Ok ok, calm down. Let's go back in history to oooh, 1998 - the year Half-Life was released.
Lets think of it's from 1997-1998 - Quake 2, Unreal, Sin, Shogo, Hexen 2... to name just a few. Technically, Half-Life was a little above Quake 2 and somewhere on-par with Sin in the graphics department. But still, Unreal held the eye-candy award. With physics, Half-Life didn't surpass either. With sound (processing), Half-Life just held it's head against the rest.
According to your argument, it was average. Graphics are average, physics average.
So then, why was it so successful? Was it because all it's players were blind with no sense of true physics? If so, why didn't the other games succeed?
It was successful because it had not just a story, but something more than that. It had characters you could believe in, fantastic soundtrack, and gameplay much better tha the keycard-based mechanisms of Quake and competitors. It supported most configurations. Never before did people care more about what was going on in the game world than how long it was until they beat the game. We became attached.
Ok, looking back at the game now after X replays we can see there were only 3 or 4 scientists, they all had the same voice, and Barney the security guard was everywhere. Soem of the level design was TERRIBLE. But we could look past all that because we were hooked on the gameplay and the atmosphere. Second, third time around we would spot the flaws, but the first play was the one that counts.
So, out of all the games, Half-Life won not on it's engine or its looks, but on its execution of a story in a game environment. It did it brilliantly. It set a benchmark. It proved graphics aren't everything, just as 3D cards were becoming standard.
As for the HL2 battle, I expect it to win again, but lets not predict the winner based on tech alone, eh?