Mind the Rounding Error — and Who’s Driving the Ferrari?

The world of DeFi had a bit of a growing-up moment recently. The decentralised finance (DeFi) protocol Balancer suffered a catastrophic exploit that reportedly saw over USD 100 million disappear into the digital ether.

The culprits didn’t brute-force the system or crack the code. They exploited a rounding error — a tiny, almost invisible vulnerability in Balancer’s “Stable Pools.” By skimming micro-gains across thousands of transactions, they siphoned off fractions of cents that quickly added up.

In other words, they stole the rounding error.

When I mentioned this to some friends in traditional finance, a few smirked and said, “See? DeFi’s not ready for serious transactions”. But I’m not so sure that’s fair.

Because this isn’t exactly a new problem. Those of a certain vintage might remember Superman III, where Richard Pryor’s character discovers that the fractions of cents missing from everyone’s pay packets are just floating around inside the company’s computer system. Naturally, he reroutes them to his own account — and would have gotten away with it too, if he hadn’t arrived at work the next morning in a brand-new red Ferrari.

But Pryor’s caper made a sharper point: the company was already skimming those fractions from employees without anyone noticing. Nobody measured it properly, so nobody knew.

And that’s really the heart of it — whether we’re talking about 1980s payroll systems, futuristic DeFi protocols, or today’s FX markets, the small errors matter. If you don’t measure accurately and independently, you might never notice what’s quietly leaking away until it’s too late.

Automation in foreign exchange is inevitable — and, frankly, desirable. Everything we do is becoming more data-driven and algorithmic. To paraphrase the Borg from Star Trek: “Resistance is futile.”

And in DeFi New Change is already providing high-quality data to power smart-contracts as the industry looks for increasingly innovative solutions to deliver efficiency gains and fairness through automation.

But here’s the catch — automation only works if your data is right and not everybody insists on using high-quality and independent data like ours. Garbage in, garbage out… or, worse, someone else’s Ferrari in your car park.

At New Change, we’ve built our entire business around that simple truth: automation depends on independent, accurate, and high-quality data. That’s why we’ve been named Best Market Data Provider two years running at the e-FX Awards, and why Euromoney crowned us World’s Best FX Benchmark Provider and Best Solution for FX Forwards.

Because how can you automate your forward execution without knowing where the forward points really are, not just the special dates, but all of them? Or your spot execution if your “market” data only comes from the same system you’re trading on? Even Richard Pryor’s character could have spotted that opportunity — and you don’t want to be the punchline in that joke.

As foreign exchange becomes increasingly automated, tiny discrepancies add up faster. Small errors can become a big issue. And unless your data is truly independent, someone else might be quietly collecting your fractions of a cent — until they drive past in a Ferrari that you paid for.

So yes — automate boldly. But measure wisely. Because in the world of data, the devil — and the Ferrari — is always in the decimals.