← All methodology
Methodology · Stress tests

What "BTC drops 30%" would actually do to your portfolio.

Three scenarios on the Analyze tab. Beta-weighted impact projections from first principles. No Monte Carlo, no hidden assumptions.

The three scenarios

  • BTC −30% correction. Classic mid-cycle drawdown. Single-trigger market-wide decline.
  • BTC −50% crash. Cycle-defining bear-onset event (2022-style).
  • Broad market selloff (−40%). Macro risk-off contagion. Different from a single-asset decline because correlations spike during these.

Each scenario shows the projected portfolio impact in % and dollars. Negative numbers are losses.

How the projection works

For each holding, we estimate beta-to-BTC and multiply by the scenario's BTC move:

impact(token) = weight × beta_to_btc × btc_move% impact(portfolio) = Σ over all tokens = sum of (weight × beta × btc_move%) dollarLoss = (impact% / 100) × totalPortfolioValue

Per-symbol beta lookup

v1 uses a hand-tuned beta table by asset class:

Stablecoins (USDT, USDC, etc.) → β = 0 BTC → β = 1.0 (reference) ETH → β = 1.15 SOL, AVAX, MATIC, ADA, BNB → β = 1.1 Everything else (default alts) → β = 1.4

These reflect typical 90-day rolling correlations during the 2022 to 2025 cycle. A real BTC −30% event historically takes ETH down ~35% and most alts down 40 to 50%.

What this gets right

  • Captures the first-order effect. High-beta portfolios suffer more in BTC drawdowns. Self-evident, but cleanly quantified.
  • Counts stablecoin buffer correctly. A portfolio with 30% stables shows ~30% less impact than a fully-deployed one.
  • Composable with the regime fit context. A 'Conflict' regime score plus a deep stress test result tells the same story.

What it does NOT do (yet)

  • Calculate proper time-series beta per token. The default β = 1.4 for all unknown alts oversimplifies. Some alts have β closer to 2.0+ in crashes, some closer to 1.0.
  • Model correlation spikes during stress. In a 2022-style event, correlations approach 1.0 across all crypto. Our linear model understates this tail risk.
  • Account for liquidity-driven slippage. Selling into a -50% move costs more than mid-cycle slippage. Not modeled.
  • Run Monte Carlo or VaR. We give point estimates, not distributions. v2 will add a sensitivity range.
  • Model second-order effects (cascading liquidations, MEV during volatility). These would deepen the projected losses.

How to read the output

Treat these numbers as a floor on potential drawdown, not a precise forecast. If our model says BTC −30% would take your portfolio down 35%, a real event might do 38 to 45% once correlations and slippage are factored in. The number tells you the structural exposure you have to the scenario, not the exact dollar outcome.

Stress Test v2 · coming next

The next iteration adds narrative-shock scenarios ("if AI sector cools 30%, what happens?") and an interactive slider so you can drag a sector heat by ±X% and see the portfolio impact in real time. Same first-principles math, extended along the sector axis.