CARROLL COUNTY, Ind. -- The only piece of hard evidence presented before jurors in the Richard Allen double murder trial in Delphi that supposedly directly ties the defendant to the killings of Abby Williams and Libby German in 2017 is an unspent bullet found at the murder scene. If that’s the best the State has, FOX59/CBS4 legal analyst John Tompkins said the prosecution has a problem. Delphi murders: Forensic firearms examiner discusses toolmark methodology, ‘unspent bullet’ evidence "Well, the first thing you do is you check the protocol that was used to conduct the test and determine if that was an accepted protocol relevant to the scientific community,” said Tompkins, an Indianapolis defense attorney. “That’s the magic language in court.” But Tompkins said the State’s firearms expert may have pulled a magic trick. Former Indiana State Police forensic firearms examiner Melissa Oberg testified all day in front of a jury of seven women and five men in Carroll County Circuit Court. Oberg told jurors the steps she went through to determine that an unspent cartridge found near the bodies of Libby and Abby on Feb. 14, 2017, near the northside bank of Deer Creek below the Monon High Bridge, came from a gun seized during a search warrant at Richard Allen’s house in October of 2022. “Sufficient agreement,” is the industry-accepted term when confirming a match, said Oberg. Earlier this week jurors learned that following the initial search of the crime scene, conducted 24 hours after the girls disappeared while walking across the bridge on a day off from school, late at night an investigator spotted a glittering object in the dirt near where the bodies were discovered, their throats slashed. Detectives recovered an unfired .40 Winchester Smith & Wesson bullet. On Feb. 17, three days after the bodies were found, that bullet and a .22 Glock handgun were delivered to the ISP Forensics Laboratory in Indianapolis for testing. Later that month, two more .40 handguns were also delivered to the lab. It was never revealed to jurors why those guns were referred for testing. Oberg testified that after the bullet was examined for DNA and fingerprint traces, she examined the base of the round which revealed three slight scratches, presumably made when the cartridge was ejected from the gun of the killer. Police have theorized that perhaps the killer, commonly assumed to be Bridge Guy who was captured on Libby’s cell phone video stalking the girls from the bridge and commanding them down the creek bank and to their eventual deaths, may have racked a bullet from the chamber of the gun in an attempt to frighten his victims. For more than five years that bullet languished in investigators’ files, unmatched with any weapon and devoid of DNA or fingerprint or fiber evidence that would lead detectives to the killer. In August of 2022, in an investigation linked to the Delphi case, ISP divers spent several weeks scouring the bottom of the Wabash River in Peru, coming up with knives, cell phones, debris and four handguns that were also tested for a match to the recovered 2017 bullet but to no avail. Then, a month later, a volunteer clerical assistant found a copy of an interview conducted by a DNR officer with Allen just a couple of days after the murders where he admitted to being on the bridge trail that day and seeing some young woman walking the path as he made his way to the High Bridge. That report and its accompanying tip, on which the word “cleared” was written, were lost in a file folder on a desk that was being cleaned out. Once that report and Allen’s admission he was on the trail that day was brought to the attention of Carroll County Chief Deputy Tony Liggett, who was running for sheriff, investigators began building their case against the local man, matching his car to a vehicle seen near the site of the murders. By mid-October, Allen was interviewed, told he was a suspect, advised of his Miranda Rights and sat by as investigators conducted a search warrant at his home where they recovered scores of knives, more than 20 cell phones and devices and a .40 Sig Sauer handgun. Missing was the cell phone Allen showed to the DNR officer during his initial interview in 2017. As he waited in a car while troopers searched his home, Allen allegedly said to an officer, “It doesn’t matter. It's over.” Twice. Oberg testified that once she received Allen’s .40 semi-automatic, she attempted to test it to determine if it was the firearm that ejected the unspent bullet at the site of the killings. The analyst told jurors that six times she ejected unfired bullets from the gun, but the examination did not result in marks on the base of the test bullets that matched the crime scene round. Defense Attorney Bradley Rozzi pushed back at Oberg’s testimony, citing professional doubts about the accuracy of firearms forensics, that the analyst did not submit her findings to an outside reviewer but only to peer examiners at the ISP Lab and that firing the gun would alter the composition of the test bullets metal, rendering the test results meaningless. Tompkins agreed that fired bullets cannot be accurately compared to unspent cartridges and that Oberg apparently rewrote the testing rules when her first experiments with six dry cycled rounds failed to replicate the evidence on the murder scene bullet, so she changed protocols to test for the answer investigators were seeking. "When you have a fired round that has been exposed to heat and pressure, metal becomes soft, easier to mark and put markings on, so when you compare a dry cycled casing to a fired casing, you are not doing a comparison between comparable items. Forensically speaking, they are not the same. The markings wouldn’t have been put on under the same conditions and so comparing those markings is not a valid comparison,” said Tompkins, who is not connected to the case. "And she’s only doing it on this one firearm. So now we have double protocol testing on only the firearm that is going to go to trial. Did she go back to the other weapons that she dry-cycled after the fact? Did she still have those weapons and fire those and compare those to see if they matched too? How much did she structure this final…quote…match…just for this trial? "She’s not doing the same test that she did with the other firearms. So if you just take the six rounds that were dry cycled through the weapon and compare those rounds to the dry-cycled rounds from the other weapons, do those match? So do all .40s leave the same kind of the same thing with dry cycled rounds on them? That’s the kind of comparison that I would want to know.” Rozzi complained that Oberg had not provided the defense team with full reports of her examinations and results. "Did she issue a report after the dry-cycled rounds did not match the recovered rounds?” asked Tompkins. “Did she tell the State and create a report that had to be given to the defense, ‘Hey, your theory does not match the data that I’ve collected,’ and then get a request to do additional testing or did she do this additional testing which is live fire testing on her own independently with no request to develop a new protocol?” Complete Delphi coverage Special Judge Fran Gull turned down a defense request for the testimony of a metallurgist to explain to jurors the significance of the scratches on the so-called “Mystery Bullet” and how there could be an alternative explanation compared to the State’s expert findings. The metallurgist is not a firearms analyst, ruled the judge. "It certainly strips them of the ability of someone other than just the defense on cross-exposing the weakness in the protocol,” said Tompkins. “The metallurgist is not there to testify about firearms ballistics. The metallurgist is there to testify about the properties in the metal in the casing that was tested versus the properties of the metal in the casing that was dry cycled. That is not ballistics. That is metallurgy.” While Oberg concluded her testimony, the defense will have an opportunity to call its own expert witness to counter the State’s claim that its crime scene bullet matches Allen’s gun.