The numbers I was trying to put together had nothing to do with paladin-versus-warrior or paladin-versus-druid. In fact I lumped warrior/paladin into the same vein because ultimately their effective health works out to be about the same over a "four hits in <5 seconds" scenario. Since the case being assumed is that all four hits are not avoided, it's guarenteed that the warrior's shield block charges
will be gone after the first two, leaving them at a ~40% chance to take an unblocked swing to a paladin's ~10% (numbers vary with gear and redoubt uptime). Playing odds, it's fairly likely that the warrior will take at least one less block than a paladin in this case, and the 300-400 blocked damage will compensate for the 4% gap on IRF vs Def Stance over the course of the 4 swings. As has been pointed out, see WWS entries if you need hard field data.
What I was trying to do was benchmark the point at which adding more hp doesn't do you any good for surviving stomp + 2x2. At some threshold value, you effectively need one less major heal to survive stomp. Once your HP hits that mark, more HP doesn't really reduce the amount of heals-per-attack that you need. Below that point you'd need either a substantial amount of HoT/quick healing, or an additional large heal. I'm not saying 23k without seed is absolutely essential, I was more getting at passing that mark is fairly worthless; if you get X big heals you live instead of needing X+1, but if you get X-1 you're still dead, just from a higher base health %. However, I believe I used a poor modeling scenario for healing, and that the catchpoint should actually be about 1k lower. Unfortunately we didn't raid last night so I didn't get to observe any numbers/patterns to try and crunch out a more feasible scenario.
-Splug
_________________
(Usually) Prot warrior:
http://www.wowarmory.com/character-sheet.xml?r=Illidan&n=Splug