Ranking Cross-Venue Funding Carries under an Estimation-Error Budget

Cross-venue funding-carry rankings compare noisy current scenarios, and selecting the maximum can change the relevant error distribution. We derive an accounting error budget: for current scenario value S0, realized net accounting Pi and overestimation D=S0-Pi, Pi>0 if and only if D<S0. We separate this baseline from a preference- and distribution-dependent decision objective that remains unavailable when joint errors or capital inputs are missing. Using 9,192 public hourly funding records, a previous-settled benchmark has holdout MAE of 0.0126-0.0298 basis points across six series. Calibration-period 90% intervals attain 68.39% coverage for Hyperliquid ETH; a one-sided rule covers 94.03% of all holdout spread counterfactuals but 88.39% of selected maxima. In a five-route public-book paper illustration, current and subsequent scenario values are all negative and reconcile with negligible numerical residuals. The results establish an exact ranking and audit framework, not positive-expectation evidence or a calibrated decision rule.