Buying used probes is like buying used underwear. I'm not kidding; who would sell a perfectly good probe or their comfortable underwear? There is so much that can be wrong with a probe that you're always better off buying new ones. As soon as a probe starts to develop intermittant problems I throw the whole thing in the bin after I cut the wire so nobody gets tempted to pick it up and use it again.
You can apply this same argument to other types of used equipment. How would sell a perfectly good Fluke 87? Who would sell a perfectly good Tek 2465? People that bought new equipment, no longer need it (company downsized, outsourced engineering/production, hobbyists changing priorities), or companies that go bankrupt / people that die.
I've bought quite a number of used probes, some of them Tek probes from the eighties, and the only intermittent one was a cheap 100 MHz switchable one that was worth less than $50 when new. A lot can go wrong with a probe during its life, but a lot can also go wrong during design/manufacturing. Designing and manufacturing a passive probe, especially one with a bandwidth exceeding 100 MHz, is a non-trivial task. The tip of those Tek probes does not contain a simple of the shelf SMT resistor/cap, for example. The wire isn't RG-174 either. I would not trust an 'off-brand' probe without careful testing. Brands I would trust are HP/Agilent, Tektronix, Lecroy and PMK.