When AFP Assistant Commissioner Colin Winchester was shot dead outside his Canberra house on the night of January 10, 1989, the instinct was to see it as an audacious hit by Ndrangheta – the Calabrian Mafia-style organisation heavily involved in the cannabis trade.

Attorney-General Michael Tate voiced the suspicion on the night of the murder. So did, in the days after, senior police, other politicians and men and women who had been gathering intelligence for years about organised crime. Like other newspapers around Australia, The Canberra Times of January 11 connected the murder to the Mafia. The immediate dedication of unlimited resources and the rush of offers for assistance from other police forces echoed the suspicion that not only was the killing of a senior police officer a cooly planned professional job but that it represented an open challenge to the system. That challenge would be accepted, and the law vindicated, with Winchester avenged.

From that day until today, I and many other journalists who have reported on the case were inundated by people who speak of strange incidents pointing at particular individuals, particularly corrupt police, or statements said to have been made by people such as prisoners, drug figures, police and judges.

No doubt police were the recipients of a similar number of reports – although a good many coming to journalists were from people who did not know who to trust in the police. Likewise, civilian experts on organised crime were free with theories and recollections of events that could be material. The very plethora of such suspicions – and the fact that, at most, only one could be true – bedevilled investigations into the organised crime link.

But the investigations continued and, as the secret reports prepared by organised crime intelligence operations indicate, those conducting the investigation were convinced of organised crime involvement and the need for redoubled efforts, even as they admitted that there was no "smoking gun" proving who actually did the hit, or how. There were a number of candidates, Italian-Australian and Italian, and some suspicious intelligence against all. There was even more information pointing at particular figures who had been involved in organising the hit. Some alibis did not fit; some movements, not least a return flight to Italy on the morning after the murder, seemed inherently suspicious. The case – such as it was – needed a good deal of resources and work.

It never got it. Progressively the investigation was shut down, and at the insistence of ACT detectives investigating their theory that the murder was done by a former public servant with a history of mental problems, with an animus against Winchester because he had refused to shut down a prosecution.

But it was here that another factor emerged.

The very existence of the AFP had been a messy and unpopular marriage of an original ACT Police Force, a Commonwealth Police Force and elements of the Narcotics Squad in the old Customs Department. There were heavy jealousies and difficulties of integration. Former ACT Police held old Commonwealth Police – whom they called "plastics" – in contempt, believing them to be badly trained and inexperienced in dealing with "real crime". This was, in part, because service on the beat in Canberra or the investigation of ordinary Canberra community crime was the only form of service regarded as similar to the way state police forces operated. Winchester, already a senior officer in the ACT, had prospered in the AFP but shared the general old ACT tribalism and disregard for the "plastics".

The Winchester investigation was strongly affected by this rivalry, and the "national" investigation into an organised crime link was being conducted by people of a non-ACT background, by detectives that many ACT-based police held in open contempt. Similarly, the ACT-focused investigation was being led and carried out by close colleagues and mates of Colin Winchester. Most had a serious emotional involvement in the case, one which seriously affected their detachment.

Despite their self-belief, most ACT detectives had little to boast about by way of sophistication, experience or record. ACT detectives had a poor record of solving murders, of resolving disappearances, or in preparing briefs for prosecution that have convinced judges and juries. It's a problem that was evident years before the death of Winchester and which continues to this day, with a large number of unresolved cases.

Some mismanagement of cases had led to further murders that could not have occurred had police done their job properly in the first place. Two young sisters were incinerated when a car "caught fire" after crashing. But it was the night of the detectives' Christmas party, and the examination of the scene was cursory, despite the intrinsic suspiciousness of the events. In due course the man who claimed to have escaped from the wreck murdered another two members of the victims' family. Then, after a call by this newspaper, the victims were exhumed, and clear bullet holes were found in their skulls. In another case, mismeasurement of a murdered woman thrown into a ditch led police to dismiss multiple reports properly identifying her, on the basis that the victim was too small. When, finally, the mistake was detected, the perpetrator was said to have confessed – he denied it and, oddly, the confession conformed to the early police theory of the way the strangling was done, rather than the ultimate view of the pathologist. Such ambiguities caused the jury to convict the (fairly plainly guilty) man of manslaughter, not murder. The man was released from jail within a year, soon married another woman and, very shortly after, killed her in exactly the same way he had killed his first wife. That could not have happened had police done a better job with the first murder.

In all too many criminal cases, moreover, police prosecutions had depended on alleged (but disputed) confessions and unsigned records of interview. ACT police, like NSW Police used to the power of the "verbal'' and the short cut, resisted campaigns to have interviews with prisoners taped, then videotaped, as well as the installation of closed circuit television in police stations. Until it came anyway, as a result of pressure from lawyers, this police resistance shrouded them from some scrutiny or made misbehaviour more difficult to prove. It did not, however, prevent a healthy suspicion, particularly within the legal profession, of the reliability and honesty of particular detectives, some of whom were involved in the Winchester investigation.

There is no evidence to suggest that there was any conspiracy of the AFP to frame Eastman, or that significant evidence was ever manufactured against him. (There is, however, ample evidence that some police investigators were less than truthful to courts, investigations or to the ultimate inquiry when their own actions were under review.) To the contrary, some evidence emerged early that suggested that David Eastman had an animus against Winchester and that he was an angry and mentally ill man who had made a nuisance of himself on many occasions.

He was not helpful to police. Asked about his whereabouts on the night of the murder, he was vague. He did not volunteer something that police had to discover for themselves – that at some stage soon after Winchester was shot, Eastman had been in a brothel at Fyshwick. Police, prosecutor and judge have always thought his failure to volunteer an alibi immediately was of itself suspicious, though some have wondered how many people, particularly those with an animus against police, would readily volunteer an alibi of being with a prostitute.

It seemed to police that the more they looked at Eastman, the more evidence that showed up suggested that it could have been him. Ric Niness was fairly sure within a week or so. As others found supporting evidence, they tended to become dismissive of what they called "also ran'' theories of the crime, including organised crime theories, especially ones being developed by detectives other than old ACT Police hands, or ones that reflected ill on the club of the old ACT Police – the club to which so many had belonged.

Increasingly, moreover, they regarded the murder of their mate as "our" case, and moved in to close off, often by leading questions, "loose ends" in investigations originally led by others. One way or another, they were going to be the ones leading a prosecution.

Moreover, once evidence fitted their theories, some detectives seemed reluctant to go further, lest it complicate the case. I have known journalists to do something similar, afraid to ask obvious questions or to speak to the other side, less material emerge that makes a preconceived view untenable. Thus leads suggesting that the Queanbeyan gun dealer's dealings were far more extensive than the first indications aroused only cursory examination, and quick dismissal, though more careful work would have found some confirmation and a line of inquiry that should have been followed up, if only to resolve hanging questions.

Likewise, refusals to identify Eastman as the gun buyer were treated as evidence of an anti-police attitude, which, in the end, could prove the negative – that Eastman was the buyer. The manner by which some critical evidence – able conveniently to resolve some gaps – emerged, through the medium of family and police social networks, was also suspicious to some observers. That a strong circumstantial case was gathered was clear – but no one, until recently, ever did much to examine the strength of some of the links in the chain.

Investigators received advice from a psychiatrist that Eastman was a paranoid loner who was dangerous. He might, it was thought, crack under pressure and do something actually proving, as opposed to pointing to, his guilt. The senior investigator thought that he might be the solid rock against which Eastman might dash himself and then dissolve. A campaign of surveillance became in-your-face surveillance, and plain harassment and police impropriety.

Detectives – even two who subsequently ended up in the top ranks of police forces – made harassing telephone calls to Eastman at home at night. When, inconveniently, Telecom was able to prove these calls came from the police station, the inquiry leaders were thought by the inquiry to have concocted a false explanation. Bland denials by police that Eastman was being harassed seriously hurt their public credibility, because the public could see it happening on television. Eastman alienated almost everyone, including his friends, but that did not mean that the public assumed he was guilty. It confirmed only that he was difficult – a test for whether the justice system could provide a fair trial.

The harassment reached into trying to poison the well of Eastman's defence – aggravating problems that Eastman always had in finding advisers who would agree with his strange case priorities. (Eastman, in particular, always wanted to focus on the harassment.) Those who accepted his briefs would be warned by police that they could be in danger from him.

Meanwhile, police were paying all too little attention to the safety of what they thought to be the central forensic evidence thought to prove Eastman's guilt – by showing that gunshot residue found in the boot of Eastman's car was the same as residue found at the murder scene. The work was being performed by a member of the Victorian Police forensic science unit, but with almost no supervision, and in a manner well outside the recognised protocols – about documentation, reasoning, reproducibility and so on – already developed in the wake of forensic scandals and miscarriages of justice such as the Splatt and Chamberlain cases. The scientist exaggerated his qualifications and had a well-developed tendency to adopt a case, becoming an advocate of one side or another.

He was also a very difficult character (like Eastman), angry if his confident conclusions were questioned, deeply resistant to any form of peer review, or review by independent outsiders, and very slow to commit anything to paper. We now know he could not have performed some tests he said he did; he may have done others but there are no contemporaneous records. In some cases he seems to have mixed up exhibits from different scenes. Nor was the jury told that he had recently been dismissed from his job for professional and scientific misconduct. All of his evidence must now be thought discredited. Rejection of his theories as not being based on "facts" is not some mere technicality but the underlining of a fundamental flaw in the argument that Eastman was guilty.

Because he was a difficult character, police were "managing'' this witness but not supervising him. He would fly into tempers and threaten to walk away from the case. He would damn the prosecutors to police, and play AFP internal politics with others.

So busy were police in trying to keep him onside that they failed in their job of determining whether he was a reliable witness, or whether his conclusions were right. The prosecution was to claim that his conclusions had been checked by and approved by leading international experts, including from British, Israeli and FBI laboratories. In fact, these were asked to assume the expert's claims of the work he had done, and to assume his findings, and asked whether his conclusions were right.

Concentrated attention was amassing an apparently circumstantial case against Eastman, but a good many of its components were less than impressive.

There's a phrase for the pattern. It's called tunnel vision. It's the tendency to look for material that supports one's case and not to look, sceptically, for evidence that might disprove it. The tendency to dismiss, or not even look, at material that does not help. To regard sceptics as the enemy; to embrace witnesses whose evidence supports one's theories as friends of the prosecution. The tendency to filter "good evidence that helps build the case, and to reject, ignore or forget material that does not, is a human failing, to which anyone can fall prey, even without bad motives.

But it is particularly common when people are emotionally involved, too close to the case, prone to lack detachment, objectivity and the capacity to see things from a distance. It comes, as one author put it, "from police and prosecutorial cultures in which challenging prior investigative work is discouraged, procedures are not in place to probe already gathered evidence and little concern is given to learning from past errors''.

When, of course, past errors are not even acknowledged because investigators "know" what the outcome ought to have been even before they have found the material to prove it, prosecutions run awry.

Tunnel vision has siblings. One is denial, much aggravated when there are traditions of group think, a lack of introspection, conscious anti-intellectualism, and a want of systems of reviewing and learning from mistakes. The old ACT police culture – with its in-groups and out-groups, and focus on loyalty to the group at all costs – is another problem, not yet eradicated, even as there are no longer any members of the old service. ACT detectives still have no obvious tradition of critical case review, or external review, other than occasionally on its own terms. Failures are instead attributed to oddity in the law, flaws in the judiciary, or the cupidity of juries. The Martin inquiry demonstrated that the AFP adheres today to everything it thought and did at the time.

Colin Winchester was a decent and straight, if human, cop but he was himself also a product of some of the weaknesses in the AFP system based on country town policing. From a force in the process of growing up, he deserved a better and more objective investigation. No one could be sure that the investigation found the right man for the job. And, just as most of the weaknesses of the investigation were being discussed at the time of the conviction, there is no simple truth in the hint, sometimes given, that the police had other evidence they were never able to put forward. Indeed they got in much more than would be found admissible in the modern day.

To this day, no one can say that alternative hypotheses of guilt were eliminated by investigation. At least one, indeed, still exists, according to recent evidence given to acting Justice Martin. This evidence may (or may not) have come from an informant who spoke to Victoria Police some time ago. His information was passed on to the AFP, which ignored it because it did not conform to what they "knew" of the Winchester case. Only belatedly was it passed on to the inquiry as possibly relevant. The judge has given no detail but says that what he was told could have founded an alternative hypothesis consistent with Eastman's innocence.

Whether anything is to be salvaged, 25 years on, is doubtful. Only a serious objective review, conducted by outsiders, could decide. But few who saw the system fail could be sure that all of the mistakes would not be repeated by a completely new generation of ACT police officers. Though it is perhaps significant that none of this new generation was invited into the strategy teams defending, to the last, the investigation fiascos.