Hey bugger you've had a bad run with gaskets.
This is from my own experience;-
Cometic certainly are the bee's knee's of head gaskets, normally re-useable (you'd want them to be at the price!) however they are not 'kind' to the deck face of alloy heads. I'll add that these gaskets are usually only warranted on boosted applications. Also, if there is another issue / cause of failure, then even a Cometic will give trouble.
The current range of those marked BK450 (black in colour) are utter rubbish. Maybe they were OK at some point, but certainly those that I've seen and tried in the past 3 - 4 years have failed in very short time. Well, that's been my experience, though some say they have had no problems.
I've used the Cooper Gaskets with great success for a very long time and still currently use them, though they are different these days since the removal of asbestos from them about 20 years back. I'll also add to that that preparation is everything though if success is to be assured, and if everything is not at 110% or better, then a copper gasket will give trouble in very quick time. I've also found in every case that the engines run just that tiniest bit cooler with the copper gaskets as they allow heat transfer through them where as the black ones are a heat insulator. This has a broader effect than maybe first realised, in that the block and head decks are going to expand at a considerably closer rate to each other with one that will transfer heat. Since they expand at the same rate, less chaffing, and better sealing results. In 90 times out of 100, the copper gasket is re-useable too.
Just coming back to preparation, in many cases of 'bad gaskets' - not all, just many - I've found that poor or inadequate or over-looked preparation has been the cause.
Deck flatness we all know is important, but even more so is finish. A milled finish, particularly in the case of copper gaskets, will give trouble. I generally grind finish the decks, however I have also oil stoned finished the decks too (prior to being able to grind them), while giving awesome results, it usually takes around 2 - 3 hours to get there, if the (mill) finish was pretty good to start with.
Next up, remove the top 2 - 3 threads from those for the studs in the block. This spreads the area of effect of the stud when torqued up, roughly by a factor of 3 to 4. Lightly chamfer the holes. Which actually reminds me that putting an overly big chamfer on the bores in the block can also lead to gasket failure. They only need be big enough to remove any burr from the decking operations and to remove that sharp edge and glow plug that can occur.
Head studs and nuts. Be absolutely certain that the studs are all the same material / treatment and age / number of cycles. If not, then there will be variations between the clamping forces of each stud. I know ARP Studs and nuts are good, however, I've probably had better success with the very late factory types along with the flanged nuts. I have come across a few so call genuine ones in recent times that 'wobbled' when screwed in to the block, clearly they were not and I have seen a few case where these break off in the block. This comes about because as they are being torqued up, they are also being pulled straight, bending and loading one side of the stud, instead of even loading through the core of the stud. If you find any that wobble as they go in, I'd suggest binning them. I torque the late factory items to 50 ft/lb over 5 steps. I also 'bed' in the threads by first torquing them up dry 4 times, then finally with oil. I think ARP have some good advice on this. When we build an engine here, we have log sheets for recording lots of info, including the length (at the time of fitting) of the head studs. Then at later times, if the head is ever removed, we can measure the studs again to check for stretch, which would be a fail of course. I have to say, I haven't had a fail in a genuine stud yet.
Also, don't overlook the rocker pillars, the quality of them or the deck pads on the head, again, I grind finish them. Also, the locking plate for the screw that locates the rocker shaft. It seems that there were 2 types, on that was short and fitted only under the rocker stud and the long one which also went under the head stud. If you have the latter, give it the chop so that it only fits up under the rocker stud.
After the initial run up to operating temp, I let them cool overnight then re-torque again. I usually do them again around the 1000 km mark as well, but not always (Yeah, OK, I'm lazy).
So, that's been my experience.
Opps, sorry for waffling,,,,,,,,,,
PM also sent.
) I've had a look at the gasket and there may be evidence of number 4 touching it 
