If you get all the data on fast SSDs in a single chassis, you probably still beat EMR over S3. But then you have a whole dedicated server to manage your 14TB of chess games.

The "EMR over S3" paradigm is based on the assumption that the data isn't read all that frequently, 1-10x a day typically, so you want your cheap S3 storage but once in a while you'll want to crank up the parallelism to run a big report over longer time periods.