They offered API that they don't want people to use. Maintaining APIs and uptime costs, but they only wanted their API to be used by their apps that collect data about you.
I wonder how this would play out if instead of serving responses for nonhuman consumption (i.e., an API), it was serving responses for human consumption (i.e, HTML documents). In that scenario, if a subset of requests were of a high-frequency nonhuman nature (i.e., polling and scraping), would a court find it equally abusive or materially different?