Tests done on new wells that boosted the fortunes of oil developers by billions of dollars during the U.S. shale boom are increasingly coming under scrutiny,Bloomberg said Jan. 23.

The one-day performance tests, which regularly spike company shares on their results, don’t provide enough data to predict future potential, according to a study by Drillinginfo, a Houston-based analytics and data firm. Additionally, few rules or standards govern the tests, industry observers say, making for inconsistent findings at best.

The result is that a practice that helped draw significant financing for drillers in an era of $100 a barrel (bbl) oil could become a liability as the price collapse leads investors to take a closer look.

“Now more than ever, it’s a priority for the industry to be more transparent and do a better job at communicating what the longer-term productivity of wells in basins are,” said James Sullivan, a New York-based analyst with Alembic Global Advisors. Skepticism, he said, “is only going to grow.”

Drillinginfo’s results don’t stand alone. Other research, including by Australian mining and petroleum conglomerate BHP Billiton Ltd. (NYSE: BHP, ASX: BHP.AX) have yielded similar results.

Along with the short time frame of the initial testing, developers use a range of procedures that can boost first-day output, according to Allen Gilmer, Drillinginfo’s CEO, and more than a dozen other analysts and company officials interviewed about the tests.

Some producers open flow valves full blast for the tests, an action not generally used in regular production, they said. Others install pumps to create artificial pressure, or measure just the first eight hours of flow, then multiply that by three to represent a full day’s output.

“It’s hyperbole,” Gilmer said. “There is no relationship between those test numbers and what can be economically delivered on a sustained basis.”

Still, the tests continue to be a gauge for investors.

Last month, for instance, shares of Range Resources Corp. (NYSE: RRC) jumped by 14% over three days after the Fort Worth, Texas-based driller reported results from a one-day test that suggested one of its Pennsylvania wells had produced enough natural gas to heat 1,000 U.S. homes for a year.

The well flowed for a consecutive 24 hours and a production-boosting pump wasn’t used, according to Range spokesman Matt Pitzarella. While Pitzarella declined to say how widely the well was opened, he said pressure and other factors were designed to match conditions that will exist once the well is producing on a regular basis.

Releasing well results over a longer period isn’t always possible because pipelines and other infrastructure aren’t set, according to Pitzarella. A decade of Pennsylvania production data shows the company’s wells have been prolific, he said.

Range’s Claysville well, though, is in the Utica oil and gas field, a different and emerging formation that spans much of Appalachia, and a region that’s only beginning to be understood. It will be months before it’s known whether Range’s December production test is an accurate reflection of the well’s potential, according to Bloomberg New Energy Finance.

In the Utica, , drill sites on average are on pace to produce less than half of the 700,000 bbl/d of oil and gas that was originally forecast for a period of about two years, according to a Jan. 12 review of hundreds of wells by Tudor, Pickering, Holt & Co., a Houston- based investment bank.

In the Eagle Ford field in Texas, the shale boom’s most prolific formation, the tie between a well’s 24-hour test and its 12-month performance was found to be statistically insignificant in a Drillinginfo study by senior research analyst Chris Smith.

The 24-hour test is “a real number, but it’s a misleading number,” Drillinginfo’s Gilmer said.

In gas wells, the 24-hour test rates correlate slightly better to one-year performance, but the relationship still isn’t strong by statistical standards, according to Drillinginfo. In the Permian Basin and the Barnett Shale, the initial flow rate was only slightly better at predicting the quality of wells. A far more accurate measure is how much a well produces in its first month, according to the study.

“The press releases make many of these wells seem irresistible, but some just don’t have any life to them,” said Ted Borrego, a lawyer and lecturer at the University of Houston who has advised energy production companies for more than 40 years.

A few producers have stopped touting 24-hour tests, saying that the results are poorly understood and figure too prominently in trading and speculation about company prospects, leading to overly sharp changes in share prices. Both Apache Corp. (NYSE: APA) and Marathon Oil Corp. (NYSE: MRO), for instance, prefer to give investors a 30-day snapshot, a better predictor of long-term productivity, according to Drillinginfo.

Apache takes a “relatively conservative approach” to reporting early well results to investors, John Christmann, the Houston-based company’s CEO, said in an email.

“Reporting practices tend to vary significantly from operator to operator and a number of variables that are critical for comparability and for assessing the potential recovery from a wellbore are frequently not disclosed by industry in conjunction with 24-hour test rates,” Christmann said.

An example of just how sharply a company’s fortunes can be affected by a test result occurred in June as U.S. crude was selling for $104.41/bbl .

Halcon Resources Corp. (NYSE: HK) reported that initial tests showed that a well in Mississippi had produced an “average” of 1,548 boe of oil and gas in 24 hours, an outstanding rate. In the next week, the company’s shares jumped 12 %, fueled by the well result and the announcement of a new partner to help fund drilling.

Four months later, though, it became clear the gusher hadn’t continued at a rate even close to its initial breakneck pace. By July, the rate tumbled to about 325 bbl/d of oil, according to a Sept. 15 analyst note from KLR Group. Halcon’s stock, meanwhile, plunged 81% from its July high point.Oil prices have fallen more than 50% in that time. West Texas Intermediate fell 1.1% to $45.79 at 9:05 a.m. in New York.

Scott Zuehlke, Halcon’s vice president of investor relations, didn’t respond to calls and emails seeking comment.

A key variable when measuring production is how wide open companies run the well during the test. A set of valves, or the “choke,” on a well controls the flow just like the water from a garden hose runs faster or slower when the spigot handle is turned.

Test results released to investors rarely disclose the well’s choke setting, or whether that setting will be changed after the test.

In 2012, Aubrey McClendon, who was then CEO of Chesapeake Energy Corp. (NYSE: CHK), announced a remarkable well result. The Thurman Horn 406H well, in a largely unknown formation in North Texas called the Hogshooter, produced an average of 5,400 bbl/d of crude over eight days, he said on June 1 of that year.

What the company didn’t say was that the well’s choke was open almost to its widest limit, according to documents reported to the Texas Railroad Commission. The diameter of the valve through which oil flowed was more than 3x larger than the one used by peers such as Marathon, the documents show. Less than six months later, average daily crude output at Thurman Horn 406H had dropped 94% to about 343 bbl/d, according to state production data.

The full-blast strategy is a loser in the long term, according to a study last year by producer BHP. The Australian mining company, with access to some of the most prolific drilling land in Texas, found that many wells reporting high IP slowed significantly over time.

A two-year comparison of BHP wells to four competitors in the same field showed that those whose wells produced more in the first month of operation eventually lagged BHP in output after 24 months, according to a June 2 presentation.

“Investors need to be cautious” about using initial production [IP] tests as an indicator of the quality of a well,” Rod Skaufel, president of BHP’s shale business, said in an interview. The tests “tend to get a lot of publicity, but they aren’t necessarily a good predictor of long-term performance.”