These ratings are bull, pure and simple.
Yes, some companies have a machine that tests the switches (others don't, and just copy the number from another company that makes similar switches, or even just make something up out of nowhere), but the test conducted by this machine does not reflect real-world usage, where switches are a) subjected to a wide variety of different levels of force, b) subjected to force coming from a variety of different angles, c) sometimes held down for prolonged periods, and not always with the minimum amount of force or a force that is perfectly centered, d) subject to incursion by a wide variety of kinds of debris, vapor, liquids, etc. over a long period of time, and then used, and e) have endured the strains of being installed into a keyboard, soldered in place, re-soldered in case of defects or repairs, having keycaps changed, etc. Also, if the test is done on a lone switch, it does not capture points of failure that occur between the switch and the rest of the keyboard-- for instance, if the switch's design places undue stress on the point where the switch is soldered to the PCB.
Next, the actuation number does not capture changes in the quality of the switch over time due to improvements or degradations of the manufacturing process, unless the switch is re-tested. If a manufacturer's quality declines, they won't re-test the switch to reflect this change. How old do you think Cherry's numbers are at this point?
Finally, the number seems to be subject to a lot of manipulation. For example, there was a time when MX switches other than MX Black were rated at only 20 million keypresses; only MX Black was rated for the full 50 million. You can still see these numbers occasionally from companies that have been copy-pasting their marketing materials for a long time. As far as I have been able to identify,
no change to the design or manufacturing of the other switches was made when Cherry decided to rate them at 50 million. It's as if they simply decided that their test of MX Black switches was actually a test of "the MX switch" and so all MX switches would thereafter carry that rating.
The real question is: "How long is the warranty?"
Unfortunately, this doesn't help either, since warranties vary not only by length but also in terms of how easy they are to take advantage of, and how much of the purchase price will go towards covering it. Any company can offer a super-long warranty so long as they make it so hard to use, or so expensive to use due to non-covered shipping costs, that people with problems are generally rejected or just made to give up. Alternatively, they can just mark up the product enough to cover a longer warranty.
Case in point: tT eSports Poseidon Z models carry a longer warranty than their Cherry switch models. But do you really think that they last longer?
