[Fluffy]

I found this usual suit combo yesterday
playing 3♥ dummy had

♠K5

And I held
♠1062


LHO first lead before seeing dummy is ♠4

no matter what, third spade is gonna be ruffed in dummy and there is no comunication issues for the defence. All that amtters is to score ♠K as a trick.
Is it more likelly that LHO underled ♠A on a suit contract or ♠QJ?
Does it change if you suspect/know from the bidding that spades are A:5-3, B:4-4 or C:3-5?

[mgoetze]

Well I doubt he underled his ace at matchpoints. Too hard for me at IMPs.

[fromageGB]

Whenever I have had Kx in dummy and only xxx in my hand, I obviously play the K. It always loses. So I think underleading QJ is more likely than underleading the A.

The only distribution that makes a difference for me is 3-5 (LHO-RHO). If I felt sure about that, I would go up with the K.

[655321]

Did you know your opponents, there are people who love to underlead Aces and others that never do.

On a good day, if you call for a low card RHO may play the Ace from AJxx(x). I have seen this work (haven't tried it myself!) even when declarer had 2 small opposite the Kx in dummy.

If spades are 3-5, then LHO hasn't lead the 4 from QJ4, OTOH perhaps RHO will be more likely to play the Ace instead of the Jack when he has 5 spades instead of just 4.

[Phil]

655321, on 2010-December-31, 15:53, said:

On a good day, if you call for a low card RHO may play the Ace from AJxx(x). I have seen this work (haven't tried it myself!) even when declarer had 2 small opposite the Kx in dummy.

Indeed - especially when

- dummy has a threatening suit
- dummy doesn't need a pitch
- we make the play in tempo
- dummy didn't show strength in the auction (and an underlead would have been more attractive)

[gnasher]

Unless there's a reason to hope for RHO to play the ace from AJ, I'd always play the king. If I play the king when they've underled QJ, it's just a Grosvenor; if I play low and they've underled the ace, I've chucked a trck that was legitimately mine.

[nige1]

655321, on 2010-December-31, 15:53, said:

Did you know your opponents, there are people who love to underlead Aces and others that never do. On a good day, if you call for a low card RHO may play the Ace from AJxx(x). I have seen this work (haven't tried it myself!) even when declarer had 2 small opposite the Kx in dummy. If spades are 3-5, then LHO hasn't lead the 4 from QJ4, OTOH perhaps RHO will be more likely to play the Ace instead of the Jack when he has 5 spades instead of just 4.

gnasher, on 2010-December-31, 19:05, said:

Unless there's a reason to hope for RHO to play the ace from AJ, I'd always play the king. If I play the king when they've underled QJ, it's just a Grosvenor; if I play low and they've underled the ace, I've chucked a trck that was legitimately mine.

Oh dear I'm convinced by both these arguments

[han]

Why does it matter whether something is "just a grosvenor"? Seems a strange argument to justify getting it wrong.

[pooltuna]

nige1, on 2010-December-31, 21:29, said:

Oh dear I'm convinced by both these arguments

You are just cooked on this hand if you need a ♠ trick at least for all practical purposes. I tend never to underlead either the A or the QJ against a suit contract. Consequently you may have to hope that LHO thinks he is defending a NT contract , failing that you have to hope LHO is one of the players that underlead the QJ.

[wank]

is underleading aces really that rare? if i've got a manky hand to lead from and the auction suggests people are generally balanced, i think underleading an ace is fine - it's got a nice upside. i would much rather do that than cashing an ace unless it's a long suit.

[han]

Again in response to gnasher's comment, if there is Kx in dummy and you have Jx in hand, wouldn't you almost always play for LHO to have underled the queen? Isn't that the same, if LHO underled the queen and you play the king it may be a grosvenor, but so what? Let's just try to make the contract as often as possible.

[barmar]

Underleading aces should be rare. In Lawrence's book on opening leads he writes, "As terrible as leading an ace is, underleading an ace against a suit contract shows a new magnitude of bad judgement." He then gives 8 reasons why this is.

There are rare times when underleading an ace is right. I thought Lawrence mentioned them at the end of his book, but I can't find it. From what I've read elsewhere, the best reason to do it is BECAUSE it's so rare, and you need to trick declarer. But you have to be careful, because you could also trick your partner. Lawrence gives a great example. A player leads low from Axx, and dummy shows up with QJxx. Partner has K9xxx, and since he doesn't expect the lead to be from an ace, he assumes declarer has the singleton ace, and plays low. This allows declarer to win his singleton 10. When the opening leader gets in later, he assumes declarer started with KT, so he cashes his ace, which gets ruffed. Soon after, declarer is able to take a ruffing finesse with dummy's QJ, and get a pitch.

[straube]

Others have pointed out that the lead could be from QJx. That gives his partner a decision when dummy plays low and he holds ATx.

[barmar]

From QJx, the normal lead is the Q (or J if playing Rusinow). Leading low from this is even weirder than leading away from an ace. The only reason to do it is if the leader is desperate to get out of his hand.

[awm]

There are some reasons that Gnasher's Grosvenor versus normal trick distinction matter. Logically, the goal when you are playing is not necessarily to maximize your expected IMPs, but rather to score more IMPs than your opponents. If your team is a favorite to win the match, then there's not necessarily symmetry between winning IMPs and not losing IMPs. To take an extreme example, suppose a two board match where my team is a 90% favorite. If I win 24 IMPs on the first board, my chance of winning just went from 90% to 100%, but that's only a 10% improvement. If I lose 24 IMPs on the first board, then my chance of winning just went from 90% to 0%. The upshot is that if I'm the favorite, the benefit from winning some number of IMPs is generally less than the negative effect of losing the same number of IMPs.

Further, opponents who are trying to give me IMPs by making silly leads will usually manage to give me some IMPs eventually, even if I miss the current opportunity to take advantage. Opponents who are playing tough defense will be difficult to defeat, so it's important that I don't let them set me in a cold contract by playing them for making a silly lead.

Finally, there's a possible psychological effect. Going down in contracts that should make is more frustrating (to most of us) than going down in contracts that really have no play... even if the opponents managed to find a misdefense to the "no play" contract that might allow us to make.

To return to the original question, I don't think it's possible to answer this without taking into account the auction and the skill level of the opponents. For example, intermediate level players (especially in the US) are often taught to never underlead aces. Further, they are more likely to play the ace from AJxx in third position than a strong player would be. This combination makes ducking the percentage play against intermediates. Stronger players will sometimes underlead aces, but whether they will do so depends a lot on the bidding. If the auction has revealed that most of our side's values are in dummy (i.e. I opened a weak two) then the ace underlead becomes reasonably likely. If LHO is marked with most of their side's values or could easily be stuck with a lousy holding in every suit, then the ace underlead becomes more likely. If it seems obvious that dummy will hit with a solid (or near-solid) side suit then the ace underlead becomes more likely. If none of these are the case, then the ace underlead is very unlikely. In my experience good players don't underlead aces "randomly" on ordinary hands -- but they will do so when circumstances dictate. Of course, leading low from QJ in a suit is a very weird lead too, especially in a suit where neither dummy nor declarer has shown length.

[gnasher]

han, on 2011-January-01, 16:35, said:

Again in response to gnasher's comment, if there is Kx in dummy and you have Jx in hand, wouldn't you almost always play for LHO to have underled the queen? Isn't that the same, if LHO underled the queen and you play the king it may be a grosvenor, but so what? Let's just try to make the contract as often as possible.

It's much more common to lead low from a queen than low from an ace. With Kx-Jx, therefore, playing low gives good odds.

Leading low from QJ is (in my experience) rather less common than leading from an ace, but let's suppose that the two leads are equally likely. Superficially that makes this a random guess, but it's not, because playing low is a higher variance action than playing the king. If I play the king and it's wrong, I'm still in the same position as the declarer at the other table, so I still have a chance to outplay him later in this deal, or on another deal. If I play low and it's wrong, I now have to outplay him twice, once to get back to parity and once to win the match.

From the point of view of team morale, I'd rather have this conversation:

"Plus 50."
"Flat. Sorry, I could have made it: they underlead ♠QJ at trick one, but I put the king up."

than this one:

"Minus 420."
"10 out. Sorry, they underlead ♠A at trick one, and I played low."

Similarly, consider the effect on the opponents' morale: getting away with a random Grosvenor might amuse them, but nothing like as much as a sucessful ace-underlead that gains them 10 IMPs.

This post has been edited by gnasher: 2011-January-02, 09:28

[han]

And I thought that bridge was difficult enough without having to worry about which way to lose 10 IMPs is worse for morale.

[FrancesHinden]

I've seen low from QJxx(x) about as often as an ace underlead, but from different players, so this is where it really helps to know the opponents.

[jmcw]

FrancesHinden, on 2011-January-02, 11:06, said:

I've seen low from QJxx(x) about as often as an ace underlead, but from different players, so this is where it really helps to know the opponents.

Seems right on to me. But, what if they don't lead a ♠ at the other table?
Left to play the suit yourself there is no option but to play for the Ace onside. So pop with the King at trick 1

[JLOGIC]

I always play the king vs anyone when it's Kx. Not for morale.

I mean... it is often the best choice of a tough decision to just underlead your ace. But with QJ you always have an option that most people choose (unless dummy has shown length in the suit...but dummy has Kx). If they had the 9 or the 8 they would be even more likely to lead a high one.

If dummy has bid the suit and i have Kxxx in dummy and Tx in my hand I am always playing low since underleading the QJ is semi-normal.

Hanp, keep fighting the good fight!

Gnasher, no offense but whenever someone says they think it's exactly 50-50 and base the decision on their psyche/morale I think they are copping out of making a decision. How could one objectively decide that this decision is 50-50?! Some people just hate to make a decision they could get wrong and thus rationalize it with "well they could have beaten me anyways..." to protect themselves.