| I disagree that you have to raise pre-flop, and I disagree that Villain made a big mistake here (but he did make a mistake). Alternative analysis of this hand titled "How slow playing can be a mistake" or "Count the mistakes":
Scott all in pre-flop would likely have won a 250 pot. Scott first all-in on the flop would likely have won a 300 pot.
As played, opponent's call of 250 when the pot is 1750 is a perfectly reasonable action, and he would be a fool to fold in that situation. The fact that Villain bet on the flop after two checks is reasonable, although his bet of double the pot foolishly committed him to the pot. Only the size of his bet was the mistake.
In other words, if you hadn't doubly slow-played (call, check), you likely would have won a small pot instead of lose all your chips in a big pot. Unless you have the nuts, an uncontested pot is always preferable to a showdown, especially in a tournament when your opponent has you covered and losing the pot means tournament death.
I think limping with a big hand in the small blind is OK if you're on a stop-and-go. Call pre-flop, all in immediately on the flop. As played, you allowed a hand to price himself into a call.
Pavel, go back and review the hand -- Villain didn't just push all in with a pair of fours. He bet (although overbet) on the flop, suspecting weakness from all the checks. Only the size of his bet priced him in, and he absolutely had to call the last 250 for a 1700+ pot.
Better plays for both hands:
Scott all in on the flop.
Villain bet 150 on the flop when Scott checks. Then, when Scott moves all in and it costs another 700 to call (less than 2 to 1 pot odds), the pair of fours can fold. | 
