[noerps]

The tradeoff is the same as keysize in crypto. It is time.

If you choose to communicate a second time from the same endpoint with the same equipment you may achieve only pseudonymity.

Since Tor doesn't limit the encapsulated protocols, it depends on the implementation and awareness of the user and you can't put a number or percentage on that.

Imho the Tor-role has changed, it provides access against censorship, DPI, region-partioning and hidden services. Simply try to access youtube or any other global service via different tor exits, that may be intresting, not from an anonymity point of view.

[rane]

Out of curiosity, why does the equipment matter?

[noerps]

There may be information leaking (as in fingerprinting) from your device (that you are not aware of) that would allow very easy correlation, like in a mac address or existing session cookie, similar address in a public open network or whatever you can imagine to compromise your anonymity, that gives an adversary any advance to successfully correlate your current session with one of your previous sessions.

At this event your anonymity becomes a pseudonym.

The next step would be to try to reproduce or predict behavior and setup a trigger for that information.

If the loss (compromise) of anonymity or pseudonymity may lead to imprisonment, torture, assassination or death this maybe an issue to consider.

If you try to obfuscate your access to porn, it is a completly different story.