It depends. Do you plan huge current per cell or low?
If you have run your test with a a current equal or above your planned load later on you also know that the resistance currently is low enough to suit that load. This will work great for loads up to 1A and you dont have to sort them.
As long as you randomize the packs they will work fine and will equalize over time. Some cells will have slightly higher current draw than other but this is not a problem.
High current packs like e-bikes you SHOULD check IR and make sure they are more equal.
There are threads where this is tested and you can see the behaviour.
Am I guessing? No Im not. I have around 29 000 Cells in my powerwall and followed a set rules when doing it. Currently after 4 years I have still not swapped out any cells due to issues of above.
For those interested the Ir is very much related to the tested current and the time frame you run the test and this also shows it.
if you plan and do want to know the IR you need a 4 wire tester. The testers on for instance Opus is useless.
You should have same amount of high and low capacity cells in each pack. You should also mix the resistance cells.
For instance if you sortr all your cells in 100mAh buckets and then randomize them this will be 100% true on larger packs like 80p or above based on statistics. Statistics can for sure be used doing low current powerwalls like most of us do.
Wolf and the others will soon chime in on the importance of Ir and what they have done.