>We know what viruses are, how they work in principle. We can use that knowledge to predict what is going to happen if we disrupt how they multiply or spread or gain access to our cells.

I think that's a misapprehension of how drug development works in practice. The models can lead towards discoveries, but most drugs fall out of the pipeline due to being unexpectedly unsafe or ineffective. The "checking if it actually works" step is the most important part of medicine by a large margin.

What I meant is we know what happens if we have working vaccine. Meaning that working vaccine leads to less people being sick and that leads to less deaths from virus.

As to the vaccine itself it is usually easy to come up with a formula that can disrupt the virus (any potent poison or lava flow can do that). The difficult part is that the formula must also not disrupt anything important. No amount of knowledge about virology can help with the second part, human body is vastly more complex than the virus.

I should have been clearer. I meant that if it was shown to be effective in clinical trials, then it's unsurprising that it worked when deployed in practice.

The strains in the wild at the time of the clinical trials are different from the strains in the wild now. Furthermore, the trials did not test herd immunity, only direct immunity.