How a Simple Fair Value Model Anchors Crypto Portfolio Sizing: BTC, ETH, and ZEC in Focus

The framework is a two-regime Markov model with observable states, not a calibrated HMM

Regime 1 (Liquidity risk-on) is indicated by positive 5-day net $BTC / $ETH ETF flows, declining $DXY, funding below 0.03% per 8 hours, and BTC above 77.1k then 81.3k.

It implies overweighting ETH/ $SOL relative to BTC, with only reduced $ZEC positions.

Regime 2 (Tightening/range) is marked by negative BTC ETF flows, failure at 82k, unwinding open interest, and Fed/Senate events. It suggests focusing on core BTC, fading alt-coin extensions, and riding volatility events on XRP/ZEC.

As of September 12, the state is Regime 2 with rotation pockets in ETH. BTC is capped below 82k, flows are unstable, leverage is partially flushed, and ETH is absorbing the flow. This is neither a capitulation bear market nor a clean institutional breakout.

A simple fair model serves as an anchor, not a target: Pi = βi * P_BTC + αi + εi, where εi represents jumps from news, ETFs, or squeezes.

Empirical 30-day beta estimates for 2026 are approximately: ETH 1.3, XRP 1.5, SOL 1.6, and ZEC much greater than 2 due to its idiosyncratic privacy/ETF regime.

The ZEC residual ε has dominated its BTC signal over the past year. Volatility is estimated using public proxies, as no GARCH is used here: BTC DVOL around 40%, ETH DVOL around 54%, and realized 30-day ZEC volatility in the triple digits.

For sizing, a fractional Kelly formula is used: weight proportional to μ/σ², with a cap at 0.25 of full Kelly. For ZEC, extreme sigma keeps its structural weight low despite a high narrative mu.

Risk matrix (desk) Factors

FOMC Sep 15–16, 3–10d: high for BTC, ETH, XRP, SOL, very high for ZEC (beta); Clarity Sep 15, 1–5d: low for BTC/ETH/ZEC, dominant for XRP; ETF flows: dominant BTC, important ETH, secondary XRP/SOL, via beta ZEC; Leverage crowding 1–20d: moderate BTC/ETH, high SOL/ZEC, critical ZEC; Privacy regulation/listings 3–18m: structural, N/A for five; Weekend liquidity 48h: low BTC, medium ETH/XRP, high SOL/ZEC, very high ZEC. Illustrative VaR 1-day, 95%, normal: VaR=1.65sdweight*capital. sd=2.5% BTC, 3.5% ETH, 4.5% XRP/SOL, 7.5%+ ZEC. Equal weight book: disproportionate ZEC contribution. Thus constraint: ZEC ≤5–8% speculative sleeve, 0% core.

Reading by crypto asset

|a| BTC functions as a reserve asset under a technical ceiling. The thesis posits that institutional savings (ETF AUM ~$98B) remain intact, but marginal demand is unstable. The market has twice sold off at 82k. COT data shows leveraged funds are still net short, suggesting a potential squeeze if 82k breaks alongside ETF inflows exceeding +$300M/day for multiple sessions, which is not the current tape.

Key levels: support at $74,700–$75,000, then $72,700; pivot at $77,100–$77,500 (near max pain); resistance at $81,300–$82,400 (double failure). A bearish invalidation would be a daily close below $74,500 with three consecutive sessions of negative ETF flows. Risks include FOMC, real yields, ETF outflows, and dealer hedging around the $75k/$80k walls.

|b| ETH shows better micro-structure this week. A measurable rotation is evident: +$216M in ETH ETFs versus -$13M for BTC on September 11. ETH IV is sometimes at a premium or discount to BTC, indicating the market is paying for ETH optionality. ETH has a residual positive performance vs BTC over 30 days, though this does not yet confirm an “ETH season” regime as BTC dominance remains above 50%. Key levels: support at $2,325–$2,400; resistance at $2,545–$2,570, then $2,700. Invalidation would come from a loss of $2,325 and a reversal of ETH ETF flows.

|c| XRP faces a binary legislative catalyst, with ETF flows too thin to support the price. Cumulative ETF AUM is about $1.5–$1.7B, but days with zero flows are frequent. The dated catalyst is the Senate cloture vote on CLARITY / H.R.3633 around September 15 (requires 60 votes; Republican majority has 53). Polymarket gives ~18% odds on a “signed in 2026” contract, while Galaxy previously cited ~10% passage in 2026. This creates an asymmetry: real event-driven upside against a low base rate. Key levels: support zone at $1.25–$1.30 for a bounce; recent supply at $1.44–$1.50. Fundamental invalidation would be failure of cloture without a replacement ETF bid.

|d| SOL has high beta, with on-chain fundamentals less supportive than its 30-day price suggests. While SOL is up +36% over 30 days, TVL on apps has declined from a 2025 peak (~$11.5B) to around $5.5B as of a September note. SOL ETFs exist (AUM ~$1.4B) but with marginal flows. The roadmap (Agave 4.3 in October, Alpenglow) represents a real option for throughput, not current cash flow. On-chain tokenization of stocks remains the only structural narrative. Key levels: floor to reclaim at $97.5–$100; supply at $105–$110. A loss of $95 would signal a de-rating regime.

|e| ZEC is the best 12-month trade but has the worst crowding profile today. This is a documented thesis, not a desk legend. It is up +2,000% over 12 months, with a market cap of ~$19B, placing it in the top 10. The spot ETF ZCSH (from Grayscale conversion on August 25, listed on NYSE Arca) has seen AUM grow from ~$304M to >$400M. The halving occurred in November 2024, with inflation at ~2%. The shielded pool holds ~28–35% of supply. The SEC investigation closed in January 2026 without action. The Ironwood upgrade (2026) will migrate away from Orchard. Open interest in perpetuals reached $2.4B, with short liquidations of ~$34M at the $1,000 break. Bitwise’s Hougan has cited ZEC as a privacy sleeve alongside BTC. The trap is that the “institutionalization of privacy” trade is already priced in. EU MiCA Article 76(3) remains an overhang for regulated listings of assets with native anonymization. An invisible issuance bug is a specific tail risk for zk. The high volatility and OI make ZEC a violent mean-reversion asset, not a core pension-sized holding. Key levels: supply at $1,200–$1,250 (post-ETF spike); support at $1,000, then $880–$900. Below $1,000 with OI not rebuilding would mark the end of the squeeze regime.