes no additional provider or network requests of its own.
• Automatic email delivery is enabled explicitly per account; the rule remains stored even when the email channel is disabled.
• No old hits are delivered when the feature goes live: outbox entries created before activation expire as pre-activation events, and stale or no-longer-valid hits are not sent either.
• Delivery uses the separate STRATO SMTP path and can be checked with a test email to the verified account address.

Price alerts →

Gem discovery and the Quality evidence layer were hardened further: missing data stays unknown, independent discovery never changes DPMI eligibility, and expensive deep checks are fed from a controlled candidate pool.

• GEMS-2.0 remains the shared deep-review methodology; independent discovery finds candidates in addition to the DPMI Core but never writes into the DPMI eligibility registry.
• The central rule remains strict: missing evidence is UNKNOWN and must not make a candidate negative by itself.
• Discovery catalogs and prechecks in limited batches, keeps an exploration reserve to reduce selection bias, and schedules different recheck intervals for early, rejected, challenger and candidate states.
• The Quality evidence layer for Team/Execution, Developer Activity and Governance/Decentralization works interactively from cache only; background collectors remain separate from it.

Gems → Hot or Trash →

An independent outside review was turned into a small consistency patch: the 15% rebalance cap was verified for Sep 1, rank and live weight are clearly separated, XRP gets stronger primary-source evidence, the Bitcoin hashrate chart is smoothed, and FAQ/Product Truth now share the same public overlay state.

• The archived DPMI rebalance from Sep 1, 2026 contains 30 assets, sums to 100%, and nowhere exceeds the 15% cap at rebalance. Live weights may later drift above 15% as prices move; the methodology itself is unchanged.
• The public DPMI component table now states explicitly: rank = market-cap order at the latest rebalance, weight = current live weight. A runtime invariant checks stored rebalance weights for both the cap and a 100% total.
• For XRP, public evidence now includes official XRPL primary sources for the DEX, AMM and fungible tokens/tokenisation. Eligibility, rank and DPMI weighting are not changed by this.
• The Bitcoin network chart shows hashrate as a 7-day moving average to make the naturally noisy estimate easier to read. The current hashrate metric remains the unchanged provider value; difficulty remains raw and step-like.
• Product Truth JSON and the FAQ now use the same public overlay state for Public Preview, DPMI-BENCH-1.1, Security Radar and Public Research. Existing llms files are not overwritten; they are checked for the same core facts before installation.
• Historical Dude Report entries from Aug 30 are marked as the final Alpha state before the switch to Public Preview later that day, without rewriting the historical record.

DPMI Family → FAQ & Product Status → Product Truth JSON → Dude Report →

The public Research Hub now exposes a few current core values on the landing page; the JSON index reports timestamp and freshness per dataset, and stale Trend Radar snapshots are no longer presented as the current narrative picture.

• DPMI, Market Overview, Bitcoin Network and Security now show a small set of current key values and their timestamp directly in the Research Hub without triggering additional provider or AI calls.
• The machine-readable /data/research/ index was upgraded to schema v2 and now returns title, short description, updated_at, status, freshness age and selected key facts for each dataset.
• Trend Radar is considered stale in the public research layer after six hours without a new snapshot. The latest state remains traceable, but is visibly and machine-readably marked stale rather than phrased as the current narrative picture.
• The DPMI Fear & Greed endpoint was already current during the live check and remains linked as its own public data endpoint.
• Account features, Gems, Coin Analysis, schedulers and provider budgets are unchanged by this refinement.

Research Hub → Research JSON Index → Dude Report →

A new public read-only Research Hub publishes DPMI Family, compact market context, Trend Radar, Bitcoin Network and Security as stable HTML and JSON snapshots without opening account features or provider access.

• The Research Hub lives under /research/ and deliberately is not another primary Web App navigation item. People can open it by direct URL, search or citation; crawlers can discover it through the sitemap and LLM hints.
• DPMI Research covers the main DPMI, all seven sector indices, the daily benchmark and DPMI Fear & Greed; the full public DPMI charts remain on the existing DPMI page.
• The public Market Overview exposes only central total-market, dominance, sentiment and DPMI values plus a compact Dudelytics assessment. Interactive depth remains in the account.
• The public Trend Radar is limited to the five leading narratives and 24h/7d context; top coins, hot candidates and protocol details remain account content.
• Bitcoin Network and Security get dedicated public research pages and machine-readable JSON snapshots. Security incidents can be referenced through public chronologies.
• Opening a Research page triggers no external provider or AI calls. Account data, watchlists, admin information, internal queues, secrets and personal data are not published.
• The structure is prepared for later /research/coins/, /research/gems/ and /research/rwa-reality-check/ modules without publishing them in V1.

Research Hub → Dude Report →

Current, quarter and year counts now run directly from known incidents; updates are not double-counted, and a trend direction appears only once the observed dataset supports a meaningful comparison.

• The previously planned historical backfill is no longer planned or executed. Security is an ongoing information service, not a retrospectively reconstructed complete archive.
• The header shows Current, the running quarter and the current year from day one. An incident counts exactly once by incident ID and incident date; later updates do not increase the count.
• The 12-month monthly values may start immediately. Until a meaningful comparison exists, the trend direction stays blank instead of pretending to be stable, rising or falling.
• Discovery and adaptive recheck logic, threat levels, sources, German short summaries and the global ❗/⚠️/🔥 seen-state logic remain in place.

Security Radar → Public Security Research → Dude Report →

Automatically discovered security cases are now normalised into short German summaries; status distinguishes a newly reported case from monitoring, containment and recovery so “Current” is not inflated by merely recent historical incidents.

• Raw SlowMist text is no longer shown 1:1. Dudelytics creates a short German summary from attack method, loss, affected users and status; the original source remains available via [more].
• New discovery cases without evidence of an ongoing attack start as “Reported” rather than automatically becoming “Monitoring” merely because of their date.
• “Reported”, “Monitoring” and “Recovery” count as currently open. Contained, resolved or historical cases remain visible in the ticker but do not increase the current incident count.
• The threat level continues to describe present risk: contained or resolved cases are not held at “Serious” solely because they once involved a large loss.
• Previously stored SlowMist entries are normalised on read under the same rules without a new provider call. Seen state and alert revisions remain unchanged.

Security Radar → Dude Report →

The Gems logic is deliberately separated from Coin Quality: Candidates, Challengers and confirmed Gems are judged by exceptional idea, execution, token coupling, discovery character and asymmetry — not by maximum maturity or a blanket 80% data-coverage threshold.

• Gems are explicitly Dudelytics’ speculative high-risk area. Typical early-stage traits such as short history, low circulating share and normal unlocks are explained once as context instead of repeated as automatic exclusions for every young project.
• The new path is Candidate → Challenger → confirmed Gem. A Candidate may become visible early when discovery signals are strong; a confirmed Gem still requires strong evidence and all four gates.
• New elements are the Distinctiveness Gate, Token Gate, Kill Switch and Discovery Gate. Established large projects may score extremely well on quality and still fail Discovery because they are no longer hidden-gem cases.
• Project age is no longer rewarded by default. Young projects are judged on phase-appropriate evidence; a long history without use, development or a new turning point can instead be negative.
• Low circulating supply and ordinary early-stage unlocks feed into tokenomics and asymmetry with weights. Only exceptional liquidity, security or project-specific risks can act as a Kill Switch.
• The blanket 80% evidence threshold is removed. Dudelytics instead shows score evidence and data coverage separately; missing data stays unknown rather than becoming zero.
• PERLEN-1.x results are not carried forward as current 2.0 results after the methodology change. The existing scheduler queue reassesses them within the normal quota; database state, cost counters, AI-cache version and the running data pipeline are not reset.
• Coin Analysis adopts the same Gems 2.0 perspective and leads with why a project may be interesting; project-specific special risks and unusually clean early-stage traits are highlighted separately.

Gems → Dude Report →

The Watchlist is back to a wide working view on iPad and desktop; generic long-page navigation no longer mistakes individual coins for page sections.

• The previous two-column authenticated layout wrapper was removed from the Watchlist. Coin cards now use the available content width instead of only about two thirds of the page.
• Generic account section navigation is explicitly disabled on /watchlist/. Bitcoin, Ethereum, Solana and other watchlist coins no longer generate false section tabs or repeated Previous/Next section bars.
• The explanation of the difference between the normal Watchlist and HoS tracking remains, but now sits below the Watchlist instead of consuming a wide side column.
• Watchlist data, sorting, drag-and-drop, positions, HoS links and Coin Analysis links remain functionally unchanged.

Watchlist → Dude Report →

While the historical security backfill was not yet comparably complete, the header stopped showing provisional minimum counts or a fake trend and instead clearly said that history was still being built.

• Current active incidents remained visible immediately; quarter, year and trend figures were deferred until the historical backfill was complete enough.
• The expandable 12-month section explained that there was not yet enough comparable Dudelytics data for a robust trend assessment.
• An incomplete backfill was no longer presented as if it were a market trend. Incident timeline, threat levels, seen state and global security symbols were unchanged.

Security Radar → Dude Report →

A dedicated Security page brings incidents and updates together newest-first, separates confirmation status from current threat level and flags unseen serious or critical developments directly in Web App navigation.

• Security Radar is a dedicated Web App page covering protocols, chains, bridges, exchanges, wallets including hardware wallets, and other important crypto infrastructure.
• Each update carries its incident context in the title. The global timeline is strictly reverse chronological: the newest robust update is always on top and scrolling moves into the past.
• The current threat level is shown as Limited, Elevated, Serious or Critical. It describes today’s risk rather than merely the historical loss amount.
• Confirmed, Partially confirmed, Not officially confirmed and Conflicting reports remain separate evidence levels. A reported incident is not prematurely labelled a confirmed hack.
• External original or primary sources open via [more] in a new tab while Dudelytics remains open.
• Navigation shows only the highest unseen alert type, without a counter: ❗ new report, ⚠️ new serious report or 🔥 new critical report. Opening Security Radar marks the current alert revision as seen account-wide.
• Regular discovery runs every six hours. Known serious cases may be rechecked about every two hours and critical cases hourly. Interactive pages never trigger provider calls.
• The 12-month statistic and quarter/year counts are only presented as comparable once historical coverage is actually sufficient. Historical backfill cases do not generate new navigation alerts.
• SlowMist Hacked provides broad discovery coverage; ongoing assessment can be supplemented with curated primary sources. The GoMining wallet incident from Sep 4 deliberately starts as serious and not officially confirmed.

Security Radar → Dude Report →

Hashrate, difficulty, the latest and currently projected next difficulty adjustment, and retarget progress add Bitcoin’s technical mining dynamics to Market Overview.

• The new section shows current hashrate, 7d/30d trend, current difficulty and the latest actual difficulty adjustment.
• The running retarget period shows the current projected next difficulty change, progress, blocks remaining, estimated retarget time and average block time.
• Hashrate and difficulty are indexed to a shared starting value of 100 for 30d/90d/1y comparisons so different units are not compared through misleading dual axes.
• The Dude assessment explicitly treats network and mining dynamics as context, not as a Bitcoin price signal.
• Market Overview reads only the local Central Data Hub snapshot. Page views trigger no mempool.space provider calls; forecast and history are cached server-side and updated in a controlled way.

Market Overview → Dude Report →

Source coverage, provider resolution, usage and value data, and claim-level evidence checks now work together in one auditable quality loop.

• Point 1 — source basis: central source packages distinguish supported, missing, stale and technically unavailable information. Missing stays unknown and is not emitted as zero.
• Point 2 — provider resolution: DefiLlama, Coin Metrics, GitHub and Snapshot distinguish confirmed coverage, not applicable, missing coverage, expert clarification, deferral and technical errors.
• Point 3 — usage and value flow: users/addresses, transactions, TVL, fees, protocol revenue, holder revenue and token coupling are tracked separately. Direct value-flow claims require a measurement or a primary-source proof.
• Point 4 — evidence governance: claim-specific source contracts, bounded background queues and audited human decisions complement one another. A human source assignment never awards points directly.
• Acceptance state on Sep 11, 2026: 205/205 central quality objects are structurally valid; 152 belong to the active set and 53 to reserve. There are no open manual tasks and no invalid quality packages.
• Six decisions for Aave, Bitway, Chain-2, OpenEden, Render and Squid are logged. Four source decisions are automatically re-evaluated; Aave and Render repository mappings affect the data basis only after technical confirmation.
• Visible automatic clarification questions are ongoing system coverage work. They are neither red admin alarms nor 152 manual tasks nor invented zero values.

Data Sources & Coverage → Dude Report →

Growth Pass r3 implements the approved clean-first hero: the core question and BTC/ETH/SOL demo on the left, only the compact free registration on the right.

• The first product area on the home page now has just two clear jobs: understand and try on the left, register for free on the right.
• The previous duplicated large CTA buttons and additional explanation block are removed. Demo selection is directly visible and needs no extra click first.
• The public demo remains server-side limited to BTC, ETH and SOL; full coin search remains an account feature.
• The existing registration form is integrated compactly into the right hero column. Account logic, email confirmation, CSRF protection and privacy checkbox are unchanged.
• On narrow views the order is intentionally linear: core message, demo, registration.
• DPMI engine, DPMI methodology, scheduler, Market Data and Coin Analysis logic are unchanged.

Home → Coin Demo → Dude Report →

Growth Pass r2 reduces the entry point to one clear core question and opens only BTC, ETH and SOL as a real Dudelytics demo without an account; full coin search remains behind the free account.

• The hero now asks plainly: “How good is the coin, really?” followed by “What’s behind it — and does that success actually reach token holders?”
• The entry stays intentionally clean: a short explanation, one demo CTA, one account CTA and three clearly separated demo coins instead of extra forms and explanation cards.
• Without login, only Bitcoin, Ethereum and Solana are allowed. Other coin IDs and free-form coin search are blocked server-side; a manipulated URL does not expand the demo.
• The Coin Analysis V2 endpoints opened for the demo also enforce the fixed BTC/ETH/SOL allowlist server-side. Anonymous demo views may trigger neither a heavy Quality build nor a new external AI editorial call.
• Full coin search, Watchlist and personal features remain account features.
• Growth & Conversion counts demo selections and demo opens only as daily aggregates; no personal tracking profiles are created.
• Native iOS/iPadOS development remains under Other / Roadmap. DPMI engine, DPMI methodology and schedulers are unchanged.

Home → Coin Demo → Dude Report →

Growth Pass r1 puts Coin Analysis, Token Value Capture, market structure and Gems ahead of Public Preview and development detail, while measuring the conversion funnel privacy-first as daily aggregates.

• The hero starts with the concrete user value and leads primarily into Coin Analysis or Gems; DPMI Family remains prominent but follows the product entry.
• Public Preview remains transparently visible but is no longer sold as the main purpose. Registration and login copy now focuses on immediate product value.
• Ongoing native iOS development was removed from the top home-page flow and moved under Other / Roadmap.
• Growth Analytics counts only daily aggregates for home, key CTAs, registration, email confirmation, login and Coin Analysis. No raw IP, user agent, email, internal user ID or tracking cookie is stored.
• The admin area gets Growth & Conversion with 7/30/90-day funnels and coarse source/campaign aggregates.
• The safe source export now also includes sanitised private analytics logic; config.php and credentials remain excluded.

Home → Features → Dude Report →

The Dude now reads the latest completed ETH Exchange Supply cohort on two time scales: the longest robust horizon plus the latest real 1-day pulse from the same wallet cohort.

• If the latest 1d pulse confirms the longer trend, historical evidence is treated as stronger.
• If the latest 1d pulse runs against the longer trend, Dudelytics calls out the short-term counter-move instead of only repeating the 7d/30d primary direction.
• A stable final phase or a later new impulse is labelled separately.
• No cohort bridging, no backfill, no interpolation and no alteration of real measurement points.

Market Overview → Dude Report →

Recovered Exchange Supply history is no longer merely counted: Dudelytics calculates 1d/7d/30d within each cohort, shows a segmented path indexed to 100 per cohort, and summarises the latest completed cohort in the Dude history view.

• Earlier and current wallet cohorts remain methodologically separate; there is still no cross-cohort delta.
• Each cohort gets its own 1d/7d/30d values, segment change, lifetime and start/end balance where the real time base is sufficient.
• The new chart indexes every cohort separately to 100 and does not connect the segments. This reveals direction and scale without pretending that absolute wallet sets are directly comparable.
• The Dude history view uses the latest completed cohort as real historical evidence and clearly separates the signal then from the current comparison base.
• No retroactive history, no interpolation, no change to the 4-hour scheduler and no activation of the Exchange Supply score.

Market Overview → Dude Report →

Wallet-set changes still start a new comparable baseline, but no longer delete earlier real observations. Dudelytics segments history into cohorts and reconstructs existing older points from the Market Data DB where available.

• 1d/7d/30d values are calculated only from points in the same active wallet cohort; different PoR wallet sets are never mixed.
• Earlier scheduler observations remain as their own cohorts. A cohort change stores timestamp, old/new cohort, and going forward the affected exchanges plus wallet additions/removals.
• The existing Market Data series is used as a recovery source during migration so observations persisted before an earlier reset can become visible again.
• Market Overview shows total history and points from the current cohort separately; after a change it no longer misleadingly looks as though all history just restarted.
• No backfill, no interpolation and no change to the 4-hour scheduler or the disabled Exchange Supply score.

Market Overview → Dude Report →

The production parser was correct; the generic HTTP helper received a different body from Jina. ETF flows now use the exact request that worked in diagnostics.

• Direct Farside access from STRATO remains blocked by Cloudflare HTTP 403.
• Forensics showed that the generic production helper received a body without the Aug 31 row and without table pipes, while the same parser read the exact Jina body correctly.
• Jina ETF requests bypass the generic browser helper and use the diagnostically confirmed header/User-Agent set.
• ETF-FLOW-0.6, DPMI benchmark and schedulers remain unchanged.

After transport diagnostics, Dudelytics explicitly requests Markdown from the working Jina path and parses the Farside table robustly.

• Direct Farside access from STRATO reproducibly returns Cloudflare HTTP 403.
• r.jina.ai returns current BTC and ETH data when X-Return-Format: markdown is set; production now uses exactly that diagnostically confirmed request.
• The Markdown parser is hardened against wrappers, links, whitespace and minified tables. Provisional fund cells stay missing rather than becoming zero.
• ETF methodology, DPMI benchmark and schedulers remain unchanged.

Data Sources & Coverage → Market Overview →

A cache-busted Farside target is now preferred through the live reader relay; static/cached relay responses remain continuity fallbacks only.

• STRATO can temporarily fail to reach Farside directly. The new live-relay path varies the embedded Farside target URL each minute so an old relay cache cannot freeze the latest trading day.
• Fresh relay/direct data may revise stored trading days; stale clean-relay data may not overwrite the verified bridge.
• Aug 31, 2026 is verified in the bridge as an emergency reserve: BTC +$216.7m complete; ETH +$27.7m provisional with 7/10 funds reported.
• DPMI benchmark and schedulers remain untouched.

Data Sources & Coverage → Market Overview →

The central Farside sensor compares proxy and direct source by latest trading day, accepts later revisions date-for-date and marks partially reported new sessions as provisional.

• A compact proxy/Jina table with at least 20 rows no longer ends the fetch before the direct Farside source has been checked for a fresher trading day.
• Verified Aug 27/28 values correct old continuity data; later complete provider data may revise stored trading days.
• Provisional sessions are visible but do not affect the score or 5/20-trading-day context until fully reported.
• More than one expected US trading day of source lag becomes visible as degradation in the Dude Report.

Data Sources & Coverage → Dude Report →

The observed CoinGecko /global cutoff gets exactly one short transport retry within an eligible scheduler run without changing the 00:00 UTC target or the ±7-minute window.

• Each eligible cutoff scheduler run allows at most two CoinGecko /global fetch attempts, each with an eight-second timeout and only a short pause between them.
• The retry is provider/transport resilience only. It extends neither the ±7-minute cutoff window nor the allowed provider-side updated_at time.
• If the first fetch succeeds, no second request is made. If both fail, the total-market snapshot remains missing for that run and a later eligible cutoff run may try again.
• The 01:00 UTC late-finalise path stays staged-cutoff-only and may not substitute a 01:00 live spot value.
• DPMI-BENCH-1.1, the shared 00:00 UTC cutoff, normalisation and the main DPMI remain unchanged.

DPMI-BENCH-1.1 keeps a still-incomplete 00:00 target day open after 00:40 UTC for exactly one late-finalise attempt from 01:00 UTC, without moving the cutoff or changing observed values.

• The shared benchmark cutoff remains 00:00 UTC; the reserve does not move any measurement point to 01:00.
• During late finalisation, total market may come only from a CoinGecko /global snapshot already captured around 00:00 UTC. If that cutoff snapshot is missing, the daily point remains incomplete.
• DPMI and the completed BTC daily point may be read again from their real 00:00 target data during the single late-finalise attempt.
• Between 00:40 and 01:00 UTC the target day is held without forcing another provider finalisation every five minutes. The first scheduler run from 01:00 tries once and then closes the target day regardless of outcome.
• No backfill, no interpolation, no previous-day fill and no 01:00 live spot substitute. The main DPMI remains unchanged.

DPMI Benchmark → Benchmark JSON → Methodology →

The first real benchmark target day exposed that the planned historical CoinGecko total-market endpoint was unavailable on the accessible provider plan. DPMI-BENCH-1.1 therefore captures total market directly around 00:00 UTC going forward.

• The Sep 1, 2026 target day deliberately remains incomplete: DPMI and BTC were observed, but total market was missing. There is no retroactive backfill.
• From Sep 2, 2026, CoinGecko Total Market Cap is staged prospectively via /global at the cutoff; the provider-side updated_at timestamp must lie within ±7 minutes of 00:00 UTC.
• BTC remains the completed CoinGecko daily point for 00:00 UTC and is added from 00:10 UTC; DPMI continues to use its latest real observed point at or shortly before the cutoff.
• Target time, source time and fetch time remain separate. A later live value is never relabelled as the daily close.
• The main DPMI and its update process were not changed by this benchmark repair.

DPMI Benchmark → Benchmark JSON → Methodology →

The Web App can configure price-movement rules for Watchlist coins: coin, 15/30/45/60 minutes, 5/10/15 percent and direction up/down.

• Identical rules are not duplicated; saved rules remain assigned to the respective account.
• Evaluation uses the central 5-minute price history and treats missing or not-yet-mature reference points as unknown.
• Real delivery via push, email or webhook is still disabled in this stage; the rule base is being tested in a controlled way first.

Features → Price alerts in account →

Coin Quality can evaluate official project documentation, GitHub references, and security/governance evidence more robustly even when aggregators do not expose individual source fields.

• Official project and documentation pages may expose GitHub repositories as evidence sources without coin-specific special scoring.
• Markdown-heavy and modern documentation pages get a technical reader fallback; the original official project URL remains the cited evidence source.
• Security, audit, bug-bounty, governance, validator and staking documentation can make criteria assessable without turning unknown portions into artificial zero values.
• Quality weighting was not changed in favour of individual coins; the universal source path was improved.

Features → Data Sources & Coverage →

DPMI Family gets a daily comparison of the main DPMI, Bitcoin and the unfiltered total market. The target cutoff is 00:00 UTC; history starts observed-only with no retroactive backfill.

• The first complete daily point sets DPMI, BTC and total market to 100 together.
• BTC and total market use completed CoinGecko daily points for 00:00 UTC; the DPMI close comes from real observed DPMI history at or immediately before the cutoff.
• CoinGecko Total Market Cap deliberately includes stablecoins as an unfiltered contrast to the productively filtered DPMI.
• With fewer than 30 complete days, the page shows a warm-up state; before 90 days there is no directional Dude assessment.
• Missing daily values remain gaps and are not interpolated or filled from the previous day.

DPMI Benchmark → Benchmark JSON → Methodology →

The public product phase moves from Alpha to Public Preview: Dudelytics is publicly usable and functionally broad, while ongoing development, limitations and remaining rough edges continue to be stated openly.

• Access waves, waitlist, email confirmation and tester control remain technically unchanged.
• Internal development, admin and tester terminology may remain internally, but no longer appears as the public product positioning.
• FAQ, Product Truth, SEO/JSON-LD, llms files, login/registration, account copy and access emails now tell the same Public Preview product truth.

FAQ & Product Status → Product Truth JSON → Features →

Market Overview gets a separate A6 sensor for meme/hype momentum, speculative small-cap relative strength, turnover intensity and new DEX pools — without changing the productive DPMI methodology.

• Performance rotation means relative performance and activity; Dudelytics does not claim this directly measures capital flows.
• Missing data stays unknown and is removed from weighting instead of being scored as zero.
• The DEX pulse starts with a real baseline observation and matures over 24 hours without artificial backfill.
• High Speculative Heat is not a quality, safety or buy signal.

Features → Data Sources & Coverage → Dude Report →

This entry documents the final Alpha state on Aug 30 immediately before the switch to Public Preview later the same day. FAQ, structured data, llms files and Product Truth JSON were synchronised to that interim state at the time.

• At that point the Web App was still public Alpha; the move to Public Preview happened later the same day. Account and email confirmation remained technically in place.
• DPMI Family was consistently described as the main DPMI plus seven sector indices and separate DPMI Fear & Greed.
• Public Product Truth contained no private server paths, candidate/shadow data or internal eligibility thresholds.

FAQ & Product Status → Product Truth JSON → Features →

Data is reported late only after its own next_due plus a short grace window. Private shadow/candidate datasets do not appear in the public data-source catalogue.

• Refresh cadences and grace windows are unchanged; only the public interpretation was corrected.
• The Data Hub scheduler may execute a scheduled run shortly after next_due without the Dude Report immediately calling it an incident.
• DPMI Shadow Market is defensively filtered from public status and source views.

Dude Report → Data Sources & Coverage →

Data Hub datasets count as normal within their own scheduled refresh window. Only a genuinely missed next_due is exposed publicly as delay.

• Slower 4h/6h datasets are no longer treated like 5-minute sensors.
• Soft TTL alone does not create a public delay warning while the scheduled next run is still ahead.
• The overall status card was made responsive for iPad and long German status labels.

Dude Report → Data Sources & Coverage →

Coin Quality, Hot or Not and Trend Radar clearly separate signal strength, data coverage and what a score cannot tell you.

• No price forecast and no safety guarantee.
• Red flags, sources, data gaps and coverage remain visible.
• Longer explanations can be collapsed in frequently used work areas.

Features →

The sticky navigation now sits directly at the viewport edge without dead space; the Fear & Greed scale positions its marker correctly on the 0–100 scale and uses its real own history.

• 0 = fear/stress, 50 = neutral, 100 = greed/risk appetite.
• Anchor targets account for the compact sticky bar.
• Explanations remain collapsible while data and warnings stay visible.

DPMI Family →

In the account area, the new DPMI Family hub brings together main DPMI, Fear & Greed, seven sector indices, drivers, components and direct Coin Analysis links.

• Rebalance notes and component lists sit directly with the relevant index.
• Dude explanations are collapsible to keep the workspace compact.
• Section navigation remains pinned to the top of the viewport while scrolling.

Public DPMI Family → Why these seven sectors? →

A dedicated methodology page explains why Dudelytics uses exactly these seven productive sectors and why large market segments such as stablecoins or memecoins are not part of this measurement question.

• Market capitalisation is a weight, not an admission ticket.
• Not being in the DPMI does not mean unimportant or bad.
• New sectors require distinctness, economic relevance and productive substance.

Read sector selection →

DPMI Smart Contract Platforms was added as the seventh sector across Index Family, Web App, briefings, Dudipedia and public methodology.

• History begins only with real observation and is not invented retroactively.
• Fear & Greed includes the new sector automatically once the relevant time horizon has genuinely matured.
• All sector components remain traceable through DPMI Family.

DPMI Family → Methodology →

Source, refresh cadence, history, freshness, fallbacks and known limitations are now visible as their own transparency area.

• Missing data remains a data gap and is not turned into zero.
• Provider degradation is communicated as a state, not hidden.
• Primary and secondary sources are classified separately.

Data Sources & Coverage →

This historical entry describes the final Alpha state before the move to Public Preview later on Aug 30, 2026; it is not the current product status.

• Accessibility, device and workflow testing continued.
• The native iOS/iPadOS app remained in development.
• Known data maturity and degradations remained visible separately in system status.

Features → Data Sources →