Methodology
Every judgement in the logs comes from the public rules below. Rule changes are logged in Corrections; past logs are never rewritten.
How layer verdicts are made
L1 Macro liquidity (since 2026-10-02, based on the BTC full-cycle macro attribution study): L1 sets volatility, pacing and risk budget over weeks to months, not direction.
- Tailwind: optimism z > 0 and the BAA spread narrowing over 13 weeks, with no real-yield surge.
- Neutral-cautious: a real-yield surge (13w ≥ +0.40pp) on its own — one notch less risk budget, no adds.
- Headwind: a surge plus (BAA/NFCI tightening or optimism z < 0), or the liquidity gate (net liquidity 13w ≤ −2.72% with a stronger dollar) — nominal risk budget ×0.7, not compounded.
- Neutral: everything else. An 'alert' tag is added when the high-yield spread widens ≥ 0.30pp over 13 weeks (early warning; this site's threshold, not back-tested).
Evidence grades: verified > verified·direction > descriptive > monitor; only verified gauges may move position parameters. 13-week changes are aligned to Friday closes. Thresholds use post-2018 data only.
L2 Cycle：Cycle position = the MVRV Z-Score zone (MVRV as fallback). NUPL and realized price share its source and are not scored twice.
L3 Coin flows：Based on 7-day exchange net flow: outflow = coins leaving, inflow = watch selling. Exchange balance change is shown as trend only.
L4 Sentiment：Fear & greed (−2..+2), funding (−1..+2) and 7-day OI change (−1..+1): ≥3 hot, 1–2 warm, ≤−2 cool, otherwise neutral.
Alert thresholds
Core zone changes · L1 regime change · ≥5,000 BTC exchange net flow in a day · funding sign flip · VIX crossing 20 · Pi Cycle Top trigger · launchpad sector ±25% in a day · new #1 launchpad · deep-tracked platform ±60% DoD.
Indicators and zones
L1 · Macro Liquidity（How much risk budget？）
| Indicator | Level / evidence | How to read | Zones |
|---|---|---|---|
| Nasdaq-100 · 13w change FRED · NASDAQ100 | core Verified | First optimism gauge: ln(NDX / NDX 13 weeks ago). Correlates 0.95 with the common optimism factor. >0 = optimistic. | By 13-week log change：Weak < 0.00%；Optimistic |
| VIX FRED · VIXCLS | core Verified | US equity implied volatility. BTC follows the downside more than the upside, so a VIX spike usually hurts it more than the Nasdaq. >20 = optimism weakening. | By level：Calm < 20.00；Risk-off (>20) |
| Optimism composite z FRED（自算） | core Verified · direction | Mean of expanding-window z-scores (from 2011, no look-ahead) of Nasdaq 13w change, −VIX and the 13w change in 10y breakeven inflation. >0 = more optimistic than average. | By mean of 3 z-scores：Weak < +0.00；Strong |
| BAA credit spread · 13w change FRED · BAA10Y | core Verified | Moody's BAA corporate yield minus 10y Treasury. Main credit gauge, 0.89 correlation with the optimism factor since 2011. Rising over 13w = widening (pessimistic). | By 13-week change：Narrowing < +0.00pp；Widening |
| NFCI · 13w change FRED · NFCI（周度） | core Verified · direction | Chicago Fed National Financial Conditions Index; below 0 = looser than average. Rising over 13w = tightening. Published Wednesdays for the prior Friday. | By 13-week change：Easing < +0.00；Tightening |
| High-yield spread · 13w change FRED · BAMLH0A0HYM2 | aux Descriptive | Junk-bond spread over Treasuries. FRED only publishes ~3 years, so it is an early-warning reading only. The ±0.30pp alert band is this site's own, not back-tested. | By 13-week change：Narrowing fast < -0.30pp；Stable < +0.30pp；Widening fast (early warning) |
| 10y real yield · 4 states FRED · DFII10 | core Verified · direction | Now a state variable, not the headline: 13w ≥ +0.40pp = surging, ≤ −0.40pp = falling, otherwise high/low plateau by level ≥1.0%. A surge is an independent headwind since 2022 and is combined with credit/optimism. | By 13-week change + level：Falling < -0.40pp；Plateau < +0.40pp；Surging |
| Fed net liquidity · 13w change FRED · WALCL − WTREGEN − RRPONTSYD | core Verified · direction | Fed assets − Treasury General Account − overnight RRP. ≤ −2.72% with a stronger dollar over 13w trips the gate (risk budget ×0.7 in the framework); expansion never adds. Weakening since 2024-05; review end-2026. | By 13-week change：Sharp contraction < -2.72%；Mild contraction < 0.00%；Expanding (no add) |
| Broad dollar · 13w change FRED · DTWEXBGS | aux Verified | Trade-weighted dollar vs 26 currencies. Confirmation only; combined with net liquidity for the gate. | By 13-week change：Weaker < 0.00%；Stronger |
| Stablecoins ex-payment issuers DefiLlama（逐币） | aux Verified · direction | All USD stablecoins minus payment/institutional ones (PYUSD, RLUSD, USDG, USDGO, U, USDP, GUSD), which run −0.43 vs BTC over 13w in the ETF era. Replaces the old total-supply reading. | By 13-week change：Contracting < -1.00%；Flat < +1.00%；Expanding |
| Trading bullets · core ex-Tron DefiLlama（逐币 + Tron 链） | aux Verified · direction | (USDT+USDC+FDUSD+USD1+TUSD+USDD) minus all stablecoins on Tron (mostly EM remittances, ~uncorrelated with BTC since the ETFs). Closest proxy for exchange-side dry powder. | By 13-week change：Contracting < -1.00%；Flat < +1.00%；Expanding |
| Yield / synthetic stablecoins · 13w DefiLlama（逐币） | aux Descriptive | USDe, USDf, BFUSD, USDS and DAI combined. Size tracks funding rates and on-chain leverage — a leverage thermometer. | By 13-week change：Leverage receding < -1.00%；Flat < +1.00%；Leverage building |
| Spot ETF · 4-week net flow Farside（haturatu 镜像） | aux Verified · direction | US spot BTC ETF net flows over the last 28 days. Flows chase price with a lag (0.6 vs past 3–5 days, ~0.05 vs the future): trend confirmation only, never a contrarian/bottom signal. | By 28-day sum：Outflow (trend confirm weak) < $0；Inflow (trend confirm strong) |
| CME futures annualized basis (front) Yahoo · CME 近月合约（自己逐日积累） | aux Descriptive | (CME front month / spot at the same moment − 1) × 365 / days to expiry, rolling ~5 trading days before expiry. US institutional and basis-trade appetite. No free history: recorded daily since 2026-10, descriptive only. | By own-history percentile：own-history percentile: <20 low / 20–80 mid / >80 high (after 90 days) |
| Deribit 3-month annualized basis Deribit（前后两张插值成固定 90 天） | aux Descriptive | Crypto-native leveraged demand. Same concept as CME but a different market — levels often differ by several points, so only direction and own-history percentile matter. | By own-history percentile：own-history percentile: <20 low / 20–80 mid / >80 high (after 90 days) |
| BTC–Nasdaq 52w rolling correlation CoinMetrics + FRED（自算） | aux Descriptive | 52-week rolling correlation of weekly log returns. When it drops, Nasdaq-based rules lose weight; it fell to ~0 in 2024. 0.2 is this site's alert line. | By 52-week correlation：Low (down-weight Nasdaq rules) < 0.20；Normal linkage |
| Tokenized stocks / stablecoin cap CoinGecko 类目快照 | aux Monitor | Buying tokenized stocks usually drains stablecoins to the issuer. Only above 3% of stablecoin cap (or 5% of volume) does it enter the stablecoin adjustment; recorded only. | By cap ratio：Below threshold < 3.00%；Above 3% (adjust stablecoin lines) |
| 2Y/10Y curve FRED · T10Y2Y | watch Watch only | Weekly correlation with BTC only 0.06; BTC did better during inversions, the opposite of the folk rule. ~18 independent samples. Watch only. | By level：Inverted < 0.00%；Normal |
L2 · Cycle Position（Cheap or expensive？）
| Indicator | Level / evidence | How to read | Zones |
|---|---|---|---|
| MVRV CoinMetrics | core | Market cap / realized cap = average paper-profit multiple of all coins. | By level：Undervalued · bottom zone < 1.00；Low < 1.50；Fair < 2.40；Warm < 3.20；Overheated · top zone |
| MVRV Z-Score CoinMetrics（自算） | core | (Market cap − realized cap) ÷ historical std-dev of market cap — MVRV normalised across cycles. | By level：Bottom zone < 0.00；Low < 1.50；Fair < 3.00；Warm < 5.00；Overheated · top zone |
| Realized Price (spot premium) CoinMetrics（自算） | core | Network-wide average cost basis (price ÷ MVRV). Spot below it = the average holder is under water, historically a major bottom zone. Judged on spot's premium over it. | By spot premium：Below cost basis < 0.00%；Near cost basis < +30.00%；Fair premium < +140.00%；High premium |
| NUPL CoinMetrics（自算） | core | Net unrealized profit/loss = 1 − 1/MVRV. Same source as MVRV, used for sentiment phase only, not scored twice. | By level：Capitulation < 0.000；Hope / fear < 0.250；Optimism / anxiety < 0.500；Belief / denial < 0.750；Euphoria |
| Mayer Multiple CoinMetrics（自算） | aux | Price ÷ 200-day moving average. | By level：Deeply undervalued < 0.80；Below 200DMA < 1.00；Fair < 1.50；Warm < 2.40；Overheated |
| Puell Multiple CoinMetrics（自算） | aux | Daily miner issuance revenue ÷ its 365-day average. | By level：Miner capitulation < 0.50；Low < 1.00；Fair < 2.00；Warm < 3.50；Overheated |
| Pi Cycle Top CoinMetrics（自算） | aux | 111DMA ÷ (2 × 350DMA); ≥1 has marked past cycle tops. Top-zone warning only. | By ratio：Far from top signal < 0.900；Near top signal < 1.000；Top signal triggered |
L3 · Coin Flows（Who is moving coins on/off exchanges？）
| Indicator | Level / evidence | How to read | Zones |
|---|---|---|---|
| Exchange BTC net flow (7-day sum) CoinMetrics（flash 口径） | core | Inflow − outflow. Sustained outflow = coins leaving exchanges (hoarding); inflow = potential sell pressure. Flash data may be revised slightly next day. | By 7-day sum：Heavy outflow (withdrawals) < -10,000 BTC；Net outflow < -2,000 BTC；Balanced < +2,000 BTC；Net inflow < +10,000 BTC；Heavy inflow (sell pressure) |
| Exchange BTC balance (30d change) CoinMetrics | core | Total BTC held on exchanges — the most direct stock measure of holder behaviour. A long decline = holders self-custodying. ~2-week lag on the free tier; trend only, not scored. | By 30-day change：Falling (coins leaving) < -1.00%；Flat < +1.00%；Rising (coins returning) |
L4 · Sentiment & Derivatives（Is it about to overshoot？）
| Indicator | Level / evidence | How to read | Zones |
|---|---|---|---|
| Fear & Greed Index alternative.me | core | Sentiment thermometer from volatility, volume, social and search data. | By level：Extreme fear < 25；Fear < 45；Neutral < 56；Greed < 76；Extreme greed |
| BTC perp funding rate Hyperliquid（日均年化） | core | Who pays whom. Includes ~11% annualized base interest (same order as Binance 0.01%/8h), so 'normal' is not zero. | By annualized：Shorts crowded < -5.0%；Cool < 5.0%；Normal < 20.0%；Longs warm < 40.0%；Longs crowded |
| BTC open interest (7d change) Hyperliquid | core | BTC open interest on the largest on-chain perp venue; the 7-day change shows leverage building or flushing. Single venue. | By 7-day change：Deleveraging < -15.00%；Stable < +15.00%；Leverage piling up |
| Altseason index (own) CoinGecko（自算） | core | Share of the top-50 coins (ex stablecoins, wrapped and staked derivatives) beating BTC over 30 days. Own construction. | By share：Bitcoin season < 25.0%；BTC leaning < 50.0%；Alts leaning < 75.0%；Altseason |
| BTC dominance CoinGecko | aux | Cross-check for the altseason index. | By level：Trend only |
Claims tested and dropped
| Claim | Why |
|---|---|
| M2 / 'global liquidity leads BTC by 2–3 months' | 1 of 14 lead relations survives out-of-sample; the 0.94 level correlation is spurious (0.08 in returns) |
| Total stablecoin supply as crypto dry powder | Payment stablecoins −0.43 vs BTC, Tron ≈0; total +4.0% while trading bullets −4.2% |
| Real yield as the first L1 gauge | Weekly correlation −0.09, insignificant; now a state variable |
| Gold / 'gold highs are bullish BTC' | Insignificant since 2022; two spill-over definitions point opposite ways |
| Single FOMC decisions as direction | Direction mismatched in 7 events; volatility budget only |
| BTC tops align with real-yield lows | Median gap 96 days vs 110 random, p=0.40 |
| Oil as an optimism gauge | 0.04 correlation with the optimism factor since 2023 |
| 'Big ETF outflows = buy signal' | Four weeks after extreme outflows: −3.5% vs +1.9% baseline |
| 'Deribit basis = CME basis' | Same concept, different markets; levels differ by points |
| Pre-2018 data for any threshold | 2010–2017 macro explanatory power below the noise floor |
Not yet connected
| Layer | Indicator | Reason / plan |
|---|---|---|
| L2 | Realized Cap HODL Waves | Needs UTXO age distribution; paid Glassnode or own node |
| L3 | URPD | Paid Glassnode metric; planned self-build |
| L3 | STH-MVRV / LTH-MVRV | Needs realized cap by holding age; paid |
| L3 | STH-SOPR / aSOPR | Paid Glassnode metric |
| L3 | Miner reserves / illiquid supply | Paid Glassnode metric |
| L4 | Long/short ratio / liquidations | Coinglass needs an API key |
| L4 | Options put/call · IV skew | Free on Deribit; next version |
Log rules
Log date = UTC day of the run; each reading carries its own data date (FRED on US business days, NFCI weekly, exchange balance ~2 weeks late; stale readings get ⚠). Frozen once written; a same-day re-run only overwrites when fewer core inputs are missing, marked 'Re-run'. Weekly reviews follow ISO weeks, monthly ones calendar months.
Launchpad conventions
| Measure | Launchpads are always compared by fees, never volume (volume coverage is incomplete and understates share). |
| Period | Only 'yesterday' as a full UTC day is reported; today's data is never complete. |
| Matrix | All protocols in DefiLlama's Launchpad category, split by chain with DefiLlama's per-chain data; the sector trend uses daily top-60 fees since 2024-01. |
| Share | Share bars = % of the filtered total (chain × period × tier); deep-tracked share uses the top-60 daily total. |
| Late data | A source that has not published yesterday is marked late — never back-filled with old values or zero. |
| Flap / StonkFun | Flap is a tax-token model and StonkFun fees count only the platform's cut; compare them by revenue. |
| PONS burns | Sent to 0x…dEaD rather than burn(); market cap is always burn-adjusted. |
| Never | No 'it's dead' calls from comparing with all-time peaks; data only, no trade advice. |