Sorting similar vocabulary

I keep hitting a wall with words that look almost identical. First I isolate the radicals to find a logic. Then I check if the usage contexts are different enough to separate them. Finally I drill them in pairs to force the distinction. This system keeps my accuracy up but it takes a lot of energy. My consistency is good but these lookalikes are tricky. Anyone have a better way to sort these?