All three settings will tend to yield a figure-8 pattern or something approaching it. The old-style gaskets tended to dry out and harden over the years/decades--not specifically from use, but more as a function of age. They could become partly or completely unable to seal the openings to the acoustical chamber behind the backplate.
When you see eBay ads for three-pattern capsules, they may say that "all three patterns work"--but that doesn't tell you anything. Sure, they'll put out a signal in all three pattern settings, but what is the actual pickup pattern? What would get my attention is "The capsules were rebuilt at Schoeps in [year], and haven't had any mishaps or accidents since then" (where the year is later than the mid-1980s or so). They could also say, "The three patterns work as they should"--which forces you to rely somewhat on the seller's skill as a tester--but at least that gives you the implied right to return the capsule or mike for a refund if the claim doesn't hold up.
The thing is, many microphones and capsules have been traded around so much that the current owner has no idea of their service history. Also, there's a certain "ignorance is bliss" mentality. It's reminiscent of "if you don't make sex education and birth control available to young people, they won't have sex" (or currently, "If we didn't test so much for coronavirus, our infection rates would be lower"). All I know is, if you do any amount of live recording and you've never dropped a microphone, you're in a small, fortunate minority.
So yeah, the figure-8 setting will generally produce a real figure-8 pattern, while the setting most vulnerable to the gasket problem is the omni. But the easiest pattern to test is the cardioid. You can just plug the mike into a preamp or recorder or mixer and listen to the output over headphones as you talk into it from various angles. Pay especially close attention to the sides--there should be no particular "dip" in the signal level at 90 degrees (right where the null of a figure-8 would be) followed by a return to the expected signal level as you rotate the mike further.
The problem with testing the omni pattern is that its on-axis response differs quite audibly from its off-axis response at upper-mid and high frequencies, as in all omni microphones of this size or larger. The high-frequency response is quite audibly reduced when you rotate the mike so that the sound source is at the side or back. Actually, this is something I recommend that everyone try out with any omni that you have, to help understand why omni microphones function as they do in reverberant spaces. Our hearing is the most sensitive in this region, so the typical ~6 dB difference comes across as greater. If you're not expecting it, you could easily think that something's wrong in the omni setting, when it's actually working exactly as designed. But again, the key with the MK 6 is the sides; if you slowly rotate the capsule or mike so that you start at the front of the capsule and reach the 90 degree point--then if the signals actually increase as you go past 90 degrees, you've got a problem. It helps if you can do this test in the quietest, most acoustically "dead" location possible, because all you want to hear is the direct sound.
--edited later to add: I hope everyone is aware that a figure-8 pattern can never be tested in the way that I'm describing. This is because your voice will reach your inner ears via bone conduction at the same time as it's going from the microphone through the preamp to the headphones. The rear lobe of a figure-8 microphone has the opposite polarity from the front lobe, so on one side of the mike those two signal paths will reinforce each other, while on the other side they will interfere "destructively" and tend to cancel, but very unevenly with regard to frequency response; it will just sound very wrong and be misleading. To work around this, you can either have someone else do the listening for you, or else record your test without wearing headphones, then rewind, put on the headphones, and listen separately while you're not talking.