In follow-up interviews, the interviewer may dig into those results, and in reference checks, may try to confirm your achievements, so I do not think it's accurate to say it would be OK to use BS numbers.
But the overall sense that numbers are important, and send an important signal about a developer's work habits, patterns and practices, is true.
My point was more towards those engineers who are sticklers for accuracy and truth. It doesn't matter in this case. Mass delusion is fine as long as enough people believe it and there's a thin veneer of logic around it.
A lot of places only give yes this person worked here between those two dates.
The person giving the reference would be in no position to answer in many cases.
Legally a company doesn't want to get sued because they told another company the person didn't hit 27% when they might be able to prove they did.
In fact it doesn't even matter if the number was utter BS and unscientific in measurement. So are probably 80% of metrics in companies that are metrics driven. But the manager running the team still needs a metric to meet their goals and this developer shows that they can deliver that.