As a writer, I am always careful to run my deals through "Deep Finesse". This software (there are other similar products) can analyze a deal with 100% accuracy.
While computers still can't play bridge anywhere near the level of top humans (see my article on computers and bridge), they are perfect when they can see all four hands.
Accordingly, if I ever write "Four spades can be defeated by a heart lead," you can bet your life that four spades can be defeated by a heart lead.
Sometimes, when verifying an analysis, I stumble upon some illuminating conclusions. For example, consider this deal from the 2008 Cavendish Pairs:
|A 10 7 6 5|
Q 8 7 3
J 8 2
|K 9 8|
J 10 6 5 4
A Q 8 5
| ||Q J 2|
A K 9 2
9 7 3
10 3 2
| ||4 3|
A K Q 6 5 4
J 9 7 6 4
At many tables, South reached 5, usually doubled. West led a trump. Declarer has a slow spade loser as well as the A to lose. What do you think?
All the declarers won the trump lead in hand and played a club to West's ace. West couldn't play more trumps. So, declarer was able to win any return, come to hand, ruff a club in dummy, come back to hand and ruff another club.
If the club queen falls tripleton, declarer avoids any additional club losers. As you can see, the club queen doesn't fall, so declarer has to lose another club trick for down one.
But what (you ask), if declarer plays clubs differently? On the third round, instead of trying to drop the queen (which doesn't work), declarer can (looking at all the cards), lead the J from hand. East has only the 10 singleton remaining. If West covers with the queen, the 10 is pinned and the suit is good. If West doesn't cover, declarer discards from dummy, and still had a trump left to ruff out the queen.
Should declarer play this way? No. It is against the odds (unless there are some really strong clues from the bidding and early play). Dropping queen‐tripleton from either opponent is twice as likely as playing for East to have started with specifically tripleton 10.
But, looking at all 52 cards, the contact can be made. Right? Wrong!
Before writing up this deal, I checked with Deep Finesse to see if there is a way to beat 5. It told me there is (it also tells how).
West leads a trump (as good as anything). Declarer wins and plays a club (nothing else is better). West ducks the A! Yes. He refrains from taking his A on the first round of the suit.
Now, the contract can't be made. Declarer can come to hand, ruff a club, come to hand, ruff another club, come to hand, draw trump… but he has lost control. You can try it trick by trick yourself (it might help to lay out the cards on a table), but don't fight the computer.
The computer is 100% correct that if West ducks the A at trick two, there is no way for declarer to make 5.