Asymptotic uniqueness in long-term prediction

Vladimir Vovk

Asymptotic uniqueness in long-term prediction

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

Abstract

This paper establishes the asymptotic uniqueness of long-term probability forecasts in the following form. Consider two forecasters who repeatedly issue probability forecasts for the infinite future. The main result of the paper says that either at least one of the two forecasters will be discredited or their forecasts will converge in total variation. This can be regarded as a game-theoretic version of the classical Blackwell-Dubins result getting rid of some of its limitations. This result is further strengthened along the lines of Richard Jeffrey’s radical probabilism.

Original language	English
Title of host publication	Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications
Subtitle of host publication	Proceedings of Machine Learning Research
Editors	Simone Vantini, Matteo Fontana, Aldo Solari, Henrik Bostrom, Lars Carlsson
Pages	90-104
Number of pages	15
Volume	230
Publication status	Published - 10 Sept 2024

Keywords

Long-term prediction
Probabilistic prediction
Jeffreys's law
Radical probabilism

Access to Document

https://proceedings.mlr.press/v230/vovk24a.html

Cite this

@inbook{0d811fc50d8e49b28a95b0fba9934cfa,

title = "Asymptotic uniqueness in long-term prediction",

abstract = "This paper establishes the asymptotic uniqueness of long-term probability forecasts in the following form. Consider two forecasters who repeatedly issue probability forecasts for the infinite future. The main result of the paper says that either at least one of the two forecasters will be discredited or their forecasts will converge in total variation. This can be regarded as a game-theoretic version of the classical Blackwell-Dubins result getting rid of some of its limitations. This result is further strengthened along the lines of Richard Jeffrey{\textquoteright}s radical probabilism.",

keywords = "Long-term prediction, Probabilistic prediction, Jeffreys's law, Radical probabilism",

author = "Vladimir Vovk",

year = "2024",

month = sep,

day = "10",

language = "English",

volume = "230",

pages = "90--104",

editor = "Simone Vantini and Matteo Fontana and Aldo Solari and Henrik Bostrom and Lars Carlsson",

booktitle = "Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications",

}

Asymptotic uniqueness in long-term prediction. / Vovk, Vladimir.
Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications: Proceedings of Machine Learning Research. ed. / Simone Vantini; Matteo Fontana; Aldo Solari; Henrik Bostrom; Lars Carlsson. Vol. 230 2024. p. 90-104.

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - Asymptotic uniqueness in long-term prediction

AU - Vovk, Vladimir

PY - 2024/9/10

Y1 - 2024/9/10

N2 - This paper establishes the asymptotic uniqueness of long-term probability forecasts in the following form. Consider two forecasters who repeatedly issue probability forecasts for the infinite future. The main result of the paper says that either at least one of the two forecasters will be discredited or their forecasts will converge in total variation. This can be regarded as a game-theoretic version of the classical Blackwell-Dubins result getting rid of some of its limitations. This result is further strengthened along the lines of Richard Jeffrey’s radical probabilism.

AB - This paper establishes the asymptotic uniqueness of long-term probability forecasts in the following form. Consider two forecasters who repeatedly issue probability forecasts for the infinite future. The main result of the paper says that either at least one of the two forecasters will be discredited or their forecasts will converge in total variation. This can be regarded as a game-theoretic version of the classical Blackwell-Dubins result getting rid of some of its limitations. This result is further strengthened along the lines of Richard Jeffrey’s radical probabilism.

KW - Long-term prediction

KW - Probabilistic prediction

KW - Jeffreys's law

KW - Radical probabilism

M3 - Chapter (peer-reviewed)

VL - 230

SP - 90

EP - 104

BT - Proceedings of the Thirteenth Symposium on Conformal and Probabilistic Prediction with Applications

A2 - Vantini, Simone

A2 - Fontana, Matteo

A2 - Solari, Aldo

A2 - Bostrom, Henrik

A2 - Carlsson, Lars

ER -