So you do accept scientific evidence, then- simple (approximate) models that explain well-verified patterns should be taken as practically true, until their limits are found. Right?

Yes and no, depends on the context. In reality, some of patterns can be taken as practically true and some of it is not.

As an example, If I drop something from top of building, it's always go down to the ground; this pattern is always reproducible with the same result by all peoples who can test it. But, if I drink hot water when I'm sick and I get healthy in the next morning, that would become biased, because it's not always reproducible with the same result.

I think, it's only a matter of how someone defined the value for "well-verified" and "limit" until it become true for himself.