Betting on residual life: The caveats of conditioning
Introduction
In the process of using marker data to assess the lifetime of an item experiencing failure due to ageing, we were confronted by a dilemma that sneaked upon us as a matter of course (see Singpurwalla, 2006a). It turns out that the scenario leading to the dilemma is quite common and can arise when addressing practical issues of conditioning in the actuarial, the engineering, and the biomedical sciences. Stripped to its essentials, the scenario goes as follows.
Suppose that an item's lifetime X is judged to have a distribution function , and a survival function . We suppose that lifetime can be continuously monitored so that . Were this item supposed to survive until x, its residual (or remaining) lifetime will be . We are required to make statements of uncertainty about , so that actuarial, engineering, or medical decisions about the item can be made. That is, we are required to specify , for all . Our interpretation of probability is de Finnetian (see de Finetti, 1937), in the sense that probability reflects one's disposition to a two-sided bet. Thus, probability assessments can be seen as a device for hedging our bets on the item's survival, or some other unknown quantity of interest, such as parameters in probability models.
A solution to the problem posed is elementary and unique, given a distribution function G. Specifically, for any
Suppose now, that instead of the subjunctive, “were the item to survive until x”, we are told that the item actually did survive to x. That is, the event is no more an uncertain event; has now become observed data. What then would our assessment of the uncertainty about the residual life be? In other words, how would we bet on the event , for ? Would it continue to be , or could it be something else? If the latter, would the number to bet be unique? For a discussion of these and related questions, one may visit Freedman and Purves (1969). A more recent discourse on the different kinds of conditional beliefs is in Joyce (1999, Chapter 6).
Intuitively, it seems that there ought to be some distinction between looking at as a possibility, versus looking at it as a fact that is revealed as data. Thus, need not be the correct answer. Yet many individuals when faced with this problem would simply mimic the steps leading to Eq. (1.1) and continue to declare as their answer. In doing so they do not appear to be making a distinction between as a supposition versus a reality. Alternatively put, they may be failing to recognize the connotation that in a conditional probability statement, the word “given” does not indicate a fact; rather it indicates a supposition that the conditioning event is true. Thus, are those who declare as their answer—irrespective of the character of the conditioning event—in error, or is there a rationale for their answer?
We claim that the rationale cannot completely be within the calculus of probability, because the notion of probability—at least from a subjectivistic point of view—is germane only when the disposition of all events in question is unknown. Thus, for example, it may not make sense to say that the probability that a coin with heads on both faces when flipped will land heads, is one. This is because the disposition of the outcome is known before the flip. Consequently, a two-sided bet on the outcome heads has to be $1, which will be exchanged for a $1 when the coin lands heads, which it will. The two-sided bet of $1 is thus meaningless. The rationale therefore must come from concepts in statistics wherein the notion of a likelihood plays a signal role. By all accounts the notion of a likelihood appears to be alien to probability theory.
In what follows we point out that there are both philosophical and technical arguments which support as an answer, but that this answer is one among other possible answers. This is the main point of this article. Arguments about conditioning are common among philosophers of science. That such arguments could also be relevant to reliability, survival analysis, filtering, and forecasting seems to not have been recognized.
Section snippets
Reassessment and the principle of conditionalization
Some individuals when faced with the matter of assessing with as observed data, may chose to re-assess all probabilities treating the factual event as a part of background history; that is, they would start from ground zero, even if the observed is not a surprise. Diaconis and Zabell (1982) label a process like this, complete reassessment; however, the driving premise considered by the above authors is different from the one we are discussing here, in the sense that
Conclusion
The innocuously simple problem of assessing conditional probabilities can get riddled with issues, both philosophical and technical, when the conditioning event becomes a reality. The cleanest way to approach it is through Bayes’ Law. When this is done it can be seen that the standard answer arises as a special case under the assumption of a constant likelihood. Other forms of the likelihood will lead to other answers. Since the choice of a likelihood is an assessors prerogative—just like the
Acknowledgements
The comments of a referee are acknowledged. The referee drew attention to Joyce (1999) and made us aware that matters discussed here have been considered by philosophers of science. Supported by Grants N00014-06-1-037 by Office of Naval Research and W911NF-05-1-2009 by the US Army Research Office.
References (18)
Statistical information and likelihood
Sankhya Ser. A
(1975)- Basu, D., 1982. A note on likelihood. In: Symposio Nacional de Probabilidade e Estatisticia, Universidade De Sao Paolo,...
La Prevision: ses Lois Logiques, ses Sources Subjectives
Ann. Inst. H. Poincare (Paris)
(1937)- de Finetti, B., 1972. Probability, Induction and Statistics. Wiley, New...
- de Finetti, B., 1975. Theory of Probability, vol. 2. Wiley, New...
- et al.
Updating subjective probabilities
J. Amer. Statist. Assoc.
(1982) Likelihood
(1992)On an absolute criterion for fitting frequency curves
Mess. Math.
(1912)- et al.
Bayes method for bookies
Ann. Math. Statist.
(1969)
Cited by (9)
From Least Squares to Signal Processing and Particle Filtering
2018, TechnometricsFiltering and tracking survival propensity (reconsidering the foundations of reliability)
2016, Statistical ScienceBayesian methodology for uncertainty quantification in complex engineering systems
2015, Numerical Methods for Reliability and Safety Assessment: Multiscale and Multiphysics SystemsLikelihood-based approach for uncertainty quantification in multi-physics systems
2015, Numerical Methods for Reliability and Safety Assessment: Multiscale and Multiphysics SystemsAdaptive Percolation Using Subjective Likelihoods
2014, Econometric Reviews