Now that I'm on my break, I'll try to elaborate on the example in ToP:

Ungar and Brunson are heads up at the 1980 WSOP main event. Ungar has a significant chip lead. Brunson is dealt A-7, Ungar 2-5 (I think). Pre-flop betting wasn't significant.

Flop comes down A-7-4. Brunson bets out $17,000 (a significant amount compared to stack sizes), making the immediate pot odds poor for a call. After the match, Ungar said to the effect that he wouldn't have called for much more, but he knew that if the 3 hit, he could bust Brunson for all his chips.

When the 3 hits on the turn or river (I forget), Ungar pushes Brunson all-in and wins with the straight.

I'll leave it for the upperclassmen in this forum to say whether this is still an overall good strategy, but the point made in ToP was that given the information that Doyle had a hand that he might commit all his chips with, you can look beyond the chips in the pot to determine whether the pot odds are good for a call. (Which, counting Doyle's remaining stack, were just about right for the 11-1 odds of hitting the 3.)