What load are you running to get 2720 in a 20" barrel?
No reason to go faster than a 1-7" unless your location is extreme or your speeds are low, but they don't seem to be. Last year I ran a factory Savage 26" barrel that I was able to get 2820fps out of with a load of 24.4gn of Varget, also a 1-7". I now run a 30" Shilen select match 1-7", 6 groove button rifled barrel that I definitely can't complain about. This year I am using a light load to save brass life that really seems to work. I am using 23.5gn of N150 and get 2715fps with an ES=14 and a SD=4, can't ask for much better than that IMO. I also have a 8lb jug of Varget that I am going to open and work up some loads and see what it can do out of this 30" tube. I'm not really concerned if someone says a 30" barrel is a waste to shoot at slower speeds, it works great for me.. My next is a 30" Broughton 5c 1-7" that I have high hopes for, with claims of higher MV, easier cleaning and longer barrel life, hopefully that all holds true. Also, barrel length and contour will be determined by your allowable weight. Depending on the rest of your setup, you may be close to or over the weight limit for FTR.
IMO there is no difference in button vs cut rifled barrels as long as you stick with a reputable maker. Everyone can make real hummers, and everyone can make tomato stakes...Just reality.
I use naked bullets in a naked barrel and see no reason to change that, JMO. I don't believe you would be able to make it 3000, but if you did, you would have severely decreased your brass life. I wouldn't push them past the ~2820 node anyway, once you start getting into those high pressures, the slightest changes will show up that much sooner or more pronounced. If you want to really make the 90's fly into the 3000+ range, you should choose another cartridge that is more suited to the job.
A kernel or 2 shouldn't make a difference, but a .1gn does. I worked up my loads basically the same way as I would do for anything. I did do a test of 5 different loads with the only difference being a .1gn of powder and you can see the groups shrink and then open back up over the course of the test. The largest group group was approx .6MOA and smallest was in the .2's, so both would still be SUB-MOA but it was neat to see a consistent fall and rise of group size, a real show of barrel consistency IMO to see real results like that. That test was repeated 3 different occasions and revealed the same results each time.
Chamber will be personal preference, some run a Wylde chamber etc.. I use a .223 Rem match chamber with .252" neck which prefers Lapua brass. You can extend the throat of your barrel to suit you. On my Savage barrel I throated it so I could seat the body-boattail junction of the bullet above the neck-shoulder junction of the brass for an approximate OAL of 2.700". With the barrel I use now, I went only half the distance as I wanted to also shoot the 80SMK and 80AMAX and still have them seated far enough into the case. My OAL for the 90's now is 2.580" and since I neck turned my brass, I don't have any worry about donut issues. I may yet still have it pushed back another .125" to get into the next velocity node, but we'll see.
Anyway, that's just my take on things, YMMV.