My need to experiment, I might test a little more before calling it done, though you're in pretty good stead where you are, and could just keep it as it is.
My thought process, but take it with not just a grain but a helping of salt...id test swr on simplex 15-21, and again on the repeater channels. If it's tending downward on as you go up in simplex, and rising as you go up the repeater Channels, you're likely pretty close, and your lowest swr is in the gap between simplex and repeater frequencies. If you're in that dip and the repeater channels (overall) are higher swr than the simplex, you could probably take a small amount more off to bias it a tiny bit more toward the repeater channels. If the repeater channels are already lower swr than the simplex, or the top of simplex (22) and the bottom of the repeaters (15) are about the same, I wouldn't take any more off.
And if I'm off base, by all means, let me know...I'm approaching from a logic/theory side of it, and know the real world doesn't always follow, and I'm always willing to learn. I know my 5/8 wave laird instructions said I should trim a good 3/4" to be centered where I want it, but nanovna and swr meter said it was almost spot on as is.