Ace for Attitude, King for Count

[Finch]

In another topic, mikeh said:

 

BTW, I was interested to read the suggestion that North could chose what message he wanted to get on the opening lead: A for attitudde and K for count (or reversed). I have never played that and would need to think about it. My preliminary reaction is that I don't like it: my choice of A or K against a suit partial shows different combinations in my hand: the lead of the K denies the A...

 

The idea of A for attitude, K for count is extremely common round here. As it clearly isn't so internationally, perhaps it would be helpful to explain the arguments for it.

Suit Contracts

Start from the premise that you are playing normal leads, i.e. top of touching honours. What signal do you expect partner to give at trick one? Leaving aside a few obscure exceptions (such as when you give a suit preference signal at trick one) there seem to be three traditional approaches:

 

i) Always gives count (or always gives attitude)

ii) Give count or attitude based on a set of rules (e.g. usually attitude but count if the Queen is in dummy)

iii) Tell partner what he wants to know

 

i) is too inflexible for me. iii) may be theoretically best, but I find that all too often I don't know what partner wants to know. Many people who start off playing iii turn it into a complicated form of ii.

 

So I end up wanting to have rules rather than having to use judgement every time partner leads. You can discuss all sorts of more complicated rules, but we've found that you usually want to see attitude at trick one, except that quite often you want partner to give count when you lead the King. So a nice easy rule becomes "give count on a king lead, otherwise attitude".

 

Having got that far, it becomes obvious to give the opening leader a little more flexibility and allow them to lead the king from AK when they know they want count in the suit - usually when leading from AKQ or AK to length. Very occasionally we take this the logical stage further and lead the Q from KQ looking for attitude, but that's very rare - you have to be confident first that partner won't put the ace on your queen lead and secondly that partner won't get confused about your holding in the suit.

 

I'm not claiming this method is perfect (it's easy to improve by adding more rules) but it's our belief that it's the best rule that is that simple.

 

NT contracts are different. As it happens I still (sort of) play K for count, others for attitude but the logic - and the choice of cards to lead from particular holdings - is different.

[Chamaco]

A set of questions here, if you use Ace = Att, K= Count.

 

1. what do you lead for attitude from KQxx(x) ?

 

2. if you lead Q for attitude from KQxx(x), what do you lead for attitude from QJ9x(x) ?

 

3. Does the scheme fit well with the common convention lead "Jack denies, 10/9 promises (or top of nothing)" ?

 

 

Thanks a lot !

 

Mauro

[Codo]

It works much better for me the other way round: The Ace ask for count, the King for attitude.

 

These are the upsites:

 

1. No need to change something when leading from KQxxx

 

2. The most frequent situations for needing count are:

a. You have a suit which is running, if pd has enough cards

b. You have to take the right number of tricks in a suit before declarer can ruff

c. You want pd to ruff.

 

In nearly all this cases, you have at least AK(x), so you are able to take the King or the ace for count asking.

 

The downside is: You cannot ask for count with KQJ(xx). I think this is the lesser evil compared with the problems you get when you use the King for Count.

[lowerline]

I have always thought that "ace for attitude, king for count" is an agreement for defending NT contracts, where "king for count" actually means: unblock the queen or give count...

 

Against suit contracts we always signal attitude on an ace or king lead. There are two exceptions I can think of: 1. signal count with a doubleton on an ace lead 2. signal count on a king lead against six

[ArcLight]

In Kantars Modern Bridge Defense and Woolseys Partnership defense the authors suggest that against No Trump contracts:

 

The lead of the Ace asks pard to unblock any honors, else signal count.

This is typically made from near solid suits like AKJT6.

 

They also advise that the lead of the Q asks pard to unblock the Jack.

Thus you lead Q from KQT94.

[david_c]

A set of questions here, if you use Ace = Att, K= Count.

 

1. what do you lead for attitude from KQxx(x) ?

The way most people seem to play it, you don't have a choice with KQxx(x) - you always lead the K. If that means you aren't getting the signal you want, then that's just hard luck. It's only when you have AK that you can choose.

[sfbp]

There is another way, Frances. Mikeh seems to allude to it indirectly.

 

The opening lead is special (because Dummy is not revealed) and your post appears to acknowledge that. However many years ago, Vinje (I think) came up with the concept that Dummy's holding can be used to determine the a priori flavour of partner's signal.

 

Now the use of A and K by the leader is to show specific honour combinations, which is what mikeh alludes to.

 

After all, how often is it when we are sure what information we want? It's like the difference between blackwood and cue bids - telling partner our hand is more efficient most of the time, because he has choices and is not robotically answering. The discussion about the VERY SLOW play of a card highlights this.

 

There will be cases (in this scheme I describe) where leader's partner has to think, but these will be legitimate and the leader is unlikely to get any unfair information, in my experience.

 

Here's the idea: we make a lead that this quite specific to our holding. Indeed with dummy exposed and his own hand the signaller can usually tell what our holding is.

 

Partner decides how to signal based on dummy's holding in the led suit, roughly as follows (assuming that at the moment he signals, leader is winning the trick - partnerships can still have a different rule about whether they signal attitude or count when dummy wins the first trick):

 

1. If dummy has 0,1 or 2 cards, he signals attitude

2. If dummy has 3 or more cards, he signals count.

3. If dummy is clearly winning the second trick, or declarer can from the auction have no more than a singleton (at suit contract), he signals suit preference (this is basically standard).

 

This seems to work quite well. There's more, to do with the lead of the King which shows a specific holding, and requests unblock in certain situations, but that's a topic for another day.

 

I actually use this (as hinted above) against both suit and NT contracts, though the meaning of attitude will of course be quite different! The poster child for using count signals against NT is as follows:

 

You lead A from AKxx, dummy holds xxx. Partner shows even count if he has 2 or 4. As long as you know declarer cannot have 4 then you know to play the K next. This wins because if partner has Jxxx then declarer had Qx. Of course if he shows odd count, then declarer has 1 or 3, and you play a small one next. Under standard signalling partner is unclear whether to signal when he holds the J.

 

On the other hand if dummy has xx, partner now signals attitude, and "normal" signalling is used.

 

There will be cases where both partners needs to think about the meaning, but having the agreement that in principle we know what we are signalling for seemes to work pretty well. Try it with your favourite partner for 2 weeks and keep track of how often the a priori meaning assigned by this rule is actually wrong. Failing that, do it anyway and see how many times in the next 2 weeks partner yells at you - my guess is it will not be often. The boot is on the other foot usually, partner will say to himself "I wish I could be sure that partner is signalling <whatever>".

 

To recap what I am saying is, the default meaning under this rule turns out to be the one you want most of the time.

[Finch]

A couple of comments about your long post:

 

I specifically said I was only talking about suit contracts. Not only do I play something different against NT contracts, but it varies by partnership.

 

Of course there's another way: there are lots of other ways. I said as much. You are describing a different (and slightly more complicated) rule-based approach. I started this post because someone mentioned they hadn't heard of the idea of A/Attitude K/Count. I'm not particularly trying to sell it, just explain why it has some merit.

 

That said, I'm not going to take your suggestion and try out something different for two weeks, for two reasons. The first is that the success leading/signalling methods depends IMO 90% on partnership harmony and about 10% on method. I have partnership harmony at present and it is very, very, very rare that we have an accident that couldn't have been avoided by the opening leader applying brain before leading. Any method will have particular holdings for which it fails; the more complicated methods will have fewer (look at Meckwell's notes on signalling if you want complicated!).

 

The other reason is that I'm not playing any non-serious-competition bridge for, erm, months.

[sfbp]

By 'You' I didnt specifically mean "You, Frances". I'm sure that whatever methods you use, they are well-honed :)

 

Sorry for the unintended slight to you (and your partner's) competence.

 

Stephen

[Finch]

No problem.

Note I carefully didn't say we never screwed up, only that our screw-ups aren't related to the methods we use....

[mikeh]

I am still not convinced :D

 

Please note that I am NOT speaking about leads against notrump or high-level contracts, but against suit partials and 4-level games.

 

It seems to me that leading either the A or the K from AK, depending on the signal desired, is going to create terrible problems on routine hands.

 

Thus after [1♠] [2♠] (as the one of the simplest opposition sequences), let's say your partner leads the K♦. Dummy hits with xxx in the suit.

 

 

You hold Jxxx

 

Is partner leading the K from AK, wanting to know if you hold the Q? If so, declarer has Qxx and you need to discourage. Or is he leading from KQx, in which case you need to encourage.

 

Now, I appreciate that there are other, subtle, issues at play, including the potential for declarer pitching losers if we don't establish our tricks in time, and the harm that may arise from an inopportune switch. But, within the confines of the opening lead suit, surely this lack of clarity about the leader's holding is dangerous?

 

And matters are made worse, not better, when we 'solve' the problem by leading the Q from KQx.... at least, they are made worse if we also lead the Q from QJ10xx. Thus on the given auction, change the lead to the Q and partner's holding to the Axxx.

 

If leader holds KQx, and we overtake with the A, we have established a power trick for opener's J10x opposite xxx... not best :) . If leader held QJ10xx and we duck, declarer wins his stiff K.

 

And I simply fail to see that the upside of this approach outweighs the downside. This thread arose out of an 'assign the blame' post re defence, but, as I hope my post showed, the error was an unforced error by opening leader, unrelated to any true ambiguity in signalling methods. In other words, this K for attitude, A for count (or reverse) duality was not needed to avoid the posted defensive disaster.

 

Note that switching to rusinow solves the ambiguity issue, but at the cost of losing the flexibility of being able to ask for count or attitude...

 

Having said all of this, I confess that the idea still intrigues me. Maybe I will play against Frances in Verona next month and the lead will come up.....

[Finch]

I am still not convinced :)

 

Please note that I am NOT speaking about leads against notrump or high-level contracts, but against suit partials and 4-level games.

 

It seems to me that leading either the A or the K from AK, depending on the signal desired, is going to create terrible problems on routine hands.

 

Thus after [1♠] [2♠] (as the one of the simplest opposition sequences), let's say your partner leads the K♦. Dummy hits with xxx in the suit.

 

 

You hold Jxxx

 

Is partner leading the K from AK, wanting to know if you hold the Q? If so, declarer has Qxx and you need to discourage. Or is he leading from KQx, in which case you need to encourage.

The method does have issues, but this isn't one of them!

 

If partner has the AK and wants to know if you have the Queen, he would have led the A from AK, asking for attitude.

 

In fact, he's led the king and you give him count.

 

It's true there's a potential problem when he has led from KQx(x), dummy has xxx and declarer has either Axx(x) or AJx(x). This is the biggest problem with the method, I absolutely agree (although much of the time count will give sufficient information anyway).

 

I've got too much work to do to come up with all the sample hands where you gain from being able to select which to ask for. Most of them are soluble with a different set of rules based on dummy's holding - that's absolutely true. But there are also cases where it's not clear from partner's hand what you want, and it is clear from yours.

[mike777]

Obvious Shift "OS" play:

1) Versus a slam on the lead of A or K we give count.

2) Versus the five level, when we are known to hold 5 vards or more in the suit led we add to our Attitude/Obvious shift signal a count card by playing an odd card for an odd number or an even card for an even number.

3) On the lead of a trump we give suit preference.

 

In almost all other cases attitude is shown at trick one, this includes stiff and voids in dummy on opening lead. Yes partner can figure out what to shift to with an "OS" attitude signal from us and not suit pref. :). I note many experts prefer to play all three signals at trick one, I expect partner to figure out what I am telling her.

 

The main argument for all of the above is to solve 99.9% of the issues at the table not 100% with a set of easy to remember rules.

[Walddk]

It's true there's a potential problem when he has led from KQx(x), dummy has xxx and declarer has either Axx(x) or AJx(x). This is the biggest problem with the method, I absolutely agree

That is enough evidence for me to label the method "flawed". MikeH has come up with an example, and here is another one. Partner leads the queen, asking for attitude in your method, and dummy tables

 

xxx

 

You hold 10xx. Now, if the lead is from QJ, you want to encourage. On the other hand, if it's from KQ empty, you must discourage in order to prevent partner from continuing the suit into declarer's AJx (Bath Coup) when he ducks at trick 1.

 

With this said, I'd like to add that the method is fine against 5- and 6-level contracts, especially in competitive auctions. Ace for attitude, king for count. The king will often be the right card because you want to know how many tricks you can cash in that suit. Then count is essential and attitude useless.

 

It hurts, but I actually agree with Mike when he says that generally speaking this method will create problems on routine hands. Easy for me to say if I hadn't tried it out at the table, but I have. I had a Scottish partner for 3 years, and she insisted on A for attitude and K for count always.

 

We did face problems that would not otherwise have occurred if we had subscribed to "standard" methods.

 

Roland

[david_c]

People still seem to misunderstand what this method is all about.

 

First of all, it is based around the "standard" method of leading top from sequences. So from KQ empty the lead is the king. Not the queen! You only lead the queen from this holding if you are prepared to risk the consequences of partner playing you for QJ.

 

Secondly, no matter what your leading style is, you will have to decide,

 

- What is our primary signal on the lead of the ace?

- What is our primary signal on the lead of the king?

 

Frances said in her original post that on balance it seems more useful to get count on a king lead (assuming we always lead the K from KQ). Of course, once you've made that decision, there are going to be some hands where it is not successful. The lead from KQx(x) when declarer might have Axx or AJx is one such problem. But once you've decided to show count you can't get away from that. This problem is not related to what you choose to lead from AK.

 

But once you've decided that you generally want attitude on an ace lead and count on a king lead, then it makes sense to vary your lead from AK if you think you know what signal you want.

[Walddk]

People still seem to misunderstand what this method is all about.

 

First of all, it is based around the "standard" method of leading top from sequences. So from KQ empty the lead is the king. Not the queen!

That does not solve your problem. Who has the ace? If partner has it, you will likely want to continue the suit, whereas if declarer has it you run the risk of leading into his remaining AJ tenace.

 

No matter what, it's a guess. Allow me to point out that if you play standard methods where ace promises the king and king shows the queen but not the ace, it will make life so much easier for yourself and not least your partner.

 

Roland

[1eyedjack]

It works much better for me the other way round: The Ace ask for count, the King for attitude.

 

These are the upsites:

 

1.  No need to change something when leading from KQxxx

 

2. The most frequent situations for needing count are:

    a. You have a suit which is running, if pd has enough cards

    b. You have to take the right number of tricks in a suit before declarer can ruff

    c. You want pd to ruff.

 

In nearly all this cases, you have at least AK(x), so you are able to take the King or the ace for count asking.

 

The downside is: You cannot ask for count with KQJ(xx). I think this is the lesser evil compared with the problems you get when you use the King for Count.

I didn't entirely understand the arguments in this post, but I would add another "downside" to swapping. You might want to lead an unsupported Ace. It is also I suppose possible but far less likely that you would want to lead an unsupported King. If you have an unsupported Ace and wish to lead it, an attitude signal is most likely going to be more useful to you than count.

[cherdano]

It works much better for me the other way round: The Ace ask for count, the King for attitude.

 

These are the upsites:

 

1. No need to change something when leading from KQxxx

 

2. The most frequent situations for needing count are:

a. You have a suit which is running, if pd has enough cards

b. You have to take the right number of tricks in a suit before declarer can ruff

c. You want pd to ruff.

 

In nearly all this cases, you have at least AK(x), so you are able to take the King or the ace for count asking.

 

The downside is: You cannot ask for count with KQJ(xx). I think this is the lesser evil compared with the problems you get when you use the King for Count.

Isn't the biggest downside that when the K lead can show AK.. or KQ..., partner does not know what to signal with the J?

 

Arend

[david_c]

People still seem to misunderstand what this method is all about.

 

First of all, it is based around the "standard" method of leading top from sequences. So from KQ empty the lead is the king. Not the queen!

That does not solve your problem. Who has the ace? If partner has it, you will likely want to continue the suit, whereas if declarer has it you run the risk of leading into his remaining AJ tenace.

 

No matter what, it's a guess. Allow me to point out that if you play standard methods where ace promises the king and king shows the queen but not the ace, it will make life so much easier for yourself and not least your partner.

I find it hard to believe that you could write this if you had actually read the whole of my post. I will repeat it here for convenience.

 

Secondly, no matter what your leading style is, you will have to decide,

 

- What is our primary signal on the lead of the ace?

- What is our primary signal on the lead of the king?

 

Frances said in her original post that on balance it seems more useful to get count on a king lead (assuming we always lead the K from KQ). Of course, once you've made that decision, there are going to be some hands where it is not successful. The lead from KQx(x) when declarer might have Axx or AJx is one such problem. But once you've decided to show count you can't get away from that. This problem is not related to what you choose to lead from AK.

 

But once you've decided that you generally want attitude on an ace lead and count on a king lead, then it makes sense to vary your lead from AK if you think you know what signal you want.

[Walddk]

People still seem to misunderstand what this method is all about.

 

First of all, it is based around the "standard" method of leading top from sequences. So from KQ empty the lead is the king. Not the queen!

That does not solve your problem. Who has the ace? If partner has it, you will likely want to continue the suit, whereas if declarer has it you run the risk of leading into his remaining AJ tenace.

 

No matter what, it's a guess. Allow me to point out that if you play standard methods where ace promises the king and king shows the queen but not the ace, it will make life so much easier for yourself and not least your partner.

I find it hard to believe that you could write this if you had actually read the whole of my post. I will repeat it here for convenience.

 

Secondly, no matter what your leading style is, you will have to decide,

 

- What is our primary signal on the lead of the ace?

- What is our primary signal on the lead of the king?

 

Frances said in her original post that on balance it seems more useful to get count on a king lead (assuming we always lead the K from KQ). Of course, once you've made that decision, there are going to be some hands where it is not successful. The lead from KQx(x) when declarer might have Axx or AJx is one such problem. But once you've decided to show count you can't get away from that. This problem is not related to what you choose to lead from AK.

 

But once you've decided that you generally want attitude on an ace lead and count on a king lead, then it makes sense to vary your lead from AK if you think you know what signal you want.

I did read the whole of your post, and I do not find it convincing. This is a partnership game, so it's often not enough when one player decides what's right and wrong.

 

If you find the method appealing, fine with me. As I have pointed out in an earlier post, I know it's popular in Britain, but that does not change the fact that I think it's flawed. That would also be my view if I moved to Southampton or Glasgow.

 

It looks pretty silly when you get it wrong and either don't continue the suit when it's right and continue when it's wrong because you can't read your partner's signal. If you have never come across that situation, good for you.

 

Roland

[Codo]

It works much better for me the other way round: The Ace ask for count, the King for attitude.

 

These are the upsites:

 

1.  No need to change something when leading from KQxxx

 

2. The most frequent situations for needing count are:

    a. You have a suit which is running, if pd has enough cards

    b. You have to take the right number of tricks in a suit before declarer can ruff

    c. You want pd to ruff.

 

In nearly all this cases, you have at least AK(x), so you are able to take the King or the ace for count asking.

 

The downside is: You cannot ask for count with KQJ(xx). I think this is the lesser evil compared with the problems you get when you use the King for Count.

Isn't the biggest downside that when the K lead can show AK.. or KQ..., partner does not know what to signal with the J?

 

Arend

Not really.

 

To loose a trick, you need some quite specific holdings.

If Pd has just Jxx and declarer QTx, declarer will always succeed, no matter what you signal or lead at your first trick, even if you lead another suit.

If pd has JT9x, he may drop the jack, so everything is fine again.

If declarer has Qxx opps. Txx the suit is frozen anyway and you cannot get more then two tricks if you have to play the suit by yourself.

 

So, you only have the problem when pd has JTx and declarer Qxx.

 

This is a problem, but I don´t see this often at the table.

 

At least it seems to be a much less frequent then the problem while getting count signals leading from KQTx....

[Finch]

This discussion is interesting, because I am coming to understand that the debate has come down to what signal should you give on a king lead. 'Ace for attitude king for count' evolved from the idea that you are most likely to want count on a king lead. All the objections I've seen in this thread revolve around cases when you'd prefer to have attitude (exactly one case, though it's not uncommon).

 

Now, if you think you'd like to get (in principle) an attitude signal on every honour lead, then you should stick to standard leads because - as many people have pointed out - partner then knows what you have and can signal accordingly. Same applies if you always want count (though that seems to be extremely rare among top partnerships).

 

If you sometimes like attitude and sometimes count then either you need some rules based on dummy's holding, or some rules based on the lead (possibly with additional rule(s) based on level or type of contract e.g. count at 5-level+, count against pre-empts).

 

I think always giving attitude is also flawed, although it does have the advantage that you can decide in the context of your hand what your attitude is rather than just looking at the suit led.

 

A complicated set of rules based on dummy may be better, if you don't mind the complication. And even then there will be times when it's flawed.

 

The simplest example is with xxx in dummy. If you've led from KQJ(x)(x) you want count. From AKQ(x)(x) you want count. From KQ(x)(x) you want attitude (and ideally you want to know if it's the ace or the jack). From AKx(x) you want attitude (encourage with either a doubleton or the queen). From AKJx you want attitude on the queen only, particularly if you might want to underlead or partner would be ruffing with a trump trick. From AKxxx you ideally want to distinguish between Qx, xx and any 3-card holding. From QJxx you want attitude. From QJ10x you probably want count or unblock. From a speculative ace-from-length lead (AQxxxx) you want attitude on the King, not the queen. I could continue, but you get the idea.

 

Of course, any method that isn't "tell me what I want to know" will be flawed in some circumstances. And "tell me what I want to know" is flawed due to lack of telepathy.

 

I had a quick look at some of the CCs from the last Bermuda Bowl, and all I can say is that they aren't very well filled in! Apart from the English pairs who all indeed play K from either AK or KQ for count, the majority say they give attitude as primary signal, count as secondary signal on an honour lead, but the majority do not define when, if ever, they give count. The only one I saw that does is Fantoni/Nunes who I'm pleased to say give attitude except on a king lead when they give count, but I'm sure other pairs have other agreements.

[Flame]

I never liked A attitude K clunt although its very common here.

How do you protect yourself from declarer ducking with AJX ?

I would consider playing it if we playing Q from KQ J from QJ.

[han]

With Arend I play ace-for-attitude and king-for-count in combination with Rusinow. We lead the king only from very strong holdings where we want partner to give count (or unblock). From AKx we lead the ace, but from AKJ10x we lead the king. From KQx we lead the queen, but from KQJ10x we lead the king.

 

The lead of the ace, queen or jack is never ambiguous. The lead of the king is ambiguous but it is only lead from a very strong holding so partner doesn't have a problem. I don't think that this has any of the drawbacks mentioned here.

[cherdano]

It works much better for me the other way round: The Ace ask for count, the King for attitude.

 

(...)

 

The downside is: You cannot ask for count with KQJ(xx). I think this is the lesser evil compared with the problems you get when you use the King for Count.

Isn't the biggest downside that when the K lead can show AK.. or KQ..., partner does not know what to signal with the J?

 

Arend

Not really.

 

To loose a trick, you need some quite specific holdings.

If Pd has just Jxx and declarer QTx, declarer will always succeed, no matter what you signal or lead at your first trick, even if you lead another suit.

If pd has JT9x, he may drop the jack, so everything is fine again.

If declarer has Qxx opps. Txx the suit is frozen anyway and you cannot get more then two tricks if you have to play the suit by yourself.

 

So, you only have the problem when pd has JTx and declarer Qxx.

 

This is a problem, but I don´t see this often at the table.

 

At least it seems to be a much less frequent then the problem while getting count signals leading from KQTx....

Well there are more problems. Either you encourage with a low doubleton on a king lead and risk the bath coup, or you don't and miss out on a ruff if partner has led from AK (hoping for the queen). Also, even if you don't lose a trick in the suit by continuing (in your Jxx vs QTx example), you are losing the lead which may cost you more than one trick.

 

For some reason the first leading style I learned was standard honor leads but K from AK (A shows unsupported ace). Of course it doesn't make sense, but I don't think the "was it KQ or AK"-problem was that infrequent.