tsbootstrap: Distribution-Free Uncertainty Quantification and Conformal Prediction for Time Series — explained | StockTools Quant Research

quant research // summarizing the paperLOADING

tsbootstrap: Distribution-Free Uncertainty Quantification and Conformal Prediction for Time Series

Sankalp Gilda · 2026-07-07

what this means for investors // AI summary of the abstractSituationally useful

This paper introduces tsbootstrap, an open-source Python library that combines time-series resampling methods (block, residual, sieve, wild bootstrap) with adaptive conformal prediction calibrators in one unified interface. In a controlled study it shows that the standard IID bootstrap badly underestimates uncertainty when data are dependent over time, while dependence-aware methods (the sieve method being closest to target under short-memory linear dependence) come closer to the correct coverage.

Why it matters: Financial time series are serially dependent, so naive resampling can produce prediction and confidence intervals that are too narrow and overstate certainty. This tool offers dependence-aware uncertainty quantification and prediction intervals that could give practitioners more honest error bars for forecasts, risk estimates, or backtested statistics.

The IID bootstrap sharply undercovers (intervals too narrow) when time series data are dependent — a common pitfall in finance.
Dependence-aware resampling reduces the coverage gap; the sieve method was closest to nominal under short-memory linear dependence.
The library bundles both bootstrap resampling and adaptive conformal calibrators (EnbPI, ACI, NexCP, AgACI) in one API, and runs faster and more memory-efficiently than a compared backend (arch).
It is free and MIT-licensed (v0.6.1), making rigorous uncertainty quantification accessible.

Applies to: uncertainty quantification and prediction intervals for dependent time series (forecasting, risk estimation)

⚠ This is a software/tooling paper with a controlled synthetic coverage study, not evidence of improved trading returns, and results depend on the assumed dependence structure matching your data.

ABSTRACT

Finance, sensing, and demand streams violate the exchangeability that IID conformal prediction and the IID bootstrap assume, and existing libraries implement either a general resampling engine or conformal calibration without the other. tsbootstrap provides block, residual, sieve, and wild resampling, classical bootstrap confidence intervals, and adaptive conformal calibrators (EnbPI, ACI, NexCP, AgACI) through a single typed API in which a specification object selects each method. In a controlled coverage study the IID bootstrap undercovers sharply under dependence; dependence-aware methods reduce the coverage deficit, the sieve nearest to nominal under short-memory linear dependence. On the shared fixed-statistic path a compiled backend runs several times faster than arch, and a streaming reduce avoids materializing the $O(Bn)$ replicate tensor, limiting peak extra memory to $O(B)$ for the statistic array. The software is MIT licensed (v0.6.1).

★ PRO · COMING SOON

Go deeper: a full research-committee breakdown of this paper, its assumptions and failure modes, and how its method would apply to a specific ticker or your watchlist. See StockTools AI →

AI summary generated from the paper’s public abstract via arXiv; it may miss nuance — read the source before relying on it. Thank you to arXiv for its open-access interoperability; StockTools is not affiliated with arXiv, and all rights remain with the authors. Educational only, not financial advice.