The ideas of an 18th-century English clergyman have been used by two management scientists to forecast the success of new movies with unprecedented accuracy.

Two American experts have turned to the predictive power of Bayes's Theorem - a powerful analytical technique developed by the Presbyterian minister Thomas Bayes - to help Hollywood work out a film's potential popularity.

The mathematical model created by Ramya Neelamegham, of the University of Colorado, Boulder, and Pradeep Chintagunta, of the University of Chicago, is proving to be far superior to the methods major movie studios use to predict box-office returns.

Unlike classical statistics, which are based purely on hard facts, Bayes starts with what amounts to an educated guess as to the likely outcome and then, as more facts come in, alters the projection to give an increasingly accurate picture of the future.

ÁñÁ«ÊÓƵ

Harnessing it for the movie industry would allow a studio to study initial box-office figures and then work out how the film is likely to perform in subsequent releases abroad, revising those predictions every time new data became available.

The model was developed from a study of receipts for 35 movies between 1994 and 1996 in 13 countries, including the United Kingdom and United States.

ÁñÁ«ÊÓƵ

While it will not prevent the creation and flop of such turkeys as The Avengers, the research, published in the latest issue of the journal Marketing Science, will allow movie makers to plan marketing strategies, negotiate contracts - even develop damage-limitation strategies - with greater confidence than at present.

"By computing information available to a studio, such as star quality, past performance of similar films and the number of screens that will debut a particular film, you can make a reasonably accurate forecast of how a film will perform," said Dr Neelamegham.

"Our modelling system aims to predict viewership after the movie is produced so it will not reduce any element of creativity in the movie production process."