The problem with stating that they should be as close as possible is that you do not acknowledge the option of the samples getting better with a second iteration round. Mito has only stated that the next samples will probably be better, nothing about the ultimate accuracy you are assuming is going to happen. Maybe you have mixed his intentions and hopes with his expectations.
'Should be' is the operative word here, by saying 'should be' I'm fully aware of the possibilities that the sample might get even worse. If I said 'will be' then your argument would have some legs to stand on. And in all of my working life, never has 'as close to... as possible' or 'as soon as possible' or 'as much as possible' been construed as the 'ultimate accuracy' perceived by you.
You can keep on disputing my writing as much as you like, I thank you for keep bumping the thread. Peace out!
'Should be' is a 'will be' with a higher than 50% probability as percieved by the writer. 'As close to ... as possible' when inverted is the same as 'impossible to get closer to ...'.
What your sentence suggests is that there is a more than 50% chance that it is, by human means, impossible to imporove upon whatever gets printed on the second samples.
Anyway, I'll stop here. I guess you understand my point and I get it that you use superlatives more casually than I would.