We live in a country where rape conviction rates remain stubbornly low. In 2023, the conviction rate for rapes was recorded at about 22% – one of the lowest amongst all violent crimes in India. And that statistic tells only half the story. Many cases never reach a courtroom. Many never even become FIRs. Many are withdrawn under pressure, buried under stigma, or abandoned under the sheer weight of the process. Yet every time a rape conviction does emerge from this maze, our first instinct often seems to be to doubt it.The criticism surrounding the recent Tarun Tejpal judgment got me thinking. Not about the idea of the “perfect rape victim” which feminist scholars have critiqued for decades, but about something else entirely: What is a perfect rape conviction? Because if public reactions are anything to go by, a perfect rape conviction appears to require an impossible level of proof. The victim must have reported immediately. Her memory must be flawless. Her conduct before and after the assault must be textbook. There must be witnesses. CCTV footage. Medical evidence. Digital trails. Corroboration from every conceivable source. We look for evidence that even the law of evidence does not demand in such cases. In short, the conviction should be so airtight that one starts wondering: if there were this many witnesses, cameras and records, how did the assault happen in the first place?Take the Tejpal case. For some critics, perhaps the conviction would have been more acceptable if there had been a CCTV camera not just outside but inside the elevator recording the incident in real time. But pause for a moment. Why would a perpetrator commit sexual assault in front of a CCTV camera? The very absence of such evidence is often a feature of the crime, not a flaw in the prosecution.The breadcrumbs that don’t existRape is unlike most crimes we instinctively compare it to. A theft leaves missing property. A murder leaves a body. A financial fraud leaves transactions. Sexual assault frequently leaves none of these. There may be no eyewitnesses because the perpetrator ensured privacy. There may be no CCTV because the location was chosen precisely to avoid surveillance. There may be no medical evidence because not all sexual assaults cause visible injuries. There may be delayed reporting because trauma does not operate according to procedural timelines. Sometimes there is only one piece of evidence available: the testimony of the survivor. And that reality makes us deeply uncomfortable.When evidence ignores trauma We like evidence that is mechanical. Fingerprints. Videos. DNA. We trust machines because machines do not cry, hesitate, forget or contradict themselves. Humans do. Particularly humans who have experienced trauma. One of the least understood aspects of sexual violence is the way trauma affects memory. Popular imagination assumes that a truly devastating event should be remembered with photographic precision. In reality, traumatic memories are often fragmented. Some details become painfully vivid while others blur. Sequences get confused. Timelines shift. Certain memories surface much later. Yet courts and public commentators alike frequently treat inconsistencies as evidence of fabrication rather than evidence of trauma.Also read: Lessons in Male Rage from a 13-Year Trial and BeyondThe survivor is expected not merely to tell the truth, but to tell it in a way that resembles a well-written screenplay. The irony is striking. When a witness remembers every detail perfectly, sceptics ask whether the account has been rehearsed. When the witness remembers imperfectly, sceptics ask why the story keeps changing. The standard becomes impossible.The law recognises this. Why don’t we?This is precisely why Indian law has long recognised that the testimony of a rape survivor can be sufficient for conviction if found credible and trustworthy. In State of Punjab v. Gurmit Singh, the Supreme Court cautioned against viewing a rape survivor’s testimony with unnecessary suspicion. Similarly, in Bharwada Bhoginbhai Hirjibhai v. State of Gujarat, the Court noted that insisting on independent corroboration in every case would often make convictions impossible. More recently, in Aparna Bhat v. State of Madhya Pradesh, the Supreme Court reminded courts that harmful stereotypes about how survivors should behave have no place in judicial reasoning. The principle makes sense. If corroboration were mandatory in every case, many sexual offences would become practically impossible to prosecute.But this solution raises a different question. What exactly makes testimony “credible”? And who gets to decide? One judge may see consistency. Another may see contradiction. One may interpret post-assault communication as understandable human behaviour. Another may view it as creating doubt. One may understand trauma. Another may expect coherence. The same testimony can appear convincing to one mind and insufficient to another. That is not a criticism of judges. It is a recognition of the profound difficulty at the heart of these cases.We often talk about the perfect victim. The victim who reacts exactly as society expects her to. Perhaps we should also talk about the perfect conviction. The conviction that satisfies our almost cinematic demand for certainty. The problem is that such convictions may exist only in fiction. Real-world sexual violence is messy. It takes place behind closed doors, in homes, in hotel rooms, offices, elevators, and cars. It produces incomplete evidence. It leaves survivors with memories that may not fit neatly into legal narratives. If we insist that every rape conviction look like a crime thriller, we risk creating a standard that the very nature of the offence can never meet. The uncomfortable truth is that justice in sexual violence cases will often require us to make decisions amid uncertainty not because the law is failing but because that is the nature of the crime itself.Tanima Kishore is a human rights lawyer and an Advocate-on-Record in the Supreme Court of India.