Do those two things, plus elevate the muzzle before each shot, and I believe you will have nice groups. Do not stick the bullet so far down into the case that the base intrudes into the powder area: US NRA published pictures years ago showing really ugly looking errosion all over the portion of the bullet that intruded.
Your bullets are apparently filling the grooves now, based on no Leading and the bullets stabilizing. Gas checks, more than anything, act as a structural reinforcement of the base of the bullet. Without that reinforcement, bad things happen to the base at higher pressures, and groups go wild. My Scientific Wild Guess is your muzzle velocity is up in the 1600 to 1800 fps range. With that high a velocity and the pressures necessary to generate that velocity, generally expect bad accuracy without a gas check.
Regarding nose position of the bullet, I prefer slightly into the rifling. I am sure you have used bolt-action .22 rimfire rifles. Virtually all manually operated .22 rimfire rifles I have examined thrust the bullet slightly into the rifling. You can see this on withdrawing an unfired cartridge from the chamber. With a relatively pure Lead bullet, it takes little force to do this. This keeps the nose area of the relatively soft bullet from slumping to one side or the other on firing.
The Tubal 2000 powder you are using is almost identical in burning rate to Accurate Arms 5744, at least according to this burning rate chart (see (122 & 123):
http://home.hiwaay.net/~stargate/powder/powder.htm There is a bit of .303 data using Lead bullets and 5744:
http://www.reloadersnest.com/query_pw.a ... r=Accurate 5744
Vectan apparently classes Tu2000 as being next to H4198 and IMR4198:
http://www.nobelsport.fr/nobelsport/fr/ ... e_2006.pdfSmall quantities of Tubal 2000 have leaked across the water to the US, and friends of mine class it as very close to H4198 and good with cast bullets:
http://castboolits.gunloads.com/showthr ... Tubal+2000There is some 4198 data for the .303 on the net:
http://www.public.asu.edu/~roblewis/SML ... a11a4.htmlhttp://castboolits.gunloads.com/showthr ... =4198+.303http://castboolits.gunloads.com/showthr ... =4198+.303http://castboolits.gunloads.com/showthr ... =4198+.303Accurate Arms 5744 and the two 4198s have excellent reputations for cast bullet shooting, but that does not mean Tubal 2000 will. The closeness in burn rate says that the data should not be too different.
The only .303 Tubal loading data I find on the Internet is here (only Tubal 3000, 5000, and SP7 listed for the .303):
http://www.scribd.com/doc/7278645/Table ... Sport-2006 It would be nice to know what your velocities were. That would tell you for certain whether you were getting erratic ignition. The fact that your groups have a huge vertical aspect, and a small latteral aspect, makes me suspect erratic ignition. It may be you can't get away without using a pinch of dacron, but using it scares me too much to use it. If you do, use just a grain or two.