First, the human eye can distinguish millions of colors when they're laid right next to each other.
However, in terms of seeing colors separately and trying to remember their matches, we are positively hopeless. Further, even small changes in light cause significant changes in perceived color even if you had a color sampling device, so cameras are as blind as we are.
The Smithsonian recently restored a Jim Crow era passenger coach. Naturally, their color research was pristine. However, having nailed the right color, they then altered the color to correct for the flaws in the fluorescent or whatever lighting in their display hall. Except the lighting industry was innovating like crazy, and three years later, you could buy true-daylight lights at Lowes. But now, that would make the color wrong, eh? So they are stuck using their 1970s tier lighting.
I am not Smithsonian tier, but I usually nail it. But even with my book of 1700 color chips, careful matching with photos and Photoshop analysis, I've ordered a $300 gallon of paint and had it not match the exemplar. This was burgundy, and my mentor had the same problem with another burgundy.
All this to say, this is a hard problem. It ain't you.
That said, you did err by expecting two cans to come out the same. The problem is, there is "rounding error" on the tint shots. Computer controlled tinters are an attempt to solve it, but even that is not perfect. This problem is much worse on smaller cans because rounding error matters a lot more when the shot is 1/4 or 1/16 the size... so you greatly amplified the issue with 1/2 pint sampler cans.
Yes, you must buy all your paint and "average the error" by mixing all the cans together. At that point sampler cans stop making sense.
At least you are working in "latex" (emulsion) paint, which is basically nontoxic and not hard to dispose of. Lucky you!