SatoshiMacro Model: Ethereum (SMM-ETH)

Phase 3 ships all six tiers. Tier 2 (Validator/Staking) sources validator queue, slashings, and MEV per block from Beaconcha.in (with seeded historical anchors). Beacon APR and staking ratio are shown as informational because their series grow / decline monotonically with adoption and so carry no cycle-position signal under expanding-window percentile rank.

The SatoshiMacro Model for Bitcoin proved that institutional-grade cycle analysis can be published free and still hold up against the paywalled tools. SMM-ETH extends the same architecture to Ether - the asset that powers most of the institutional DeFi flow and the second-largest crypto market by capitalisation - because Australian-resident investors holding ETH deserve the same lens that Bitcoin holders now have.

Phase 1 ships what we have clean data for. Cycle timing and valuation are the two tiers with the longest, cleanest signal history for ETH, and they alone produce 4-of-5 in-zone calls across the 2018, 2021, and 2022 cycle inflections. The remaining tiers (validator/staking, sentiment, rotation, macro) will ship as their data sources land. The same transparency rules apply: every input value, every normalized score, every tier weight, every historical reading disclosed on this page. No paywall. No signup.

Bitcoin gets most of the cycle-reading attention because it's the largest and the simplest. But Ether is the asset where Australian retail allocation often actually sits - via the Coinbase / Binance defaults, via the staking yields, via the spot ETFs. If retail is going to hold ETH alongside BTC, retail should have the same cycle-confluence lens for both. Govind Satoshi · Former Institutional Trader · Founder, SatoshiMacro

For the canonical BTC version, see the SatoshiMacro Model (BTC). More context on the framework and why it's published free is on the About page.

SMM-ETH is a composite Ethereum cycle confluence model built from 28 signals across all six tiers. Each signal is computed at every historical date back to ETH's mainnet launch (2015-07-30), normalized to its own expanding-window percentile rank, and aggregated into tier scores. The tier scores are then weight-summed across live tiers only (renormalizing the weights for any tier whose data is undefined at that date - relevant for validator/staking signals before beacon-chain genesis, and for the ETH ETF signals before 2024-07-23), producing the final 0-100 SMM-ETH score.

Tier structure

• Tier 0: Cycle Timing & Mass Psychology (30%) - LIVE. Three signals: drawdown from ATH, days since ATH, and a rolling profitable-days proxy. The BTC-specific 4-year halving-cycle component is omitted because there is no analogous deterministic supply event for post-merge ETH. Days-since-merge and post-merge issuance dynamics will likely feed in as a future signal here.
• Tier 1: Valuation & Cycle Position (25%) - LIVE. Six ETH-calibrated valuation signals: Mayer Multiple (ETH price / 200-day MA), 200-Week MA Distance, 2-Year MA Multiplier, ETH Power Law deviation (regression of log(days since 2015-07-30) vs log(price) across the full ETH history), Rainbow Chart Position (mapped from the power-law deviation), and the ETH Risk Metric. Each gets an expanding-window percentile rank to keep historical readings lookahead-free.
• Tier 2: Validator & Staking (10%) - LIVE. Post-merge ETH is no longer mined, so the BTC miner-stress tier is replaced with three cycling validator-economics signals: validator entry queue length (high queue = retail euphoria entering = late-cycle), 30-day slashings count (network stress proxy that spikes during volatility events), and MEV per block (late-cycle on-chain activity / leverage). Two additional fields - beacon APR and staking ratio - are surfaced in the indicator panel as informational but excluded from the composite because their series grow / decline monotonically as ETH staking adoption increases. Monotonic series carry no cycle-position information under expanding-window percentile rank (they always score at the extremes), so including them would add bias without signal. Validator data begins at beacon-chain genesis (2020-12-01); pre-genesis dates renormalize across the five non-staking tiers.
• Tier 3: Sentiment & Positioning (20%) - LIVE. Seven sentiment + positioning signals: Crypto Fear & Greed Index, Google Trends (Bitcoin search interest, Buy Bitcoin search interest - both serve as broad-crypto risk-on proxies because crypto search demand moves as one block), plus four ETH-native derivatives: ETH funding rate aggregate (Binance + Bybit + OKX volume-weighted), ETH open interest (Binance ETH-PERP scaled cross-venue), Coinbase ETH-USD vs Binance ETH-USDT spot premium, and Deribit ETH options put/call ratio.
• Tier 4: Rotation & Institutional Flow (10%) - LIVE. Five signals: ETH/BTC ratio percentile (rising = ETH leading late-cycle), BTC Dominance (inverse - low dominance = ETH/alts running), Altcoin Season Index (direct - alt season = late cycle), US spot Ethereum ETF cumulative net flow percentile, and rolling 30-day net flow percentile. The ETF signals are undefined before launch (2024-07-23) and only the three rotation-only components contribute at pre-launch dates.
• Tier 5: Macro Headwinds/Tailwinds (5%) - LIVE. Four shared signals: USD Index (inverse - weak USD = risk-on = bullish ETH), M2 YoY growth (direct - liquidity = bullish ETH), Yield Curve 2s10s (steeper = accommodative = bullish ETH), VIX (inverse - low VIX = calm risk-on = bullish ETH).

Why renormalize weights across live tiers only?

For dates before beacon-chain genesis (2020-12-01), the validator/staking tier has no defined inputs. The alternative would be to set it to 50 (neutral midpoint), which silently drags the composite toward 50 regardless of where the live tiers actually point. Renormalizing means a pre-2020 cycle reading of 95 from the five non-staking tiers shows up as 95, not as some diluted value. The output JSON explicitly records which tiers contributed at each date so a consumer can interrogate the state.

ETH power-law fit

The power-law regression fits log(price) = a * log(days since 2015-07-30) + b across ETH's full price history. The current fit has slope around 2.3 (less steep than Bitcoin's ~5.7, reflecting ETH's shorter compounding history and different supply dynamics). The fit refreshes on every build, so the "fair value" reference line evolves as more history accumulates and the power-law deviation stays comparable across the historical chart.

ETH/BTC ratio anchoring

The ETH/BTC ratio is the canonical "is ETH leading or BTC leading" signal. ETH/BTC rising into and past 0.06 historically aligns with late-ETH-cycle phases (e.g. Jan 2018, Nov 2021). Below 0.04 historically aligns with BTC-led phases where ETH lags. SMM-ETH percentile-ranks the ETH/BTC ratio over expanding-window history; a high percentile contributes a high Tier 4 score, which contributes to a higher composite SMM-ETH reading (consistent with late-cycle ETH strength signalling distribution territory).

Calibration curve

SMM-ETH applies the same piecewise calibration curve as SMM-BTC after computing the raw weighted composite: identity below raw 40 (preserves bottoms), 1.5x slope from 40-60 (mid-cycle stretch), 2.0x slope from 60-70 (late-cycle stretch), 1.0x slope from 70-80 (top capping), and clamp to 100 above raw 80. The curve stretches the upper half of the scale so historical ETH cycle tops register firmly in the Cycle Top zone (85-100) while bottoms pass through unchanged. Both the calibrated smm and raw pre-calibration raw_smm are emitted in the output JSON. The curve is identical to SMM-BTC for cross-model consistency - same shape, same anchor points.

Caveats and limitations

• SMM-ETH starts at July 2015, not earlier. Ether mainnet launched 30 July 2015; the model has no pre-mainnet data to compute against. The earliest months also reflect raw price discovery with thin liquidity, so the first ~6 months of readings should be treated as price-history context rather than load-bearing model output. Same logic the SMM-BTC page applies to the pre-2013 Bitcoin period.
• Phase 4 = ETH-native data wired through Tier 3 + Tier 4. Tier 3 derivatives (funding, open interest, Coinbase premium, options put/call) now read from ETH-USDT / ETH-USD markets directly via assets/data/eth-derivatives.json. Tier 4 picks up two US spot Ethereum ETF flow signals (cumulative + rolling 30-day) on top of the original ETH/BTC ratio + dominance + altcoin-season trio. The BTC-side proxy phase ended in May 2026.
• 2022 cycle bottom still reads Neutral (~38). Even with full ETH-native data, the June 2022 reading lands around 38 - very close to the Accumulation boundary at 30 but not over it. The residual is driven by the rotation tier: ETH/BTC ratio at June 2022 was ~0.05, which sat in the upper half of the 2016-2022 distribution even though absolute ETH price had drawn down 75% from the Nov 2021 top. The signal is structurally honest (it reads "ETH still leading BTC" because that was true) and the broader composite is correctly Neutral, just one zone above the calibration target.
• Google Trends signals remain BTC-side proxies. "Bitcoin" and "Buy Bitcoin" search interest stand in for ETH-specific search demand because crypto search volume moves as one block. ETH-specific Google Trends time series with comparable historical depth would be a future refinement; the BTC proxies are documented on the indicator panel with the suffix "(broad-crypto proxy)".
• Percentile ranks use expanding-window history. Each historical date's score is computed only against indicator values available up to that date. No lookahead bias. Same approach as SMM-BTC.
• This is not financial advice. SMM-ETH is a research tool for ETH cycle position context. It does not predict prices. Cycle indicators have failed before and will fail again. Use as one input among many.