Tank is a relative concept, in this case. It is not the thickness of the steel wall of the case that determines the situation, but the accuracy of the processing of connections and adjacencies.
According to the classics, the crown gasket, the height of the tube and the pressing force of the CG lever are checked. Also, we check the gasket and tightness of the case and caseback threaded connection.
From the factory, the geometry of some parts may be disturbed in such a way that a sufficiently tight abutment in the above parts will not be tight enough, and not even compensated by a rubber gasket.
Although in modern VSFs this is rather an exception, and such a situation may arise more likely in the case of a marriage, with old REPS this can occur on a large scale.
Further, there is an even more vulnerable part, this is the adjoining of the crystal to the case. If the caseback is transparent, then it is also there. As far as I understand, on a well-made watch, this is the most common problem, and at the same time I do not see everywhere that within the framework of the "waterproofing service" the crystal is republished as standard as it should, just prophylactically.
This problem no longer refers to the quality of the parts, which is really good, or the material, but to the quality of the assembly, the problem with which is much more likely.
And I'm talking about the most common and well-made cases from large factories, it is clear that the further from the standard version, the higher the risks.
Of course, if the crystal is defective, it may simply burst, or the entire batch of the model, or even the entire release, may or may not have such a negative property. And you will understand only when you check.
Thus, I wanted to say that in addition to the external solidity provided by a massive metal case, thick crystal, solid caseback, there are still "weak" spots, the strength of which determines the quality of water resistance as a result. And it is not easy to fully control them.
And in conclusion: I understand that if my conditional rep PAM285 had a water resistance of 2500m as a GEN, it would clearly amuse my vanity. But to be honest, I find it difficult to imagine how the hours tested at 50m flow when swimming in the pool, or swimming in the sea, but for most this is enough. If you are not part of this majority, then you will have to risk more than that same majority.