Methodology
How we measure
Open, timestamped, reproducible — not a black-box score.
Micro-quote baseline
A $200 sell quote establishes the execution price baseline. All ladder rungs measure relative impact against this baseline, not against the external oracle price. This separates size-dependent slippage from premium/discount to reference.
Sell ladder
Seven notionals: $1k, $10k, $50k, $100k, $250k, $500k, $1M. For each, we request a Jupiter sell quote (token → USDC) and take the median of 3 samples.
Cliff bisection
When adjacent rungs show >15% efficiency drop AND route composition changes, we bisect between them to refine the cliff boundary.
Scoring
Exit capacity at 99%/95%/90% thresholds via log-linear interpolation. No-route ceiling, mcap-to-exit ratio, RFQ dependence, cliff detection, and letter grades A–F.
Guards against known bugs
Decimals always from Jupiter price/v3 API — never assumed. Flat-across-sizes curves with nonzero offset are rejected (external-price baseline error). Regression fixtures lock SPYx, STONK, STRCx cliff, and FTR no-route.
Research tool only. Not investment advice. Executable quotes reflect router state at probe time and may change.
This tool measures executable router quotes at probe time. It does not predict future liquidity, guarantee execution, or constitute investment advice. Past probe results do not guarantee future exit capacity.
Measured 13 Sep 2026, US equity markets closed · Jupiter router quotes