The Minecraft Wiki article on Spawning is a good start. Specifically;
Every world has a "spawn point" (really, an area) located within a few hundred blocks of the origin (0x, 0y, 0z). This point is stored in the level.dat file's SpawnX, SpawnY, SpawnZ attributes (although SpawnY seems to be ignored). If the spawn point is deemed unacceptable, a new, temporary spawn point will be selected (although the criteria for "unacceptable" is unknown).
So you always spawn "somewhere" near the center of the map (0,0). Generally the game will try to spawn you on a sand tile near water (The "washed ashore" effect). So, if you always load the same seed, then the algorithm should always plop you down in the same spot that is "close to" the center of the map.
The rules for multiplayer are somewhat different however, and the article is worth the read on that aspect.
This is an issue of multiple flukes combined into one giant mess.
The map was originally generated prior to 1.7.10, however the chunks containing the stronghold were generated after an update. Because of this, the End Portal was in a different location than what the Eyes of Ender reported.
To resolve this, I took the seed, -4365712075075430210
, created a new world in Creative Mode, and navigated to the above coordinates. When I used an Eye of Ender, it proceeded into a different direction about 100 meters away. Upon digging down, sure enough there was the End Portal in my creative world in a completely isolated section of the stronghold. Writing down the coordinates, I went to them in my hardcore world, and the portal was there as well.
In short, if you create a world in a previous version of Minecraft, upgrade to a later version, then explore the chunks containing the stronghold, there's a solid chance that the End Portal is going to be in a slightly different location than what the Eyes of Ender report. You could generally find it just by navigating the stronghold, however I was not able to locate my specific End Portal because it was isolated and walled off all by itself in my specific seed.
Best Answer
Yes - beds explosion behaves just like other explosions, the wiki contains a detailed, lengthy section discussing the exact procedure, but the very essence applicable here can be boiled to 'the more rays hit, the more damage is caused', and 'the closer the entity to the explosion, the more rays will hit'. As the dragon perches on top of the bedrock column, placing the bed closer to the dragon will cause more damage, simple as that.
What is more important in speedrunning though, is the knockback of explosion and AI of the dragon. Perched dragon upon losing 1/3 of its hitpoints will invariably, immediately take off and start circling again, costing the speedrunner plenty of time. Depending on the random attacks, hunting the dragon with a bow, randomly placed beds, waiting for the dragon to finish perching and then blasting the beds - these are all old strats that just take a lot of time. The current meta is to detonate the beds while the dragon is in the process of descending, just before it perches over the pillar. The knockback of the explosion throws the dragon a short distance upwards, giving the player time to place another bed, and detonate it again - since the dragon never gets to perch, the 1/3 hitpoints count is never initiated, and the player can kill the dragon in a single perching cycle. It requires very good timing, as detonating the bed too early won't deal enough knockback, giving the dragon enough time to perch before next bed is placed, and obviously doing it too late will put the dragon in the 'perched' state, meaning it will take off after losing just 1/3 health.
The closer it is to being perched the more damage it will take, making it possible to finish it off with as little as 4 beds, but it obviously increases the risk of allowing it to perch. Detonating the bed sooner is safer but may take 8 and more beds.