We have seen in the transport RTT NQE experiment that when using transport RTT, much fewer sessions are categorized as Slow2G, 2G. The very likely reason for this is that the ETC thresholds for HTTP RTT do not match the ECT thresholds when using transport RTT.

Now that there is some data on transport RTTs from stable channel users, the ECT thresholds for transport RTT should be updated so that approximately similar number of sessions are categorized as slow for both transport RTT and HTTP RTT.