How CoinNudge crypto market signals work

Methodology version 2.1 and correction log

Effective September 13, 2026. Version 2.1 makes the three event-study sampling procedures independently reproducible. Each JSON publishes the circular moving-block bootstrap method, block size 3, 1,000 resamples, uint32 SHA-256-derived seed, xorshift32 transitions and linear quantile interpolation. It also publishes exact matching fields, normalization divisors, unit weights, missing-component penalty, null caliper, no-replacement rule, processing order, tie-break and same-symbol overlap exclusion, plus the complete candidate-pool download and a SHA-256 digest of its ordered observation IDs. Version 2.0 standardized ISO UTC, row evidence, price-path efficiency naming, Kaplan–Meier recovery, buying-pressure components, IV/RV coverage and the 20-event publication gate. Version 2.1 additionally publishes Greenwood log-log intervals for recovery probabilities, Fisher-z intervals for buying-pressure correlations and explicit zero coverage cells. These controls improve reproducibility but do not establish causal effects, executable profits or point-in-time capture.

Batch-six historical-window maturity contract

UTC-hour volume uses a fixed BTCUSDT, ETHUSDT and SOLUSDT basket. A date enters only when all 24 hours exist for all three assets. The 30-day profile requires at least 27 complete common days and 90% of requested asset-hours; the 90-day profile requires at least 80 days and 90%. Weekend analysis applies the same 27/80-day and 90% gates per asset, compares median one-hour Saturday/Sunday UTC observations with Monday-Friday, and excludes every Warming 90-day asset from the cross-asset conclusion. Its volume-ratio 95% robustness interval uses 1,000 deterministic resamples of complete UTC-day blocks, keeping all 24 hours inside each sampled block. Each row publishes seed = 100 × sum of uppercase displayed asset-symbol Unicode code points + requested days, the unsigned 32-bit xorshift32 state transitions, and sorted-sample linear interpolation at (n-1)×q for q=0.025/0.975. Requested days, available days, complete days, expected and accepted hours, coverage, first/last observation and rejected count are public.

Deribit implied-move and actual-expiry curve contract

For each actual listed expiry, CoinNudge requires an open call and put at one shared strike closest to the expiry's median underlying price. Straddle reference move equals call mark plus put mark in underlying units; the separate IV approximation equals mean call/put mark IV multiplied by square root of ACT/365 time. Instrument names, marks, bid/ask, strike distance, OI and volume remain auditable. Term structure plots every qualifying actual expiry from 12 hours to 120 days on an uneven days-to-expiry axis. Endpoint slope is back ATM IV minus front ATM IV; no constant-maturity interpolation or synthetic prior is inserted.

ATH identity and altcoin-beta sample contract

ATH rows are exact unique-ticker CoinGecko-to-Binance base-asset crossmatches, not contract-address verification. Duplicate CoinGecko tickers, non-liquid matches and invalid price/ATH rows remain in the selection audit. Temporal scope separately identifies the current-price observation, up-to-90-candle local high and included lifetime ATH date span. Drawdown is current/ATH minus one; recovery is ATH/current minus one. Altcoin beta uses timestamp-aligned completed daily simple returns and OLS covariance(asset,BTC)/variance(BTC). It publishes a normal-theory 95% interval from residual and benchmark variance, explicitly not heteroskedasticity- or autocorrelation-robust. The page also publishes correlation, R-squared, reliable/weak-fit status, upside/downside observation counts and low-confidence labels when a directional subset has fewer than 10 observations.

Bitcoin candle-binned volume-profile contract 2.0

The BTC proxy uses a fixed $1 logarithmic origin with successive edges multiplied by 1.0025, so current price cannot move unchanged historical bins. Each completed candle's quote volume is allocated uniformly through every touched high-low bin. POC is the largest allocated bin. Value area starts at POC and expands contiguously toward the higher-volume adjacent bin until at least 70% is covered. The page publishes bin origin, step, edge rule, POC, VAL, VAH, actual share, candles and both 30-day/5-minute and 90-day/1-hour histograms. It is not tick-level trade-by-price data.

Same-venue spot-versus-perpetual volume contract

Products must share venue and base asset, each exceed $1 million rolling volume and be observed within 60 seconds. OKX spot uses quote-currency volCcy24h; swap volCcy24h is base volume multiplied by last price. MEXC spot uses quoteVolume; perpetual volume uses amount24, while OI uses holdVol × contractSize × fairPrice. Coinbase spot uses approximate quote volume or base volume × price; International perpetuals use notional_24hr and OI × mark price. USD, USDT and USDC are treated at face value without FX/depeg adjustment. Candidate, identity, product, floor, unit, timing and successful-pair counts are public.

Collection cadence and endpoints

Binance spot tickers are polled every 15 seconds and closed klines every 60 seconds. Hyperliquid metaAndAssetCtxs is polled every 30 seconds. MEXC, Coinbase and OKX spot directories and the three supplemental derivatives directories are polled about every 45 seconds so strict 60-second comparison gates can be enforced. Binance and MEXC announcements refresh every 120 seconds; Fear and Greed hourly. Deribit options refresh about every 15 minutes, DefiLlama protocol economics hourly and the restricted ETF preview every six hours. Coinbase BTC and ETH trades and OKX liquidation-orders use public WebSocket streams. Backoff applies after failure, so cadence is not an uptime guarantee. Page views only read local materialized snapshots.

Freshness, alignment and gaps

A five-minute candle must be evaluated within 120 seconds of close and a 15-minute technical or flow candle within 300 seconds. Closed 1-hour and 4-hour context must align to the event time; candle sequences are truncated after an interval gap. Coinbase confirmation must be within 90 seconds. Perpetual-versus-spot and cross-exchange price comparisons reject any row whose source observations span more than 60 seconds; rejected rows cannot become a Top result. Delivery expiry varies by category so stale events are not replayed as fresh opportunities.

Telegram promotion and suppression

Standalone RSI requires 25 or below or 75 or above. MACD or EMA needs aligned 1-hour and 4-hour context, ADX at least 20, non-countertrend DMI, relative volume at least 1.5 times and a candle move of at least 0.25 percent. Volume, Bollinger and derivatives apply separate materiality rules. Lower-information records remain in the feed or are suppressed from Telegram; they are not delivery failures.

Default coin selection

Pro includes BTC and ETH plus 10 platform-managed popular coins for 12 default coins total, and up to 10 personal coins. Eligible defaults require an allowed Binance USDT pair and at least 300,000 dollars in venue-reported 24-hour quote volume, prefer a maintained popular-asset universe and remain stable for the UTC day.

Price and volume

Price compares consecutive closed 5-minute candles. Volume compares the latest closed quote volume with the previous 20 candles, excluding itself. With fewer than 288 valid historical ratios, the volume threshold is the fixed 3× fallback; otherwise it is max(3×, the trailing 99th percentile). Public ranking is alert-qualified first, then rule events, then context; within each state, volume ratio is sorted from highest to lowest. Platform liquidity and asset-class filters apply, and users do not configure raw thresholds.

Breadth universe and denominator changes

The breadth universe is Binance USDT spot with at least 1 million dollars in rolling 24-hour quote volume. One shared asset policy reads Binance product tags and excludes bStocks, tokenized ETFs, stable-value products and tokenized commodities; leveraged-token rules and a maintained fallback catalog provide additional protection. Every eligible crypto market receives one equal vote. CoinNudge stores the UTC-day opening universe under the current policy version and publishes coins added to or removed from the current denominator so a liquidity-driven membership change is not mistaken for price participation.

Sector basket composition

The 13 CoinNudge baskets are curated, overlapping research categories rather than official Binance sectors. Display baskets require at least five current members with at least 300,000 dollars each in 24-hour quote volume; alert baskets separately require at least 1 million dollars per member. Membership is frozen for the UTC day, median return limits outlier dominance, and every live sector row exposes its complete eligible constituent list.

Technical indicators

RSI uses Wilder smoothing over 14 periods. EMA uses 9 and 21 periods; MACD uses 12, 26 and 9. Bollinger bands use 20 periods and two population standard deviations. Crossovers require two closed candles. ADX and DMI, ATR percentile and UTC-day VWAP add trend, volatility and cost-basis context.

Spot order flow and Coinbase confirmation

Active buying and selling are derived from executed Binance Spot candles, not deposits or withdrawals. BTC and ETH also receive time-aligned Coinbase USD aggressive-flow and price-premium context. Neither measurement is described as total market capital inflow.

Perpetual positioning

The Pro leverage monitor keeps venue-defined hourly funding, aligned OI, price and volume separate. Squeeze risk publishes five component states and points: funding 20, rising OI 15, adverse price 25, same-side OKX liquidations 25 and opposing Binance Spot CVD 15. Thresholds are 0.01%, +3%, 1%, $100,000 and 5%; a component contributes half its cap at threshold and reaches its cap at twice that threshold. OI points are shared by both scores and cannot determine direction. A risk_side is assigned only when scores differ at the published 0.1-point precision and the leader has funding, price, liquidation or CVD support; otherwise it remains null in the direction-undetermined table. Falling-OI rows are excluded as active deleveraging. Perpetual basis is (perpetual / Binance Spot - 1) × 100 for at least $1 million OI and a source span no wider than 60 seconds. Neither score nor basis is probability or guaranteed carry.

Liquidation pressure

OKX forced-liquidation events are grouped over a rolling 15-minute window. Contract units equal size × contract value × contract multiplier. If the contract is inverse or its value currency is USD, USDT or USDC, those units are treated as quote notional; otherwise units are multiplied by bankruptcy price. Invalid metadata and explicit non-crypto underlyings are rejected. Private alerts require an extreme amount, side concentration, a matching price move and supporting open-interest or spot-flow evidence.

Announcement classification and schedule priority

Binance and MEXC announcements come from their official domains, and the stored exchange must match the URL host. Listing and removal event types are mutually exclusive; Convert and symbol-less rows are quarantined from public listings and Telegram. Spot, futures, new pairs, Meme+ and pre-market remain distinct. Time priority is a labeled Launch/Listing/Trading Time table, then a labeled body value, then an action-bearing title. UTC offsets are normalized to UTC; conflicting explicit times are labeled schedule conflict instead of silently choosing one.

Token-unlock identity and order-book calculations

Token unlock coverage is limited to NetSupply schedules verified from tracked on-chain objects. A ticker receives CoinGecko price and volume context only when it maps to exactly one current top-500 market-cap asset; ambiguous identities remain unpriced. Binance order-book midpoint is (best bid + best ask) / 2, spread is ask/bid minus one in basis points, band depth sums price × quantity inside 0.5% or 1%, and the $10,000 simulation walks returned levels into a volume-weighted execution price. Only the top 500 levels per side are observed.

CVD, volatility, ATR and correlation

Binance Spot CVD per completed five-minute candle is 2 × taker-buy quote volume minus total quote volume. Short windows require at least 90% coverage and 24 hours requires at least 260 of 288 candles. Realized volatility annualizes completed hourly log-return dispersion; correlation uses timestamp-aligned completed daily returns with at least 80% coverage. For Deribit skew, T is ACT/365 years to expiration_timestamp, σ is mark_iv / 100 and r is instrument-level interest_rate. Delta is derived from d1=(ln(S/K)+(r+σ²/2)T)/(σ√T), with call=N(d1) and put=N(d1)-1. The nearest +0.25 call and -0.25 put are rejected if more than 0.12 away; selected put, call and ATM instruments and rates are published.

Dominance, DEX volume and trending attention

CoinGecko dominance history uses closest hourly observations and leaves unavailable windows Warming. Altcoin relative strength and new highs/lows use completed Binance Spot data; the highs/lows page identifies 20/50/90-day level events while Breakout Scanner adds ATR and volume confirmation. Cross-exchange spreads require identity-matched USD-like pairs and a maximum 60-second observation span. Exchange volume share is explicitly a four-venue common basket, not global share. The ETF page exposes only three delayed sessions per asset. DefiLlama total24h is current and total48hto24h is prior; change is (current/prior-1)×100, with percentages suppressed below a $10,000 prior base. Fees, protocol revenue and token-holder value remain separate.

Listing-performance checkpoints and exclusions

Each eligible Binance or MEXC record keeps venue, symbol, official URL, parsed opening time, requested USDT candle pair, HTTP status, usability, exclusion reason and actual first completed Kline. The first close must be within one hour of the expected first-hour checkpoint. Later 24-hour, seven-day and 30-day closes use the same one-hour tolerance and the first close as baseline. Missing series stay in the audit trail, never become zero returns, and an aggregate 24-hour median is withheld until at least five listings mature.

Noise control

Same-direction repeats are cooled down, related events can be grouped, and conflicts remain visible. Sentiment notifications have a one-hour cooldown and a six-hour delivery-expiry boundary; routine daily readings remain in the briefing. Expired market alerts are not replayed as fresh opportunities. Polling and upstream outages can still delay or omit events.

Source coverage and retention

Binance spot, Coinbase confirmation, Hyperliquid positioning and OKX liquidations are separate venue-specific inputs. None represents every asset or the whole market. Public research signals are retained for historical analysis, without a fixed age-based deletion window. Notification and delivery-job records are normally removed after 30 days; they are not the same dataset as research signals.

Delivery controls

A deliverable message includes the coin, venue, timeframe, event time, trigger and available supporting or conflicting evidence. Users select a notification language and timezone; the website does not need to stay open after Telegram is connected.

No fabricated proof

CoinNudge does not claim a success rate, backtested return, customer count or guaranteed profit. A triggered condition reports what happened in the observed data; it does not tell the user what will happen next.

APPLY THE DOCUMENTED METHOD

Choose alert coverage after you understand the rules.

The methodology stays public. A CoinNudge plan adds continuous monitoring, noise controls and private Telegram delivery using the rules documented above.

9

event-driven alert categories

12 + 10

default and personal coin coverage

1 · 12

daily briefing · market modules

12

notification languages

What the purchase pays for

• Platform-managed rules reduce setup work while closed-window and freshness checks suppress lower-quality events.
• Every deliverable message keeps the venue, timeframe, trigger, evidence conflicts and practical limitation visible.
• No exchange trading key is needed; CoinNudge does not hold funds, size positions or execute trades.
• Pro is $8.99 for 30 days, paid once with supported USDT networks and no automatic renewal.

Your path from research to alerts

1. 01. Inspect the methodology — Verify the sources, calculations, promotion rules and known limitations on this page.
2. 02. Compare coverage — Review which categories are included in Free and which require Pro.
3. 03. Connect Telegram — Choose notification language and timezone, then receive only qualifying private alerts.

See plans and pay · Preview Telegram alerts

A triggered condition reports an observation, not a prediction. CoinNudge does not provide personalized advice or guarantee returns.

Working with structured observations? Explore CoinNudge Data or inspect the Market Events sample. Data subscriptions are separate from Telegram Pro; check the published dataset scope before purchase.