As a rule, yes, but it's not absolute. I believe there have been instances where a not-very-nice person posted something to FR, and included a graphic file for the purpose of tracking the IP of everyone who accessed the image.
All he had to do was include the image (even a 1x1 "web bug" would suffice), and then check the headers when requests come in to his server. He'd get a log of IPs of people who visited the FR thread that had his post.
Of course "all he had to do" is not that trivial for the majority of users, but it is possible, and short of blocking images, there's really no way to stop it. Still, a decent firewall (zone alarm) will prevent him from doing anything with that IP number.
My point (that I seemed to dance around) is that yes, it is possible to have a probe come from a third party as a direct result of merely looking at a thread on FR. And the unpleasant reality is that, thanks to the relatively nontrivial effort required to produce that effect, it's very likely (IOW "probable") that the party engaging in the practice is doing so with a malicious intent.
Moral of the story is to download zone alarm and install it! And if you're running XP, which I do not believe zone alarm supports as of yet, dig your way into the networking management and set up the built-in firewall. It's not as pretty as zone alarm, and not as configureable, but it will keep the wolf from busting through the door. (And if you're running a Mac, I don't know of any firewalls, so my "standard solution" applies: get a computer. :))