A film forecasting fog

Boffo frame surprises pic prognosticators

By IAN MOHR

Was there something in the water?

For some reason, filmgoers last week either changed their minds about what movies they intended to see -- or they deliberately faked out the researchers.

Sony's "Ghost Rider" rode off with $45 million in three days, en route to a Presidents Day weekend tally of $55.1 million. The take was higher than what most tracking services had predicted. One , in fact, had forecast that "Rider's" tank would top out at $40 million.

Further down in the frame's B.O. charts the tracking was even further off.

Universal's "Breach" outperformed expectations by 30%, taking in $12.2 million over four days on fewer than 1,500 screens.

Tracking services also underestimated Disney's "Bridge to Terabithia," the kidlit adaptation that flew to $22.5 million after three days and more than $28.5 after four, and Warner Bros.' "Music and Lyrics."

The weekend wound up being the biggest ever for Presidents Day by a full $10 million. But with the "Breach" numbers so far off, some B.O. watchers wondered yet again about the accuracy of tracking data.

"Tracking puts you in the ballpark, and other times it misses the mark," says one studio vet soberly of the weekend's developments.

Last year, for example, Fox's "The Devil Wears Prada" posted socko numbers despite tracking that suggested it wouldn't catch on as well as it did. The trackers also caught flak when they predicted "MI3" would score bigtime at the B.O. but the pic delivered a disappointing $48.5 million bow.

"There's no science to how a film will perform," says Warner Bros. distribution chief Dan Fellman. "I wish there was. It's a barometer, not something to take to the bank."

Tracking can be especially off the mark for pics and genres that don't fit the mold. Animated fare, for example, typically tracks lower than such pics will ultimately perform, since the pics target such young auds. Romantic comedies can also be tough to call.

And sometimes when a pic performs better than tracking suggests, it's because a studio changes its strategy late in the game, after seeing encouraging data, to campaign more aggressively.

Still, despite such discrepancies, studio pros are sticking to their formulas, reminding that there's no scientific way to predict hits.

"Generally speaking, tracking is a good enough indicator," says Fellman. "It's still right more times than it's wrong."