The Almost-Perfect Bank Heist

Jul 3, 2012 - by Liz Porter

July 3, 2012

An edited extract from Cold Case Files: Past crimes solved by new forensic science

An edited extract from Cold Case Files: Past crimes solved by new forensic scienceavailable for Kindle in the United States on Hard copies available at

by Liz Porter

Mark Chrystie relished the challenge of hunting criminals with some ability. But as an armed robbery squad detective in the Australian state of Victoria, his day-to-day work involved the investigation of jobs carried out by lummoxes with no idea of planning, beyond the purchase of balaclavas and overalls. They would then barge through banks’ front doors waving shotguns and escape in stolen cars.

So he enjoyed tracking down the crew that had planned its arrival at a bank in Melbourne’s posh Toorak to the second.  Arriving one minute after a cash van delivered $250,000, they walked out of the bank with the cash in cardboard boxes on their shoulders, strolling down the street with the insouciance of men carrying a wealthy customer’s bulk order of groceries to her Mercedes.

The detective also happily matched wits with the men who held up a cash van by hiding in a car boot equipped with a spy hole in its panel work, and then leaping out to surprise a driver who thought he was pulling up in an empty car park.


The Bank of Melbourne Heist – A “Soft” Target

The 1996 daylight bank robbery of a suburban branch of the Bank of Melbourne was one of the best-planned jobs that Chrystie had ever investigated. The robbers’ haul was not huge – $137,000 in cash and a pile of blank bank and travellers’ checks that would be difficult to pass.  But the men had done their armed robbery homework thoroughly, researching every aspect of the bank’s operations in advance. If they had given the same attention to their science homework, they might never have been caught – and this story would never have been told. As it was, they were not unmasked for another eight years. When they were, it was cutting-edge forensic science that brought them undone.

The robbers’ choice of venue had been no accident. The Bank of Melbourne was clearly a “soft” target. Eleven of the 28 bank holdups in the state of Victoria between July 1995 and February 1996 had been at its outlets, with four of its suburban branches hit in January 1996.  This branch, in the inner northern Melbourne suburb of Glenroy, would have looked particularly inviting, with security  shortcomings so obvious that the management had recognized them and had scheduled a move to a more secure building for a  day  two weeks after the break-in.

Located in a building with a back door and rear courtyard, the premises offered the robbers handy access to a rear lane, while the next door property’s rear yard allowed them a subtle side-entry. Climbing up its fence and on to the roof of the bank’s outdoor toilet, they could then jump down into the bank’s rear courtyard before entering via the bank’s back door.

The establishment’s Saturday-morning opening hours were an ideal time for a robbery, with the nearby station and surrounding streets free of the weekday commuter rush, meaning less risk of passers-by overhearing the noise of the break-in. The chosen day was the second Saturday of the football finals, and the suburban streets of football-mad Melbourne were even emptier than usual. Real estate agents tended to cut back on auctions and open inspections during this period, while police worked extra afternoon shifts at the football, directing traffic and controlling crowds.

The thieves made their entry more efficient by weakening the bank’s rear door before the theft. Looking over the back fence, they had watched staff opening it when they visited the courtyard toilet, and had gauged its thickness and relative fragility. Visiting the site the night before the break-in, they used some kind of hand-held power saw to cut a square hole in the outer section of the door’s lower half. They drilled only halfway through, making an incision that was invisible from the inside but inflicted serious structural damage.

They had cased their target well enough to know that its employees waited outside each morning while one staff member opened the building from the front, walking in to check for signs of intrusion before giving the all clear and admitting his colleagues. That employee could not be given any reason for suspicion, so the preparatory work on the back door had to be undetectable from the inside. But when they rammed the door with several large steel fencing pickets bound together, the pre-cut outline would function like a perforated “open here” line on a cardboard box, a neat square section of the door would collapse inwards, and they would make their way in.

The robbers knew that the bank’s safe was opened only a few minutes before the start of business, and that a series of piercing beeps, audible outside the bank, signaled the end of the standard five-minute time delay between a bank employee inserting the key into the safe’s lock and the actual opening of the safe. Then the tellers would be given their personal money tins and drawers, and the time would be right for the criminals to enter, order the tellers to lie on the floor, and take their money.


Two Men in Masks

Aware that their arrival would trigger some kind of alarm and activate a security video camera, the two men wore overalls, gloves and balaclavas. Each man also wore a black theatrical Janus mask over his face, thereby depriving cameras of clues about his eye color or teeth. One was a tragedy mask, its mouth down-turned; the other a smiling comedy one.

The duo wanted to be in and out before opening time. And they were.

Bank employee Sam Ferraro was in the bank’s kitchen, laying out the cash that was to go into the teller drawers, when he heard a huge crash behind him.  A square section of the back door had collapsed.  First he saw the star pickets that had done the damage. Then an arm, a leg, and then a man’s head appeared in the hole in the door. Within seconds two men, wearing full length overalls, balaclavas, gloves and masks were standing in the bank’s kitchen.

“Get on the fucking floor!” they shouted, then “Don’t move. You’re not going to get hurt.”

One man had a black handgun with a small nozzle, which he continued to point at the three employees, herding them down on to the floor under the kitchen table.

Meanwhile the second robber was emptying the cash boxes in the safe, stuffing bank notes into a small back pack he had brought with him – and keeping an eye on the staff. He noticed that one of the women lying on the floor was studying him, her training for armed robberies having taught her to not resist but to memorize as much detail as possible about robbers.

“Don’t look,” he shouted at her – an instruction repeated, with emphasis, by the man with the gun.

After emptying the safe, the second man rifled the tellers’ drawers, roughly pulling out notes from the note clips, oblivious to the fact that the clips  set off a silent security alarm – and the bank’s cameras –if they are not handled appropriately,

But the intruders seemed to be expecting to be filmed, and did their best to thwart the security cameras’ eye by ducking down. As a result the cameras recorded only the tops of their balaclavas, so police could not draw conclusions about their height and build.

The speed of their operation indicated their expectation of an alarm. They didn’t know for certain until they were on the way out and spotted the tiny orange light on a black box mounted at shoulder height on the kitchen wall. It was flashing.

The first man rushed at the back door, while the gunman stayed an extra 20 seconds, shouting  a final “Don’t move” at the terrified bank staff, before he grabbed  the big vinyl security guard mail bag, that was delivered to the bank daily. He passed it through the hole in the door to his partner and followed it through. The whole operation, from the moment the back door caved in, until the men fled over the fence, had taken little more than two minutes.


Looking for Clues

The local police arrived shortly after nine. Detective Mark Chrystie had been only 10 kilometres away and was there by 9:30.

The bandits’ masks and disguises meant that the three tellers could provide only basic details about their physical appearance. Both were of “scrawny” build and medium height, with teller Sam Ferraro estimating them as about his height of “five feet six.”   Their accents were “Australian” not foreign and the gunman’s overalls were green German army issue, with a small yellow, red and black German flag on one sleeve. His colleague wore plain overalls with an elasticized waist and pockets, one of which held his mobile – one of the small “brick”-like variety in use back in 1996. Both men wore black balaclavas and ski gloves with red stitching, although the man in the safe took off his left black glove, in order to open cash tins, revealing a white surgical glove underneath. 

The police were soon able to check the tellers’ recollections against the actual clothing worn by the robbers. A police dog squad handler had been called to the scene, in the hope that his dog, “Boss,” might be able to track the robbers’ path away from the bank.

It seemed that the duo had run towards the railway station.  But instead of running on to a platform, they had crossed the train-line and headed into a street on the other side of the line. “Boss” lost their trail in the station car park – a fact suggesting that the men were picked up by an accomplice in a car.  There were too many pedestrians about, and too many scents for him to keep trying. Instead the dog handler set his animal on a property search. “Boss” immediately headed for a dumpster in the lane near the rear. A balaclava and a pair of overalls were inside.

The police also discovered a second mask and a baseball cap in the backyard of the property next door to the bank, and another two gloves, one inside the other, on its concrete footpath. Shoe impressions in the dirt at the base of the wall inside the neighboring backyard told the investigators that the robbers had climbed that wall, stepped down onto the corrugated-iron roof of the bank’s toilet and jumped from there into the bank’s courtyard. Inside the bank itself, the tip of a rubber glove was found – and it appeared to have come from one of the dumpster gloves.


Interpreting the Clues

With three detectives assigned to help him cover the case, Mark Chrystie began the investigation in a mood of cautious optimism. To his experienced eye, the style of the break-in was a clue in itself. The robbers had been patient and calm. They had sat outside waiting for the time-delay to sound rather than rushing in at the earliest opportunity. This sangfroid told him they were professionals, meaning that a survey of Melbourne’s experienced criminals, just seeing what they were up to at the time, might help to identify them.

Another reason for hope was that, for all the careful planning that had gone into the heist, the robbers had still left the detectives plenty of leads. An obvious starting point was the mobile phone spotted in the overall pocket of the man who had emptied the safe. One bank teller had seen a red light flash on the handset, presumably denoting a call received or a message left. Why would the robber have brought the phone unless he needed it to make a call – perhaps to the driver of a getaway car? The local service providers, Telstra and Optus, should be able to tell police what calls had gone through their towers in the minutes before nine o’clock.

The masks also looked promising. Specialty theatrical items, both lined with cloth, they should be available only from a limited range of shops. The overalls, military issue with serial numbers on them, should be traceable to an individual consignment, sold to a particular army disposals store. The saw used on the back door was another potential lead. It had to be portable, and had been either petrol or battery-operated – probably the latter, as it had been quiet enough not to attract attention. Such tools were rare in the mid-1990s. The bank checks bore serial numbers, enabling them to be traced when they were cashed. The robbers weren’t stupid enough to pass them personally; they would sell them on to a middle man, who would sell them again. The detectives would pounce when these checks were passed, hoping that those caught with them would name their source – bringing the detectives one step closer to the robbers.

But, to Chrystie’s frustration, all enquiries led to dead ends. The detectives looked at all robbers known to have done jobs in which they waited for the safe’s time delay to go off. Nothing. The mobile phone enquiries went nowhere: the detectives could find no call that fitted the bill. The quest for the supplier of theatre-quality Janus masks turned out to be a marathon, involving personal visits or phone calls to hundreds of outlets around Australia. But the source of the masks was never identified. The serial numbers on the overalls revealed that they had been used in training and then discarded and presumably picked up by a disposal store. The detectives spent many hours checking such stores, with no result. The quest for the rare battery-operated saw also failed. The police could not find one capable of the kind of work done on the bank’s door. No potential suspects had been reported flaunting unexplained and sudden wealth, and none of the stolen checks surfaced.


Traces of DNA

However, there had been one gap in the criminals’ meticulous preparation. Their forensic science research had been incomplete. They obviously knew about fingerprints, so wore gloves. The thick ski gloves they had donned over the rubber gloves suggested they also knew about DNA and were aware that any blood left at the scene, perhaps by accidentally cutting themselves as they broke in, might eventually betray them. It was their masks that were the problem.

Technically speaking, they had been correct to discard such highly incriminating costumes: they couldn’t risk being held up in a random traffic stop, so their masks, gloves and overalls had to be disposed of as quickly as possible. But they would have been better off dumping the gear further away. They should certainly have thought of a more imaginative hiding place than a dumpster so close to the bank.

The cloth linings of the masks had captured skin cells and saliva droplets from their wearers. A cutting-edge scientific development, “trace DNA,” meant that scientists could obtain DNA from such microscopic specks of biological material.  Of course the robbers had never heard of trace DNA, and, even if they had been doing university-level science homework on DNA, they still would not have known about it at the time. It was their bad luck to be committing their crime less than 20 kilometers from the police forensic center that employed the world experts in this ground- breaking area of science.

From the mid-1990s, Victoria Police forensic scientists Max Jones and Roland Van Oorschot had been noticing that, when they were extracting DNA profiles from blood, semen or saliva stains on clothing, they were also registering a second source of DNA that did not always belong to the person who had shed the blood, semen or saliva they were working on. Instead, this trace DNA inevitably belonged to the person who had been wearing the clothes bearing the stains. It had been deposited in sloughed-off skin cells that were invisible to the naked eye. The two scientists began documenting their results. Their research, published in 1997 in the international science journal Nature, triggered a worldwide revolution in police forensic work.

Max Jones provided Chrystie with the first lead on the case. Jones tested the cloth lining around the mouth of each mask, and, only six days after the robbery, was able to extract DNA profiles from the minute amounts of biological material present. One distinct male profile was on each mask.

The detectives were now in possession of two separate genetic fingerprints – one for each robber. But whose were they? The profiles were useful only if the detectives had a suspect to compare them with. Unfortunately, they didn’t.

The scientists cross-checked the DNA profiles from the masks against the profiles of dozens of other criminals charged with armed robbery. None matched.  At that time, DNA data bases were still brand new. The U.K. had set one up in 1995, but the FBI didn't get the United States one running until 1998. Victoria was the first Australian state to pass legislation to set up a data base, but didn’t do so until 1997. In 1996 Victoria Police were only taking DNA from people charged with major offenses. So the number of samples available for comparison was limited to those left by – or taken from – known robbers.

But detectives are patient. Early in 1997, Chrystie and his colleagues were watching two men who had begun using the stolen bank checks. On one occasion, they had bought $22,600-worth of fillet steak, claiming they were in a hurry because they had a catering company working at the upcoming Grand Prix motor race weekend. In fact, they sold the meat to a butcher. A few days later, their purchase of $86,000-worth of alcohol from an up-market wine business was interrupted by their arrest.

It turned out that they had bought the checks from a middle man who had got them from a man who seemed to have a similar height and build to one of the robbers. A phone call had been made to the bottle shop on the morning of the alcohol buy, to check that the order was ready. It was traced to the phone of a woman with links to a man with convictions for armed robbery in four states. But none of them could be connected to the Glenroy robbery.


A Cold Case

By mid 1997, the case had ground to a halt. No more of the stolen checks had been passed. Undoubtedly, the bottle shop arrest had prompted their new owners to destroy them. Chrystie and his colleagues had spent more hours on this job than on any of the others on their office whiteboard. Other more urgent cases were now taking their attention.

The Glenroy bank robbers could have been forgiven for believing they’d got away with it.

The following year Mark Chrystie left the armed robbery squad. But he didn’t forget the Glenroy job. In fact he thought of it every time he read a newspaper report about a bank robbery.

In 2002 the detective returned to the squad, by then rebadged as the Armed Offenders Squad. He thought about the case every time he had to deal with an armed robbery in which the perpetrators had simply barged in, vaulted over the counter and demanded money. Nobody, it seemed, was doing “professional” but relatively non-violent heists like the Glenroy one.


Finally, a Match

In 1999 the DNA profiles obtained from the robbers’ masks had been entered into Victoria Police’s new database, where they waited for a match. In May 2004, one finally appeared. It was to a prisoner’s profile only recently placed on the system: its owner was Goran Stamenkovic, 39, whose record included dishonesty, violence and 27 driving offenses.

One night in October 2000, police had spotted him driving along a country highway with his number plate partly obscured. Disobeying a request to pull over, he had sped off, leading police in an hour-long pursuit during which he drove at up to 190 kilometers per hour towards Melbourne, veering into the emergency lane and onto grass verges in order to overtake other cars.

Once in town, he had run red lights and dodged around closing railway- crossing boom gates. After a minor accident, he had abandoned his car at an outer suburban railway station, where he threatened another motorist with a screwdriver, stole his car and drove it to a shopping center. There, he had pounced on a woman standing by her car with a baby in her arms, grabbed her keys and was trying to get into her car when he was tackled by security guards. He was arrested and charged with attempted robbery, theft of a car, assault with a weapon, driving while disqualified and possession of a drug of dependence. In May 2001, a judge had given him a minimum four-year sentence for this “murderous” rampage.

As a prisoner, Stamenkovic had to be DNA tested before his release. In 2004, when he had been close to the end of his minimum sentence, he had been duly tested, and his profile had been submitted to the DNA database.

This man didn’t sound like the mastermind behind the Glenroy job. But Chrystie knew of plenty of cases in which once-disciplined criminals had begun using drugs and, as a result, had abandoned carefully planned professional heists in favor of desperate street-based armed robberies, carried out to fund their next hit.

The detective was thrilled with the opportunity to reopen the case. As the officer in charge of the squad, it was his job to supervise the investigation, not do the legwork himself. But he planned to follow this one very closely.

He was pleasantly surprised to find that the 1996 file had not left the office. It was sitting on a shelf directly above the desk occupied by the investigator he had just assigned to the task of running the reopened case: John Bergin. The first task was to retrieve all of the exhibits from storage so they could be sent to Max Jones at the police forensic center for re-analysis. While the masks had yielded profiles back in 1996, the gloves and the hair samples collected from the balaclavas and a baseball cap had failed to give up any secrets.

Jones ordered a new DNA sample from Stamenkovic: a reference sample that could be used in the laboratory to compare with profiles obtained from the crime scene. He then began retesting the robbers’ clothes. He could obtain DNA profiles only from the gloves, but was able to isolate DNA from three different people. Two full profiles were present, along with one very partial profile insufficient to identify anyone. One of the full profiles on one pair of gloves matched Stamenkovic.


Finding the Second Robber

Two months later, while detectives were tracing associates of Stamenkovic for testing, the DNA database linked a second name to the Glenroy bank job: career criminal Mikael Mann. He had been arrested after committing an aggravated burglary in a flat in the Melbourne bayside suburb of Elwood. Mann’s accomplice had hooded and bound the female flat owner. Mann, who knew the victim, had held a gun to her head and robbed her safe of jewelry. The DNA profile taken from him matched the profile from the other mask and the other pair of gloves.

Mann had been committing dishonesty offenses since he was 17. By 2004, he had more than 63 convictions, including one for burglary and theft in Victoria and one for receiving stolen goods in New South Wales, for which he had been sentenced to five years, with a non-parole period of three. On his release in 2003, Mann had arranged to buy a U.S. exchange student’s $7,500 Rolex watch, but had robbed him of it at gunpoint. Arrested a few months later in Sydney, he had been taken back to Melbourne and charged with armed robbery.

It was while he was on bail and waiting for his trial for this offense that he had taken part in the Elwood burglary, which had originally been planned as an insurance job in which the female victim was supposedly part of the conspiracy. But, partway through, Mann had realized that the victim was not in on it. When he was arrested straight afterwards, he had also realized that the police had been tipped off.

But Chrystie wasn’t as happy with this news as might have been expected. Detectives don’t believe in coincidence, and this result seemed too good to be true. It had been wonderful when, eight years after the crime, the DNA database had produced one offender’s name. But when it came up with the second man’s name a mere two months later, Chrystie couldn’t help but suspect a mix-up in the files.

The problem was, the detective knew Mann. He had arrested him in Elwood only a few months earlier. Perhaps there had been a confusion with case numbers. Perhaps the DNA results should have said that Mann’s profile was a match for material at that burglary scene. Detective John Bergin had also put a lot of effort into tracing possible associates of Stamenkovic – men who might have been his accomplice at Glenroy. Mann’s name had never come up.

But exhaustive rechecking confirmed the result. Mann’s DNA was, indeed, a match for the Glenroy tragedy mask. Stamenkovic and Mann were charged with the 1996 robbery. Both gave “no comment” interviews and pleaded not guilty: a great disappointment to Chrystie, who had been hoping for full confessions, to save the time and expense of a trial.


DNA Evidence on Trial

It was unusual for an armed robbery case to proceed with no identification evidence beyond the DNA. But the prosecution could produce nobody who had seen the accused men’s faces or spotted them unmasked in a getaway car near the bank. There were no witnesses who had bought checks from them or who had seen them clutching wads of bank-fresh cash.

The pair faced trial in September 2007. The detectives were looking forward to it for a particular reason: they wanted to see the two men standing in the dock together. They were used to prosecuting robbers with obvious connections to each other: known associates heard blabbing together on tapped phones and who had done jobs – and time – together. But in this case, while the DNA evidence wasn’t in doubt, the police remained uncomfortable with the fact that, for all their research, they had not been able to find hard evidence of any significant relationship between the two accused.

Mann was four years older and had a more serious criminal record, so the detectives assumed him to be the leader and Stamenkovic the follower. But, beyond that, they knew only that both came from Serbian family backgrounds, they had been to the same school, and both had a connection to Broadmeadows, the neighboring suburb to Glenroy. There was also a prison visit record suggesting that one had visited the other.

Seeing the pair sitting comfortably together in the dock, the detectives felt a little more peace of mind. But they remained worried, as did the trial prosecutor, Darrin Cain, about the jury’s reaction to the DNA evidence. Although there had been plenty of trials in which DNA evidence had been an important feature, it was still rare for it to be the only evidence presented. Nevertheless, Cain was confident that Max Jones could present the DNA evidence in a way that the jury would find comprehensible and convincing.

But he expected the defense to present an independent DNA expert, and use him or her to confuse the jury and sow seeds of doubt in their minds. He braced himself for a clash of experts, in which the defense would raise problems with statistics and contamination. He spent hours preparing possible lines of cross-examination.

Jones was ready, Cain knew, for the defense to argue that the masks had been worn by the defendants and then passed on to others. But it was clear that there was only one sample of genetic material on each mask. The prosecutor also knew that one of the men had a brother who had been convicted of criminal offenses. Might the defense use that fact to blow smoke in the jury’s eyes?

Cain was not permitted to tell the jury that the database had solved the case. He couldn’t let on that the respectable-looking men in the dock had been DNA tested after committing crimes involving theft and violence: juries are not allowed to know if an accused person has prior convictions. Any mention of the wizardry of the DNA database would entail explaining how the defendants’ profiles got there. But it was bad storytelling to have to gloss over this key aspect of the investigation and say only that the police had been acting “on information received” when they had reopened the case in 2004 and zeroed in on Mann and Stamenkovic. It sounded like he was hiding something.

He needn’t have worried. Jones’s presentation to the jury was a triumph of clarity. The key point that he made was that full DNA profiles had been extracted from each mask and set of gloves; in each case matching one of the defendants. This result was 98 million times more likely to happen if the biological material tested had originated from the defendants rather than from random individuals in the Victorian Caucasian population. And the two defense lawyers did not put up their own DNA expert after all.

Still, the defense attorneys did their best to remind the jury of the technical possibility that their clients’ DNA had been transferred onto the masks by other items in the dumpster. They also suggested that both defendants were too wide to squeeze into the hole in the bank door, although the jury knew that 11 years had passed, meaning that the men’s girths might have expanded in the meantime.

The defense lawyers hoped that the jury would be concerned by the fact that the prosecution had been unable to find any other evidence besides DNA to link the men to the robbery – concerned enough to find that the prosecution had not proved its case beyond reasonable doubt. But they were destined for disappointment. The jury found the pair guilty, and the judge sentenced them each to seven years, with a minimum of five.

While Chrystie and Bergin were delighted with the verdict, gratified by the DNA success story and relieved that neither man chose to appeal, the closing of the case left them with several unanswered questions: How had Mann and Stamenkovic met? Was the Glenroy heist the only job they had done together? Had they been working for another criminal? Was there a get-away driver? What had they done with the money? And where had they bought those damn masks?

If the detectives ever run into these two criminals after their release, those are questions they’ll be asking.

Total views: 21942