Before its theatrical release, audience tracking surveys estimated the superhero action-comedy could gross as much as $25 million its opening weekend.

Instead, the sequel took in only $13 million, finishing far behind the civil rights drama "Lee Daniels' The Butler" and earning "Kick-Ass 2" an instant reputation as a flop.

For decades, tracking was used by studios to determine filmgoer interest ahead of a new movie's release and tell marketing executives where to spend their ad dollars.

But now trade publications, national dailies, blogs, TV newscasts and even drive-time radio shows share the once closely held numbers with everyday moviegoers. Tracking establishes financial expectations for a new film as well as an A-list star's ability to "open" a movie. The estimates effectively declare a winner before the weekly box-office battle begins.

But at a time when tracking's influence on a film's box-office fate has never been greater, chronic inaccuracies have led industry observers and some studio chiefs to conclude that tracking may no longer be a dependable box-office barometer. With a cluster of Oscar-worthy films soon heading into theaters, the pre-release surveys are increasingly coming under attack.

"Tracking is broken. There's no doubt about it," said Vincent Bruzzese, chief executive of the tracking firm Worldwide Motion Picture Group. "It's been asking the same questions since 1980. It isn't predictive anymore. And it doesn't cover the way consumers make choices anymore."

This summer, several movies were damaged by inaccurate tracking. "The Lone Ranger," "The Wolverine" and "The Hangover Part III" were said to have "underperformed" when they had openings at least $10 million below estimates. All went on to sputter domestically after bad word of mouth and speedier-than-anticipated exits from the multiplex.

Even a hit film can fall victim to bad tracking. "Cloudy With a Chance of Meatballs 2" opened as the No. 1 movie at the box office but is seen as having underperformed by grossing $10 million less than estimates predicted.

When movies exceed expectations, they generate positive buzz that can increase returns. Last weekend, "Gravity" took in $55 million — $10 million more than the most optimistic pre-release surveys indicated it would earn. Summer's superhero film "Man of Steel," horror flick "The Conjuring" and the thriller "Now You See Me" earned many millions more than the tracking predicted.

"You can say, 'The testing was great,'" said one respected studio marketer who, like other top executives interviewed for this story, declined to be identified for fear of jeopardizing his industry standing. "But you know in your heart you don't believe in the testing anymore. And if you do, you're fooling yourself."

Because of the sheer volume of movies being released — 660 last year — as well as seismic social media changes, tracking service executives say, pre-release audience awareness and anticipation have never been more difficult to gauge. This is especially true, experts say, for non-sequel films and films popular with minority moviegoers, who can be harder to survey because they are a statistically small and not reliably representative cross-sampling of respondents.

Yet Penn defends Reel Pulse's tracking surveys, pointing out its estimates are within 15% of the actual gross 75% of the time. Other major tracking services are OTX, MarketCast and National Research Group.

Even with tracking's accuracy increasingly doubted, it's such a dominant part of the Hollywood conversation that none of its studio detractors interviewed for this article voiced willingness to give up the service.

Studios receive tracking information over a three-week to two-month pre-release window. The estimates sample audience awareness, "definite interest" in seeing a movie and the proportion of respondents ranking the movie as their first choice, as well as projected breakdowns by gender and age.

Firms crunch their polling results, comparing the movies with previous titles by genre and release window — "comps," in tracking terms — to yield an estimated opening-weekend gross.