I hate people who just go with the latest trend without any real knowledge of the subject. A statistics class should be mandatory in high school.
FPS is still a very valuable tool. It provides an average frame rate, the same thing as 1/frame latency (the thing every benchmark place includes now). The thing is, it's not the ONLY thing you need to consider. What you also want to look at is a histogram of the frame latency distribution. A 99th or 95th percentile frame latency only shows you a slice of this, and is just as incomplete a picture as FPS. Ultimately, FPS is a mean. A mean is a single value designed around condensing the data gained from a population into a single number. It's not about giving the whole story, it's about providing something quick to provide a general idea of what's going on. To get a really complete picture of the situation, you'll also want to look at how those frame latencies are distributed throughout time, so you can if your latencies at the edges of the distribution are due to one area or it's a persistent problem.
TL;DR: FPS is a valuable tool for providing a concise summary of a lot of data. It doesn't provide the whole story and that's OK.