In the State vs Oscar Pistorius case, we all know what may have happened. The trial can also aspire to a record – the worst forensic evidence in South African legal history, ever. This does not relate to bloodstains and fingerprints, but simple scientific evidence that could have been tested on a yes-or-no basis but which was either never presented or was given in a fumbling way that was embarrassing. In closing, prosecutor Gerrie Nel referred to the experts called on behalf of the defence as “the worst witnesses ever”. He is not far wrong.

Witnesses, the Stipps, testified that they heard gunshots and screams. If the screams came before the gunshots, then Pistorius would have been shooting after a disagreement with his girlfriend, Reeva Steenkamp. Crucial, thus, was if the screams were of a man or a woman and the order of screams and gunshots.

The defence contended that it was Pistorius screaming after firing his gun and any subsequent apparent gunshot sounds were, in fact, Pistorius breaking down the toilet door with a cricket bat to free his girlfriend.

Gunshots? Cricket bat? A circus followed this revelation. In California, Alexander Jason set up, on a shooting range, a toilet door, a gun, a bat, microphones, recorders and a camera, and videoed himself striking the door and shooting it. He concluded that both sounds were similar. In court, defence advocate Barry Roux played recordings of sound tests of cricket bat strikes and gunshots produced by defence expert Roger Dixon, who admitted during cross examination that he did not know if the person doing the sound recordings had any experience in recording gunshots. He also admitted that he was not present when the ballistics expert conducted the gunshot tests.

The ballistics expert, Wollie Wolmarans, said they had to repeat their gunshot tests because the first try had failed as a result of the gun jamming and they could not record the sounds of rapid fire. It then emerged that bat strikes and the gunshots had been recorded on different days and the recordings spliced together. Wolmarans also testified that he had poor hearing and that, in court, gun and bat sounded similar because he was not wearing his hearing aid.

Astonishingly, no gunshot tests were conducted in the bathroom of the accused. Factually, the range tests are rubbish, since the recording and reproduction equipment is not at all designed to reproduce high- impulse noises.

Definitely not one of “the worst witnesses ever” was acoustics expert Ivan Lin. His answers were truthful and unbiased. Nel could hardly pick holes in anything.
Then the prosecutor dropped the ball. Producing South African National Standard 10103, which relates to the measurement and evaluation of noise levels for speech communication and annoyance, he pointed out that the standard states that, if a sound had a tonal component, then 5 dBA must be added to measurements. He asked Lin why he had not added 5 dBA in his calculations. Lin pointed out that the 5 dBA correction refers to an average repetitive sound, and not a single sound like a scream. He struggled manfully to explain to Nel the concept of time average noise levels, as opposed to impulse noise levels.

Another miscue was when Nel inferred that Steenkamp had gone downstairs to eat in the kitchen at about midnight without triggering the alarm. This led to speculation about whether the house alarm was on or not. However, a modern alarm system has stored in memory every trigger from every sensor for at least 30 days, whether or not the alarm is activated. A download from the alarm box would have established exactly if someone was downstairs, and when. But the State did not have this evidence.

There were many other mistakes. The Pistorius trial is a first in many respects: media coverage, high-profile people, legal heavyweights. It would be good if it also set a benchmark for expert witnesses: what qualifications expert witnesses must have to be allowed to testify and what, definitely, would disqualify them.