The title I submitted contains information: "Human fertility rates have fallen nearly 50% since 1950".
The weak original title ("'Remarkable' decline in fertility rates") contains no information.
When people extrapolate figures the way you did when rewriting a title, they often (usually?) get it wrong and make a claim that isn't supported by the article, or emphasize something that the article doesn't emphasize. On HN we try to let the content speak for itself. If you want to say what you think is important about an article, that's great, but you should do it in the comments, on a level playing field with other users.
Other users pointed out a legit reason not to use the article title: "fertility rate" is misleading. In that case we look for an alternative in the article itself. The subtitle was usable; that's common. If not, you can look for a good phrase in the opening paragraph, or possibly a photo caption. Occasionally it's necessary to dig something out of the middle of the text. But there's almost always an accurate and neutral phrase somewhere in the content that can serve as a title. When we change a title on HN, that's our search path for finding a different one. We avoid making up our own language as much as possible.
This is the relevant part of the article:
> In 1950, women were having an average of 4.7 children in their lifetime. The fertility rate all but halved to 2.4 children per woman by last year.
Note that the response you got was that others aren't reliable when extracting statistics from articles, so while you may be very reliable in this way, they can't trust users to do this and have had to draw a line.
I too hate clickbait (unless it's a headline I wrote, of course) but "remarkable" doesn't trigger me.
By some metrics, it's more clickbait than the original title.
The original title told how the rate has changed since 1950 (minus 50%), which is a piece of information.