Supreme Court: When Witnesses Agree Too Perfectly, That’s the Problem

Case: Nirmala Bai Devidas Chavhan v. State of Maharashtra and Others
Court: Supreme Court of India | Judges: Justices Sanjay Karol and Augustine George Masih
Decided: August 3, 2026 | Citation: 2026 INSC 787

Sometimes, a witness account that sounds too perfect is exactly what gets a case thrown out. That’s what happened when the Supreme Court upheld the acquittal of 20 men convicted of murder — because their accusers’ stories matched each other almost word for word.

On Holi in 2014, in a Maharashtra village, a young man named Avinash got into an argument with some people playing loud music near his sick grandmother’s house. Hours later, when he returned home with his father, brother, and cousin, a mob of around 23 people allegedly dragged all four of them out and attacked them with iron pipes, bars, and wooden planks. Avinash died on the spot; the other three were injured. A trial court convicted 20 people and sentenced them to life imprisonment.

The High Court later overturned the conviction and acquitted everyone. The victim’s mother and the state government appealed to the Supreme Court to restore the convictions — and lost.

Why “Too Perfect” Testimony Backfired
Six witnesses — including the injured victims themselves — gave accounts that matched each other in astonishing detail. Each one correctly named exactly which of the 23 attackers held which victim, and which weapon each person carried. That might sound like rock-solid proof. But courts see it differently: a chaotic, terrifying attack lasting just a couple of minutes should produce messy, slightly different memories from different people — not a word-for-word match, especially when the accounts were given days later. That level of precision is a classic sign of witnesses being coached, not genuinely remembering.

The Medical Evidence Didn’t Add Up Either
This is where the case really falls apart. Witnesses claimed four men repeatedly struck the victim’s head with iron pipes for two to three minutes straight. But the autopsy found just one single wound. The doctor who examined the body admitted, under questioning, that a single injury would only be possible if all four men — using different weapons — somehow struck the exact same spot every single time. And if that had actually happened, he said, the skull likely would have been crushed. It wasn’t. Similarly, witnesses claimed one victim’s eyeball popped out and his teeth were knocked out — but hospital records showed his eye movement was completely normal, and no such injuries were found at all.

And, A Few More Cracks in the Case
• No neutral witnesses. Even though 40-50 people reportedly gathered at the scene, every single witness who testified was related to the victim’s family.
• The accused had unexplained injuries too. Some of them had gone for medical treatment right after the incident — raising the possibility this wasn’t a one-sided ambush at all.
• Evidence handling problems. The forensic samples arrived at the lab with broken seals, so those reports couldn’t even be used.

The lesson courts keep coming back to: an eyewitness account that lines up too neatly — especially days after a chaotic, violent event — isn’t automatically more believable. Often, it’s less.

However, A Closing Thought: In our opinion this judgment raises a fair question about the standard being applied to eyewitness testimony: courts distrust accounts that differ, but here, closely matching accounts were also viewed with suspicion. This leaves little room for how victims of a fast-moving group attack might reasonably describe events. It is also worth noting that the trial court, having directly observed the witnesses testify, found their evidence credible enough to convict twenty people — a firsthand assessment that arguably deserved more weight, even though the High Court and Supreme Court were legally entitled to reach a different conclusion.