"general rpc servers for solving specific problems using in memory data structures (e.g., who qualifies for a match search given dozens of constraints and millions of users; what your match score is with 10,000 qualifying people, given you've all answered hundreds of different questions each on average) ... Great tech is available now, serving is cheaper, and you probably don't have the computational workload OkCupid does."
I know people who have used OKC before. OKC users in my social class (male, white, educated) ignore the match percentages, because the SNR is really low. They just plow through all the search results of people to find good pictures and interesting profiles.
So, I'd speculate that match-percentages are a marketing thing, and that they know they made a weak business decision which required lots of computation and now they're stuck with it.
I'm probably wrong. Maybe the long-tale users pay attention to match-percentage.