Cover traffic "helps hide who you are talking to". Why is this needed?

In setting up the xx Messenger Android app, I’m being offered to enable Cover Traffic.

It says:

It not only hides *when* you send messages...

Fair enough, it obfuscates when I’m sending a real message vs a “cover” message

...but helps hide who you are talking to

This is the bit that has me feel uneasy. Isn’t the cMix metadata obfuscation is strong enough to hide who I’m talking to?

It’s not like there is only one conversation going on, so therefore the participants must both be Alice and Bob.

To me, adding this last phrase makes the network sound weak without cover traffic. Do we really need more “help” in hiding who we are talking to than is provided by the underlying network?

Can the advantage of this “help” be quantified?

I dived into the wiki page and read:

[Cover traffic] …is important for resisting attacks on the anonymity properties of the xx network. Because your connection data leaks (true for all network connections), that can be used over time to determine who is talking to who. By activating cover traffic you are making these statistical disclosure attacks much more difficult.

Maybe add explanation on when Cover Traffic is activated (such as when the app is running in foreground).

This is important to know for traffic (when metered) and battery consumption reasons, but also from a privacy perspective too.