Keep working on a painting until you’re sure it’s finished. Then come back again a few days later and work on it some more if you realize it’s not as good as you thought it was.
That seems like a “duh” kind of statement, but it’s inconsistent with lots of art book advice. We are told that it takes two to make a painting: an artist to do the work, and someone else to hit him (or her) on the head before it gets ruined. Freshness and spontaneity above all, we are told. Never overwork the paint.
That advice was a problem for me until I realized what a crock it is. My problem isn’t a lack of freshness—it’s that I am so often tempted to stop too soon. I get parts of the painting to look really good and the rest basically not too bad, so I want to stop rather than put in the extra hours needed to get the hard parts exactly right. That whole “freshness” canard is an excuse for laziness—something seen in the work of many a marginal painter of approximate smears.
If you really want the painting to look like you got every part of it right the first time (i.e., “fresh”), then do what Sargent did and continually scrape off anything that didn’t come out exactly right and paint it again. And again. And again, until it is correct in it’s calculated appearance of perfect spontaneity. Even if you have to paint it 100 times.
If a look of freshness is not what you’re after (it’s not something I’m all that interested in, myself) then just keep painting until there isn’t anything you know how to do that will make it better.* If you’re not willing to keep at it until the difficult parts look right, then you’re not serious about painting.
*Or you realize that this painting is just a dog and trash it. You should allow yourself to do that only very rarely, however.