Around every 4th of July I reread Michael Shaara's The Killer Angels, which is a book about the battle of Gettysburg during the American Civil War. There's an early passage about the Southern general James Longstreet's unease over the Confederate push north to Pennsylvania: He had never believed in this invasion...He did not believe in offensive warfare when the enemy outnumbered you and outgunned you and would come looking for you anyway if you waited somewhere on your own ground. Longstreet, one of the finest military minds of the age, spends much of the subsequent bloody fight knowing that Union forces had a terrain advantage impossible to overcome.
There's been a lot written about battleground strategy (particularly Alterac Valley) but I think all of us have known the sinking feeling you get when you realize that your side isn't going to win. Some causes of failure are relatively easy to pinpoint; starting a battleground with a heavy numbers or healing disadvantage often seals the fate of a match. And of course the collective quality of a team's gear will always play a role; people in Season 4 are unlikely to lose to those in Season 1.
All other things being equal, what I find most fascinating are the matches -- PuG versus PuG, or premade versus premade -- where the battle can swing either way depending entirely on each team's degree of foresight and strategy. Rarely, single players can sometimes decide the outcome; I once saw a protection paladin in a 2-cap versus 2-cap Eye of the Storm prevent the opposing side from taking any flags by parking himself in the middle and simply taking forever to die, and one of my own favorite techniques is to suicide/harass heavily-defended nodes in Arathi Basin and EOTS while Horde quietly caps elsewhere (you'd be amazed at the number of players who prefer an easy kill over responding to "Inc!" calls elsewhere). But failure and success are usually collective and hard to pin down. How do you convince people to do the less-glamorous jobs -- defense, distraction, crowd-control -- more likely to result in a victory? How do you know when the battleground is lost for sure?