### Prequential probability: principles and properties

A. Philip Dawid and Vladimir G. Vovk
Source: Bernoulli Volume 5, Number 1 (1999), 125-162.

#### Abstract

Forecaster has to predict, sequentially, a string of uncertain quantities , whose values are determined and revealed, one by one, by Nature. Various criteria may be proposed to assess Forecaster's empirical performance. The weak prequential principle requires that such a criterion should depend on Forecaster's behaviour or strategy only through the actual forecasts issued. A wide variety of appealing criteria are shown to respect this principle. We further show that many such criteria also obey the strong prequential principle, which requires that, when both Nature and Forecaster make their choices in accordance with a common joint distribution for , certain stochastic properties, underlying and justifying the criterion and inferences based on it, hold regardless of the detailed specification of . In order to understand further this compliant behaviour, we introduce the prequential framework, a game-theoretic basis for probability theory in which it is impossible to violate the prequential principles, and we describe its connections with classical probability theory. In this framework, in order to show that some criterion for assessing Forecaster's empirical performance is valid, we have to exhibit a winning strategy for a third player, Statistician, in a certain perfect-information game. We demonstrate that many performance criteria can be formulated and are valid in the framework and, therefore, satisfy both prequential principles.

First Page:
Full-text: Open access

Permanent link to this document: http://projecteuclid.org/euclid.bj/1173707098
Mathematical Reviews number (MathSciNet): MR1673572
Zentralblatt MATH identifier: 0929.60001

### References

[1] Billingsley, P. (1986) Probability and Measure, 2nd edn. New York: Wiley.
Mathematical Reviews (MathSciNet): MR830424
[2] Cox, D.R. and Lewis, P.A.W. (1966) Statistical Analysis of Series of Events. London: Methuen.
Mathematical Reviews (MathSciNet): MR199942
[3] Dawid, A.P. (1982) The well-calibrated Bayesian (with discussion). J. Amer. Statist. Ass., 77, 605-613.
Mathematical Reviews (MathSciNet): MR675887
Zentralblatt MATH: 0495.62005
Digital Object Identifier: doi:10.2307/2287720
[4] Dawid, A.P. (1983) Inference, statistical: I. In S. Kotz, N.L. Johnson and C.B. Read (eds), Encyclopedia of Statistical Sciences, Vol. 4, pp. 89-105. New York: Wiley-Interscience.
Mathematical Reviews (MathSciNet): MR719029
[5] Dawid, A.P. (1984) Statistical theory: the prequential approach (with discussion). J. Roy. Statist. Soc. A, 147, 278-292.
Mathematical Reviews (MathSciNet): MR763811
Digital Object Identifier: doi:10.2307/2981683
[6] Dawid, A.P. (1985) Calibration-based empirical probability (with discussion). Ann. Statist., 13, 1251-1285.
Mathematical Reviews (MathSciNet): MR811493
Zentralblatt MATH: 0587.60002
Digital Object Identifier: doi:10.1214/aos/1176349736
Project Euclid: euclid.aos/1176349736
[7] Dawid, A.P. (1986) Probability forecasting. In S. Kotz, N.L. Johnson and C.B. Read (eds), Encyclopedia of Statistical Sciences, Vol. 7, pp. 210-218. New York: Wiley-Interscience.
Mathematical Reviews (MathSciNet): MR892738
[8] Dawid, A.P. (1991) Fisherian inference in likelihood and prequential frames of reference (with discussion). J. Roy. Statist. Soc. B, 53, 79-109.
Mathematical Reviews (MathSciNet): MR1094276
[9] Dawid, A.P. (1992a) Prequential data analysis. In M. Ghosh and P.K. Pathak (eds), Current Issues in Statistical Inference: Essays in Honor of D. Basu pp. 113-126. IMS Lecture Notes,Monograph Ser. 17. Hayward, CA: Institute of Mathematical Statistics.
Mathematical Reviews (MathSciNet): MR1194413
Zentralblatt MATH: 0850.62091
Digital Object Identifier: doi:10.1214/lnms/1215458842
[10] Dawid, A.P. (1992b) Prequential analysis, stochastic complexity and Bayesian inference (with discussion). In J.M. Bernardo, J. Berger, A.P. Dawid and A.F.M. Smith (eds), Bayesian Statistics 4, pp. 109-125. Oxford: Oxford University Press.
Mathematical Reviews (MathSciNet): MR1380273
[11] de Finetti, B. (1974) Theory of Probability, Vol 1. London: Wiley.
[12] Doob, J.L. (1953) Stochastic Processes. New York: Wiley.
Mathematical Reviews (MathSciNet): MR58896
[13] Hill, T.P. (1982) Conditional generalizations of strong laws which conclude the partial sums converge almost surely. Ann. Probab., 10, 828-830.
Mathematical Reviews (MathSciNet): MR659552
Digital Object Identifier: doi:10.1214/aop/1176993792
Project Euclid: euclid.aop/1176993792
[14] Loveland, D. (1969) On minimal program complexity measures. Proceedings of the First ACM Symposium on the Theory of Computing, pp. 61-66. New York: ACM Press.
[15] Martin, D. (1975) Borel determinacy. Ann. Math., 102, 363-371.
Mathematical Reviews (MathSciNet): MR403976
Digital Object Identifier: doi:10.2307/1971035
[16] Martin, D. (1990) An extension of Borel determinacy. Ann. Pure Appl. Logic, 49, 279-293.
Mathematical Reviews (MathSciNet): MR1077261
Zentralblatt MATH: 0721.03036
Digital Object Identifier: doi:10.1016/0168-0072(90)90029-2
[17] Minozzo, M. (1996) On some aspects of the prequential and algorithmic approaches to probability and statistical theory. PhD Thesis, University of London.
[18] Rosenblatt, M. (1952) Remarks on a multivariate transformation. Ann. Math. Statist., 23, 470-472.
Mathematical Reviews (MathSciNet): MR49525
Zentralblatt MATH: 0047.13104
Digital Object Identifier: doi:10.1214/aoms/1177729394
Project Euclid: euclid.aoms/1177729394
[19] Schnorr, C.P. (1970) Klassifikation der Zufallsgesetze nach Komplexität und Ordnung. Z. Wahrscheinlichkeitstheorie Verw. Geb., 16, 1-21.
Digital Object Identifier: doi:10.1007/BF00538763
[20] Schnorr, C.P. (1971a) A unified approach to the definition of random sequences. Math. Systems Theory, 5, 246-258.
Mathematical Reviews (MathSciNet): MR354328
Digital Object Identifier: doi:10.1007/BF01694181
[21] Schnorr, C.P. (1971b) Zufälligkeit und Wahrscheinlichkeit. Berlin: Springer-Verlag.
[22] Seillier-Moiseiwitsch, F. and Dawid, A.P. (1993) On testing the validity of sequential probability forecasts. J. Am. Statist. Assoc., 88, 355-359.
Mathematical Reviews (MathSciNet): MR1212496
Zentralblatt MATH: 0771.62058
Digital Object Identifier: doi:10.2307/2290731
[23] Shafer, G. (1976) A Mathematical Theory of Evidence. Princeton, NJ: Princeton University Press.
Mathematical Reviews (MathSciNet): MR464340
Zentralblatt MATH: 0359.62002
[24] Shafer, G. (1996) The Art of Causal Conjecture. Cambridge, MA: MIT Press.
[25] Shiryayev, A.N. (1984) Probability. New York: Springer-Verlag.
Mathematical Reviews (MathSciNet): MR737192
Zentralblatt MATH: 0536.60001
[26] Stout, W.F. (1970) A martingale analogue of Kolmogorov's law of the iterated logarithm. Z. Wahrscheinlichkeitstheorie Verw. Geb., 15, 279-290.
Mathematical Reviews (MathSciNet): MR293701
Digital Object Identifier: doi:10.1007/BF00533299
[27] Stout, W.F. (1974) Almost Sure Convergence. New York: Academic Press.
Mathematical Reviews (MathSciNet): MR455094
[28] Ville, J. (1939) Etude Critique de la Notion de Collectif. Paris: Gauthier-Villars.
[29] Vovk, V.G. (1987) The law of the iterated logarithm for random Kolmogorov, or chaotic, sequences. Theory Probab. Applic., 32, 413-425.
Mathematical Reviews (MathSciNet): MR914936
[30] Vovk, V.G. (1988) Kolmogorov-Stout law of the iterated logarithm. Math. Notes, 44, 502-507.
Mathematical Reviews (MathSciNet): MR962372
[31] Vovk, V.G. (1993a) A logic of probability, with application to the foundations of statistics (with discussion). J. Roy. Statist. Soc. B, 55, 317-351.
Mathematical Reviews (MathSciNet): MR1224399
[32] Vovk, V.G. (1993b) Forecasting point and continuous processes: prequential analysis. Test, 2, 189-217.
Mathematical Reviews (MathSciNet): MR1265490
Zentralblatt MATH: 0812.60041
Digital Object Identifier: doi:10.1007/BF02562675
[33] Vovk, V.G. (1996) A purely martingale version of Kolmogorov's strong law of large numbers. Theory Probab. Applic., 41, 605-608.
Mathematical Reviews (MathSciNet): MR1450080