REFI isolates the financing dimension of the real estate cycle. The rate complex is intentionally dominant: 30-year mortgage rates (15%), 10-year yields (12%), 2-year yields (8%), and the Fed Funds rate (EFFR, 10%) together capture the full curve affecting real estate borrowing - rates are the binding constraint for housing affordability and refinancing activity. The HYG/LQD ratio (12%) provides credit spread context, while delinquency rates (DRSFRMACBS, 10%) and business loans (BUSLOANS, 12%) signal lending health and origination activity. Employment (6%), permits (5%), and VIX (10%) round out. REFI is rate-heavy by design because it answers a specific question: 'How easy and cheap is it to finance real estate right now?' Financing 'works' only when rates, spreads, and stress all cooperate, so the weight structure forces that multi-signal confirmation. Methodology note: Uses REIT ETFs, mortgage rates, and FRED macro proxies. Strongest of the three RE indices but still equity/rate-driven — not direct mortgage origination volume, CMBS issuance, or loan-level data.

REIL captures the broad health of U.S. commercial real estate equity. VNQ (15%) and IYR (12%) provide diversified REIT exposure as the cleanest public proxies for lease duration, occupancy, and contractual rent growth, with ITB (7%) for homebuilders and SPY (4%) for relative context. Gold (6%) adds an inflation-hedge dimension. The rate complex dominates: mortgage rates (MORTGAGE30US, 14%) and the 10-year yield (DGS10, 13%) together command 27% of weight because real estate is fundamentally a rate story - when rates fall, property values rise, financing improves, and REITs appreciate. Employment (6%) and housing starts (7%) confirm the real-economy property cycle, while the dollar (6%) and VIX (10%) provide macro and risk sentiment context. The weight design forces multi-signal confirmation - leasing improvement must be accompanied by supportive rates and conditions, not just a single factor moving. Methodology note: REIL captures listed REIT equity sentiment and rate conditions, not direct property values, NOI, cap rates, or occupancy data. Treat as a directional indicator for listed real estate, not specialist-grade CRE measurement.

CMAI is a crypto macro regime composite blending Binance funding rates with traditional macro signals. The largest weights sit on BTC (10%) and ETH (10%) as the two anchor assets, with Solana (8%) measured relative to BTC to capture altcoin breadth - when alts outperform Bitcoin, it signals broad-based participation. Institutional vehicles IBIT (8%) and Coinbase (COIN, 8%) track institutional and exchange-level activity. The BTC perpetual funding rate (10%) provides real-time leverage sentiment. On the macro side, VIX (10%) captures TradFi risk contagion, short rates (DGS2, 10%) and the HYG/LQD ratio (8%) measure rate pressure and credit risk appetite, and dollar weakness (8%) and consumer credit growth (TOTALSL, 10%) provide the macro liquidity backdrop. CMAI rises when both crypto-native price action and macro liquidity permission are aligned, and fades when price is carried by a narrow cohort under tightening conditions. Methodology note: Funding rate components use raw period-over-period deltas, not interval-normalized flow. Institutional flow is proxied via IBIT ETF and COIN equity, not direct exchange deposit/withdrawal or on-chain flow data. Treat as a directional crypto activity indicator.

CVXI incorporates crypto volatility and stress regime context across nine weighted inputs. BTC price action (12%) and ETH returns (8%) provide trend context, while MARA equity (7%) serves as a miner-equity proxy — not direct hashrate or mining economics data. A daily-close volatility proxy (10%, negative polarity) incorporates recent BTC price turbulence via 60-day rolling standard deviation of log returns. VIX (15%, negative) reflects TradFi stress spillover, and BTC funding rates (12%, negative) reflect leveraged positioning. The HYG/LQD ratio (12%), short rates (10%), and the dollar (14%) provide macro context. When the volatility proxy is low, VIX is calm, and funding rates are moderate, CVXI tends to rise — historically coinciding with environments where directional crypto strategies have performed better. Methodology note: Uses daily-close volatility proxy (60-day rolling SD of log returns), not implied volatility from options markets or specialist-grade realized variance. MARA equity is a listed-stock proxy for mining sentiment, not direct on-chain or operational mining data. Funding rate components use raw deltas, not interval-normalized flow. Treat as a directional regime indicator, not a specialist volatility tool.

CRSI measures the market's pricing of credit risk - it answers 'how much risk premium is the market charging?' rather than measuring actual defaults. HY and IG spread proxies (HYG, 13%; LQD, 12%) are top-weighted because they compress high-yield and investment-grade repricing into liquid, high-frequency series without requiring licensed OAS feeds. VIX (11%) and NFCI (8%) capture liquidity and risk-aversion regimes where spreads can gap even without immediate fundamental deterioration. Rate context - the 10-year yield (8%) and T10YFF spread (10%) - helps distinguish 'spreads widening because risk premium rose' from 'everything moved because rates moved.' Gold (10%, positive) captures flight-to-safety, unemployment (5%) and credit card standards (7%) ensure the index doesn't misread purely technical spread moves as fundamental regime shifts. Bank equity health (BAC, 7%, negative) tracks banking conditions. When CRSI spikes, credit markets are pricing significant default risk, even if actual delinquencies haven't materialized yet.

DRSI measures stress in the duration/rates complex — is this a duration shock, and how bad is it?

MSPM measures whether macro data is surprising in a growth-supportive or growth-threatening direction. The highest weights go to surprise and momentum measures: employment (12%), CPI (8%), retail sales (8%), and other macro inputs use a macro-impulse transform with a 15-day half-life to capture surprise momentum before it reaches earnings. The index then translates surprises into investable risk regimes: the 10-year yield (10%), oil (10%), equities (SPY, 8%), and gold (8%) all carry positive polarity because rising prices confirm the macro surprise narrative. VIX (12%, negative) captures the uncertainty that undermines the signal. Dollar weakness (6%) adds a de-dollarization signal. The fast 15-day half-life means MSPM responds quickly to data surprises but also decays quickly, making it a real-time pulse rather than a trend measure - 'surprise' is the cleanest way to capture regime shifts before they show up in earnings.

RSKP is a bounded risk-on/risk-off posture gauge (rebased 0–100, where 50 is neutral).

Sprint 10C rebuild: PWGI now measures power availability and grid constraint conditions specifically, refocused from generic 'electrification' to the question that matters for AI infrastructure: is power becoming the bottleneck? The most important design change is making merchant power companies the core signal (30%): Constellation Energy (12%) vs XLU captures nuclear baseload premium and data center PPA demand, Vistra (10%) vs XLU captures gas/nuclear power scarcity pricing, and NRG (8%) vs XLU adds retail power and demand-side confirmation. When merchant power outperforms regulated utilities, power is scarce and valuable — this is the single best public proxy for power tightness. Grid equipment makers (20%): Eaton (10%) vs XLI for transformers/switchgear, Powell Industries (5%) vs XLI for power distribution, and Quanta Services (5%) vs XLI for utility construction and grid modernization. XLU (8%) vs SPY provides the broad utility anchor. Industrial production (10%) and durable goods (8%) confirm physical power demand and equipment orders. The financing constraint sleeve (24%): 30-year yield (10%) for long-duration project financing, EFFR (7%) for policy rate pressure, and VIX (7%) for infrastructure capex confidence. Methodology note: PWGI uses merchant power equities and macro proxies — not ISO grid data, interconnection queues, or regional wholesale power prices. Merchant power outperformance is the best available public proxy for power scarcity, but conflates energy market dynamics with true grid constraint.

Sprint 10C rebuild: ACSI now measures AI compute capex intensity specifically — the hyperscaler investment cycle that drives the entire AI infrastructure stack. Stripped of cloud/software/consumer noise (CLOU, IGV, RSAFS removed) and refocused on compute hardware capex. Semiconductor demand breadth (SMH, 15% vs SPY) is the primary signal because chip demand is the earliest-available proxy for hyperscaler capex acceleration. Two hyperscaler equities — Microsoft (10% vs QQQ) and Alphabet (10% vs QQQ) — are benchmarked against QQQ so they measure AI capex confidence specifically, not broad tech momentum. Networking capex follows compute: Arista (8% vs QQQ) derives 70%+ of revenue from hyperscalers, and Broadcom (7% vs QQQ) captures custom AI silicon (TPUs, ASICs) and networking chip demand. The physical capex confirmation sleeve (16%) — durable goods orders (8%) and industrial production (8%) — separates real equipment orders from equity multiple expansion. The constraint sleeve (34%) is intentionally heavy: 10-year yield (12%) for long-duration capex discount rates, LQD (8%) for corporate funding conditions, STLFSI4 (8%) for systemic financial stress, and VIX (6%) for capex confidence. ACSI is now the primary flagship for the AI, Power & Buildout family — it answers: 'Is the AI compute capex cycle accelerating or decelerating?' When ACSI is strong but DCAI is weak, demand exceeds buildout capacity. When ACSI weakens but DCAI remains strong, we're in a capex overshoot. Methodology note: ACSI uses hyperscaler equities and semiconductor ETFs — not direct capex filings, GPU allocation data, or chip fab utilization rates.

GTRI is the single global trade gauge — rebuilt in Sprint 10D by stripping US-specific components (UNRATE, PCEPI) and adding physical trade confirmation. Industrial commodity demand (22%) anchors the index: COPX (12%) is the most trade-sensitive industrial metal, DBC (10%) provides broad commodity demand context. EEM (10%) captures EM exporter health — most large-cap EM companies are export-oriented. The FX trade channel (22%) tracks the dollar headwind: DTWEXBGS (12%, negative = strong dollar constrains trade) and DEXCHUS (10%, negative = stronger CNY = China importing more). Physical trade confirmation (18%) separates real demand from financial speculation: INDPRO (10%) confirms manufacturing output, ISRATIO (8%, negative = falling inventory/sales = demand pulling goods through trade). LQD (8%) and T10YFF (8%, negative) capture trade-chain financing and growth-scare risk. VIX (7%, negative) is the trade disruption amplifier. Oil (DCOILWTICO, 5%) reflects physical transport demand. GTRI answers: 'Is global trade expanding or contracting?' — rising GTRI = healthy demand pull; falling GTRI = trade deterioration.