[kenrexford]

So many concepts here, but let's start:

1. Why is cover card analysis best?

Art had ♠KQJxx ♥xx ♦AKJx ♣Qx. Let us compare two possible hands for partner that are remarkably similar, with the exact same HCP count.

♠Axx ♥Qxx ♦xxx ♣Axxx

♠Axx ♥xxx ♦Qxx ♣Axxx

Obviously, a simple playing out of the hands in your head will tell you that the first hand has about a 62.5% chance of making 10 tricks, while the second has about a 100% chance of making 10 tricks. Thus, about 37.5% of a trick tuns not on the HCP count but rather on where the Queen is. If you happen to translate a "trick" into say 2.6 HCP (26 HCP divided by 10 tricks), you end up with the knowledge of whether the Queen has value of roughly 1 HCP.. But, you have a much higher impact on expectations, IMO.

Now, change the club to the King. Nothing changes in the analysis -- you still end up with the same chances of making, roughly. So, a side Ace or King in some situations carries full weight regardless of whether it is an Ace or a King. If you can enable partner to know that the King is in a solid place, in a sense, then 1 HCP of nuance is immaterial.

What about changing the red Queen into the heart Jack, but keeping the club Ace? While the Club-King-Heart-Queen combo and the Heart-Jack-Club Ace combo are both 5 HCP, the former has a 62.5% chance of game while the latter has only a straight 50% chance of game, but actually less than. Adding the club Ace instead of the club King gains nothing in trick-taking despite an extra HCP.

When 1-2 HCP can emerge or disappear with no consequence and where about 1 HCP (and extra roughly 33% chance of making) can emerge solely from location of honors rather than number or type of honors, this seems rather significant, moreso than fine-tuning the total unadjusted HCP count. While this hand is admittedly not the best example of that principle (which is why I advocated a 3♣ call because of the Q-x), the principle is stronger when the hand pattern includes a stiff. System design should not be catered to the 5422 hand, IMO, when the 5431 or 5521 or the like are in total more common.

2. Does a wide range affect accuracy?

Of course. That is why I claimed that the approach seems unworkable unless high cover-card potential hands at the top are adjusted out of the range. This is why I said that you have to be "conservative" in game tries, because the hand at the top of the range that you would need for a successful game try cannot logically exist unless the entire approach is nonsense.

3. Can you in some way "invite" aggressively in this context?

Sure, somewhat. As you suggested, you may not get the opportunity to play 2M in the real world. Opponents compete over 2M because of the LTT. OK, well then you can and should use that against then, IMO. On the very marginal hands, you have the opportunity to pass but then move toward game if the auction is in fact reopened. For example, what about this possible auction:

1♠-P-2♠-P-

P-X-3♣-P-

3♦-P-4♠-all pass

While the "I was going to stop at 2♠ but just got backed into a making game" approach is not ideal, perhaps, this option has several important principles. First, by having this approach as possible, you might have a tendency to have people sell out to 2M more often, to your benefit. On the other hand, not making aggressive game tries does not always translate into missing game. Thus, If the cost of not making an aggressive game try seems like a X value loss, the actual loss is X/y, which is lower. (Plus, many of the "aggressive" game tries are unbalanced hands, where balancing is more likely and where the balancing call selected may help both in playing the tight game, increasing your odds of making perhaps, and in evaluation of whether your two hands fit well contextually, both of which seem like good game try inputs.)

4. Will passing miss game more than bidding will cost a set?

As I just mentioned, passing only misses a game that is there when LHO does not balance, which reduces that figure somewhat. Equally, when LHO does balance, our chances of making the end result game probably increase, as does our chance of reliably bidding a game, in many cases. We are, again, talking of the marginal game tries only, however. Against my reasoning is the reality that on many hands where the marginal game try results in a declining of the invite and a set, we may well have taken away the opposition partscore and may have discouraged what might otherwise have been a lucrative double by them.

But, the discussion was primarily about methods to handle the multi-faceted game tries and some form of stacking of ranges. Adding in marginal game tries also has the negative impact of reducing descriptive options for all ranges. If you need to show three different game try base holdings, you lose the kind of detail you could have as to two ranges. I other words, I my approach, I can make a try of some descriptive type and then hear back a nuanced "in the vicinity" call with additional nuance. In the three-way approach, you end up dropping one call to use that for the "I barely have a game try" hands, costing nuance on the other ranges. When you also toss in the issue of wanting to occasionally make slam tries, the entire structure collapses into essentially HCP quantitative game tries without any effective slam tries. Thus, IMO, one cost you do not seem to give full weight to is the loss of definition with the more sane game try scenarios.

[Zelandakh]

I honestly do not understand your last argument Ken. If you play 2NT in the way suggested then you have, as Andy wrote, 5 ranges below game:

1. Pass
2. 2NT and then 3♠ if partner makes a game try
3. 2NT and then 4♠ if partner makes a game try
4. Game try and then 3♠ if partner makes a return game try
5. Game try and then 4♠ if partner makes a return game try

In addition to this you also have ranges for game (4♠) and for slam tries (assuming partner's acceptance of a game try allows for this). And it is not like you are giving up on the benefits of suit-based game tries either. In both cases this information can be exchanged. What you do lose is the ability to use two-way game tries or one of the other conventional uses for 2NT here as well as 1-2-3 stop (if you want to be able to split a heart-based game try). As others have pointed out you are also committing to the 3 level on some hands where another system would not. So you do lose something but given that the simple raise is being played as described you are probably losing less this way than keeping to a more traditional structure.

[kenrexford]

Zelandakh, on 2013-December-11, 08:24, said:

I honestly do not understand your last argument Ken. If you play 2NT in the way suggested then you have, as Andy wrote, 5 ranges below game:

1. Pass
2. 2NT and then 3♠ if partner makes a game try
3. 2NT and then 4♠ if partner makes a game try
4. Game try and then 3♠ if partner makes a return game try
5. Game try and then 4♠ if partner makes a return game try

In addition to this you also have ranges for game (4♠) and for slam tries (assuming partner's acceptance of a game try allows for this). And it is not like you are giving up on the benefits of suit-based game tries either. In both cases this information can be exchanged. What you do lose is the ability to use two-way game tries or one of the other conventional uses for 2NT here as well as 1-2-3 stop (if you want to be able to split a heart-based game try). As others have pointed out you are also committing to the 3 level on some hands where another system would not. So you do lose something but given that the simple raise is being played as described you are probably losing less this way than keeping to a more traditional structure.

The best way to explain is to provide an alternative structure and show the difference. Spades are assumed as trumps.

Suppose Opener has a heart-spade two-suiter, 5521 or 5512. If he rebids 3♥ as the game try, Responder knows that the A/K/Q of both spades and hearts are working cards. He also suspects fairly reliably that the minor Aces carry full weight. A minor King is of uncertain value, and a minor Queen of even more uncertain value. However, he has no ability to in any way bid these cards to "ask" if they have value. Additionally, he cannot distinguish between "I hate that game try," "I sort of like that game try," and "I love that game try." If, however, Opener were able to bid 2NT to show the "most preempted natural game try," this allows Responder to bid values in clubs or diamonds. Responder can also bid hearts to show perhaps a no-minor-cards minimal acceptance (maybe only two covers concentrated in the majors). If Responder shows a dubious value in either minor, Opener can "game last train" back with a repeat of hearts, and you also get a nuance of a 3♦ call after 3♣. If Opener happens to be slammish, you have gained another level of values description, which has to help. Thus, the simple switch of 2NT and 3♥ obviously seems to have a great impact on the effectiveness of a major two-suiter game try.

If the second suit bid is clubs, you are in the same situation.

If the second suit is diamonds, only a 3♥ try-back is available. However, by freeing up the 3♥ call as not for hearts, you can have two ways to show diamonds.

So, first, there is a cost to using 2NT as a specific range game try to add nuance to the strength of the game tries, namely that you lose definition on the alternative game tries.

Plus, as you can see, the nuances of "strength" can pop up after the suit-based game tries, if you want. But, the next question then is whether ranges or honor location is more important. If the concept is ranges only, to get three "ranges," you lose the ability to structure definition as to side honors.

Consider again the above structure when Opener wants to focus diamonds. As I suggested, there are three ways to bid: 3♦ and accept a 3♥ try back, 3♦ but decline, and 3♥. If these are ranges of HCP's only, you lose the ability to effectively use this start to pursue slam (and hence lose nuance to jumps as alternative slam approaches). If you use these calls as tagged to side tertiaries, however, you get more pattern development. For example:

3♥ = Diamond-oriented game try with interest in heart values.

3♦ = Diamond-oriented game try with either interest in CLUB values (3♥ by Responder would show club values to check) or no interest in either side suit (would decline 3♥).

Thus, Opener can somewhat nuance through this the difference between (1) 5-3-5-0 or 5-3-4-1, (2) 5-0-5-3 or 5-1-4-3, or (3) 6-1-5-1 as archetype game tries. (Opener would decide how to treat 5521 or 5422 types with judgment. For instance, if one of the two doubletons has the Queen, as Art had, that information might help determine which suit to entice tertiary support considerations, or strength of invite might determine whether to discourage consideration of a side King.)

Moreover, the entire concept of dedicating a call to a specific weak type of game try is dumb already. If 2NT is, for example, simply a heart-specific game try, Responder has four responses for range, plus Opener can try back three different ways. Thus, 2NT alone allows 6 different ranges of strength-based ranges. 3♣, if club-centric, allows 3 responses, with one allowing a try-back, for 4 total ranges. If 3♦ and 3♥ and heart-centric, you have three ranges you can handle. Thus, if you want three ranges and also suit-based game tries, a better scheme would be:

3♦/3♥ = three ranges of diamond-centric game tries

3♣ = 4-way club-centric game try

2NT = 3-way heart-centric game try OR 3-way no-suit-centric game try. If Responder has no heart-centric help, he bids 3♦ and then declines, 3♦ and then accepts, or 3♥. With no help no matter what, 3♠. If heart-centric help, 3♣, which allows Opener to bid 3♦ and accept, 3♦ and decline, or 3♥, whether he has hearts or not.




The structure proposed, therefore, is not efficient anyway. Plus, it fails utterly to account for secondary and tertiary side values, forcing a two-dimensional pattern analysis rather than a 4-dimensional pattern analysis.

[Zelandakh]

kenrexford, on 2013-December-11, 09:06, said:

If the concept is ranges only, to get three "ranges," you lose the ability to structure definition as to side honors.

The proposed structure is 2 invitational ranges with suit-specific game tries within each range where space allows. For the strong invite the suit for the game try comes from Opener; for the weak invite the suit comes from Responder. Then partner of the player making the suit-specific game try can make a return game try. In the case of a strong game try in hearts, Opener would bid 3♥ with the upper end and 3♠ with the lower end. You are right in as much as some sequences over 2NT (most specifically 1♠ - 2♠; 2NT - 3♣; 3♦ - 3♥) are not optimised. It would therefore probably be possible to make some improvements but my assessment is that the OP would not want the level of extra memory load and complexity that that would entail. You know I love theory but sometimes practical has to come into the equation too.

[kenrexford]

Zelandakh, on 2013-December-11, 09:24, said:

The proposed structure is 2 invitational ranges with suit-specific game tries within each range where space allows. For the strong invite the suit for the game try comes from Opener; for the weak invite the suit comes from Responder. Then partner of the player making the suit-specific game try can make a return game try. In the case of a strong game try in hearts, Opener would bid 3♥ with the upper end and 3♠ with the lower end. You are right in as much as some sequences over 2NT (most specifically 1♠ - 2♠; 2NT - 3♣; 3♦ - 3♥) are not optimised. It would therefore probably be possible to make some improvements but my assessment is that the OP would not want the level of extra memory load and complexity that that would entail. You know I love theory but sometimes practical has to come into the equation too.

How exactly does the fact that Opener has a weak invite create a side suit for Responder to show?

Now, I do personally like one aspect of this approach IF Responder "shows suits" empathetically. In other words, a 3♣ reply to 2NT as "I would accept a weak 3♣ game try" rather than "I have my own side suit in clubs." I like that not so much because of it making sense but rather because I have a fondness for this sort of theoretical approach. I dabbled for a while in a concept of cuebids of the dubious value, where a cuebid showed precisely where the dubious value was, with other suits either holding clear values or no values. As a parallel to this, then, Responder could bid, say, 3♣ with the trump King, a side Ace in either hearts or diamonds, and the club Queen -- bidding the tertiary value. Similarly, Responder would bypass 3♣ to bid 3♦ with, say, ♠Kxx ♥xxxx ♦Qxx ♣Axx, because he is identifying the dubious value.

[PhilKing]

kenrexford, on 2013-December-11, 10:38, said:

How exactly does the fact that Opener has a weak invite create a side suit for Responder to show?

Now, I do personally like one aspect of this approach IF Responder "shows suits" empathetically. In other words, a 3♣ reply to 2NT as "I would accept a weak 3♣ game try" rather than "I have my own side suit in clubs." I like that not so much because of it making sense but rather because I have a fondness for this sort of theoretical approach. I dabbled for a while in a concept of cuebids of the dubious value, where a cuebid showed precisely where the dubious value was, with other suits either holding clear values or no values. As a parallel to this, then, Responder could bid, say, 3♣ with the trump King, a side Ace in either hearts or diamonds, and the club Queen -- bidding the tertiary value. Similarly, Responder would bypass 3♣ to bid 3♦ with, say, ♠Kxx ♥xxxx ♦Qxx ♣Axx, because he is identifying the dubious value.

Virtually all top players are going in the "dummy shows" direction these days, creating more sequences where declarer does not show his hand at all. This creates an advantage in the card play that massively outweighs those texbook hands where we show our KJxx to the world and congratulate ourselves on our nuanced approach when we reach a thin game opposite QTx.

Your approach to this situation is fast becoming obselete.

[kenrexford]

PhilKing, on 2013-December-11, 12:04, said:

Virtually all top players are going in the "dummy shows" direction these days, creating more sequences where declarer does not show his hand at all. This creates an advantage in the card play that massively outweighs those texbook hands where we show our KJxx to the world and congratulate ourselves on our nuanced approach when we reach a thin game opposite QTx.

Your approach to this situation is fast becoming obselete.

Fair point, and one I actually employ most of the time myself. I also use one call in this sequence as a "random game try" and usually bid that unless I am a more traditional 5-5 or 6-4 for the natural game try.

That said, the discussion has suggested a three-way random game try, which seems like taking a good thing way too far. I would probably bid the random game try with Art's hand, actually, but that is stylistic/judgment.

[Vampyr]

Where I play, an invitational hand with 3 trumps normally starts with 2/1, then jumps or raises the opener's suit to the 3-level. Using this approach, you would be able to take the better hands out of the single raise, with the responder showing where she has values as well.

[Trinidad]

Vampyr, on 2013-December-11, 12:28, said:

Where I play, an invitational hand with 3 trumps normally starts with 2/1, then jumps or raises the opener's suit to the 3-level. Using this approach, you would be able to take the better hands out of the single raise, with the responder showing where she has values as well.

Of course, that is what we all do in one way or another. But the OP doesn't want to be at the three level with an invitational hand opposite a minimum opener (which can be a 10 HCP hand).

He plugs this leak in his system by creating a leak somewhere else: a 7 point range for the single raise. This, in turn, means that opener has to be aggressive in his invitations since responder may have an invitation (or even a hand that would be a minimum game force for many of us). The consequence is that he is forced to the three level on many hands where the field plays at the two level.

Rik

[ArtK78]

As we open light only when nonvul, it is unlikely that anyone would allow us to play in 2 of a major. So it is very rare that we will be higher than necessary.

Also, part of the purpose of the wide-ranging single raise is obstruction. If we open light and respond light, for example, the auction is up to the 2M level on much less than half the deck in high cards. This kind of cooperative preempting is very useful, and makes up for some of the inaccuracies of constructive auctions.

[mycroft]

If that's the case, then if you expect, on "half the cards" to be at the 3 level anyway, then a two-stage (or more-stage) game try system is fine; if you make a "I need a max to GT" call and partner has a minimum, yeah, you're on 14 vs 6, but you were going to be there anyway; if you make a "I have a GT opposite 6-9" call, you were going to be there anyway.

The problem with the original auction is that your partner would accept a "GT opposite 6-9" with that hand *in diamonds*, and you didn't have one. Not only that, but he had no way to guess whether you had a GT I would make, or a "only go with diamond help and a limit raise" (or even "go on a limit raise"). It's not his fault; it's not your fault; you're asking two questions knowing you can only get a single "yes" or "no" answer.

[Vampyr]

How about using a 2♣ response for some range of the 2♠ raise?

[aguahombre]

Vampyr, on 2013-December-12, 19:36, said:

How about using a 2♣ response for some range of the 2♠ raise?

Depending on jurisdiction, you could run afoul of regs regarding Drury by an unpassed hand.

[fromageGB]

Not Drury, asking, but 2♣ telling, and after a 2♦ relay 2M = 11/12. That makes the 1M 2M more manageable.

[Zelandakh]

As has been said, system is not the point of the thread but since my own system is built around light openings and the thread drift is working in this direction I will share my solution. My idea is for all of the good (invite or better) responding hands without 4 card support to bid the first step as a relay and for weaker hands or hands with strong support to use a higher call. This is essentially the reverse of 2/1, where the GF hands show and invite or worse bid the relay (1NT over 1♠ is a relay by another name).

So the complete structure is:

1♥
==
1♠ = invite or better (1NT = min without 4 spades; 2♣ = 4+ spades; others = max without 4 spades)
1NT = weak with 4+ spades (poss canape), NF
2m = weak and natural, NF
2♥ = weak raise, typically 3 card support or 4 and a low ODR
2♠ = invitational raise with shortage (mini-splinter) or a strong splinter (maxi-splinter)
2NT = GF raise
3♣ = limit raise unsuitable for mini-splinter
3♦ = mixed raise
3♥ = preemptive raise
3♠ = splinter with any side void
3NT = splinter with singleton spade
4m = singleton splinter
4♥ = preemptive raise

--
1♠
==
1NT = invite or better (2♣ = min without 4 hearts; 2♦ = 4+ hearts; others = max without 4 hearts)
2m♥ = weak and natural (poss canape), NF
2♠ = weak raise, typically 3 card support or 4 and a low ODR
2NT = invitational raise with shortage (mini-splinter) or a strong splinter (maxi-splinter)
3♣ = GF raise
3♦ = limit raise unsuitable for mini-splinter
3♥ = mixed raise
3♠ = preemptive raise
3NT = splinter with any side void
4m♥ = singleton splinter
4♠ = preemptive raise

The raise structure is thus:

3M/4M = preemptive
2M = weak raise
3M-1 = mixed raise

relay followed by 2M = 3 card limit raise (if partner shows a max then jump to 4M if no slam interest)
3M-2 = 4 card limit raise without a side shortage to show
2M+1 = mini-splinter (also maxi-splinter)

relay followed by 4M = GF raise without slam interest (can be 3 card limit if partner showed a max)
relay folowed by 3M = GF 3 card raise with slam interest (a second relay will often be used instead)
2M+2 = normal GF 4 card raise (note one less step than Jacoby after 1♠ opening)
3M+2 to 4M-1 = singleton splinters with 4+ card support
3M+1 = void splinter (any suit) with 4+ card support (you can reverse these 2 if you like)
2M+1 = maxi-splinter (also mini-splinter)

It is probably not for you and your partner but it perhaps shows that there are ways of working around the issue. A more standard solution would be to play a 2♣ response as "natural or 3 card limit raise or balanced invite". That also takes the top end out of the 2M raise and avoids some of the issues being raised here. I believe Cyberyeti has also got a system designed around very light initial action and perhaps has another solution that might fit to you and your partner. In the long run, I do think you should look to change this part of the system though.

[benlessard]

I believe that showing Q values is more important than showing the 3rd range of pts.


1S-2S
2M+1---should always denies a stiff IMO. You should never do "light" inv with no singleton.
3C = splinter in unknown suit light inv.
3D = D splinter serious inv
3H = H splinter serious inv
3S = C splinter serious inv.

1S-2S-2nt

3C = denies Qc or JTc
3D = show QC or JTc denies D values
3H = QJ values in both m
3S = really crappy hand

[fromageGB]

Zelandakh, on 2013-December-13, 12:02, said:

My idea is for all of the good (invite or better) responding hands without 4 card support to bid the first step as a relay and for weaker hands or hands with strong support to use a higher call. This is essentially the reverse of 2/1 ...

Interesting system, seems workable. Do you call it 1\2 ?

[jogs]

kenrexford, on 2013-December-11, 07:48, said:

So many concepts here, but let's

start:

1. Why is cover card analysis best?

If you happen to translate a "trick" into say 2.6 HCP (26 HCP divided by 10 tricks), you end up with the
knowledge of whether the Queen has value ......

Why is the bridge community still clinging to point count as the main method of hand evaluation? It is
an oversimplication. Point count isn't even a primary vector for estimating tricks.
Treat hand strength as a multidimensional jigsaw puzzle. Each partnership holds two multidimensional
puzzle pieces. When the two partnership pieces fit well more tricks are generated. When they fit poorly
fewer tricks are available. The shapes of the pieces are not rigidly fixed. They are pliable and can often
be molded into additional tricks.
During the auction each partner should attempt to estimate the tricks for the partnership. The
estimated tricks vary depending on the strain. The estimate is the sum of two multidemensional vectors.
The first vector is power or strength. Point count is its primary subvector. The other vector is pattern.
Pattern is the joint suit pattern of the partnership.

Queens in the trump suit are nearly almost always worth a trick. Queens in the primary side suit are
usually worth a trick. Queens in the secondary side suits are rarely worth a trick. Queens definitely do
not have a fixed value.