[straube]

Free, on 2011-April-19, 06:08, said:

It's more like a theoretical exercise. It depends a lot on how 'weak' your strong ♣ can be. The lower your HCP limit (17+HCP is not the same as 14+HCP), the more semipositives you'll get. When you put all double negatives and semipositives in 1♣-1♦ you're overloading this response and you'll be more vulnerable to preemption. That's why it's useful to describe these hands immediately (frequency principle). The rarest types of hands are double negatives, so you can waste some space in that case (again frequency principle). GF hands should stay as low as possible, so 1♣-1♦ is used for that.

The way to do it vary and there have been lengthy discussions about this in the past. I still prefer to use the principle I've just described for the given reasons. It's theoretically sound and the big advantage is that the most frequent hands (semipos) immediately limit their strength AND describe their hand shapewise (Major suit oriented). 1♣-1♥ basically denies 5M and can be a bit of everything (balanced or minor suit oriented), but 1♣-1NT+ specifically show 5M or 6M and sometimes even specific side suits. The tradeoff for all advantages is 1♣-1♠ as double negative, I'll admit. Most important to know here is that it's very infrequent, and it's most important to find some playable part score.

I believe Straube prefers 1 bid for all semipositives. This doesn't have the same advantages like I described, and I also don't see any gains over my approach, which is why we had lengthy discussions...

Using 1♦ as double negative and the rest as semipositives or GF has been discussed as well. Biggest problem here is that both opener and responder need to show somehow that the auction became GF.

Our 1C is probably stronger than yours. It can be a prime unbalanced 15, but our 1N opener is 14-16. Our semipositive is 3-5 QPs, but most of our 5 QP hands (a good 8 pts or so) are allowed a positive response. This means that we have fewer semipositives and more GF hands than you likely do. For 200 hands...

1D-32%
1H-35.5%
1S-25%
1N-2.5%
2C-3%
other-2%

or...
GF-39.5%
semi-35.5%
DN-25%

I was expecting (from other tallies) more like
GF-45%
semi-35%
DN-20%

but that's what you get for only 200 hands.

Really almost tempted to suggest our GF include all 5 QP hands and the semipositives to be 2-4 QPs (with 2 QPs being an optional semipositive). Many of the 2 QP hands have 5 or more hcps. Then maybe...

GF-47%
semi-43%
DN-15%

Had one other idea come to me while tallying hands. If 1C-1S, 2C makes sense as Stayman, perhaps 1C-1H, 2C should also be Stayman.

[wclass___]

1♦=4♥+ semipositive+
1♥=4♠+semipositive+
1♠=DN or bal GF
1N=other semipositives..bal.. minors
2♣+GF

Seems fun. . If you fear competition you must start to bid your suits.

[straube]

wclass___, on 2011-April-20, 02:20, said:

1♦=4♥+ semipositive+
1♥=4♠+semipositive+
1♠=DN or bal GF
1N=other semipositives..bal.. minors
2♣+GF

Seems fun. . If you fear competition you must start to bid your suits.

I think that's awful! The 1S response would occur maybe 50% of the time and the continuations would be a nightmare. It would leave us in a terrible spot for competition. I want to know if we're in a GF!

[wclass___]

According to your bridge logic if there were only 3 shapes..e.g.: 4333; 6322;6430. With corresponding probabilities: 60%; 32%;8% and permutations: 4;12;24. It would be poor bridge to play anything but 1♣-1♦ as 4333; 1♥ as 6322 and 1♠ as 6430...

As i understand none of your 1♦/1♥/1♠ responses tells anything about suits...
In uncontested 1♣-1♦(negative) i expect way better results. Only mayor advantage should come from competitive bidding, but i can't imagine that knowing points (not knowing suits) would help there much. If we want to play something we need to get in suits anyway.

Also if GF is set i strongly prefer for strong hand to be captain of bidding (at least initially), it saves spaces and hides information from opponents.

But this looks like you are using SP/DN so that 1♣ opener would be captain of these auctions. That i find strange, especially as opener will probably be first to show his shape anyway.

[straube]

I tried placing all 5 QP hands into positives and 2 QP hands with a useful 4+ hcps into semipositives. For 200 hands...

1D.....29%
1H.....42%
1S......19%
1N+...10%

or breaking it down....

1D with 8+...........27.5%
1D with 7...............1.5%
1H with 3-4 qps....30.5%
1H with 2 qps........11.5%
1S..........................19%
1N+.......................12.5%

or
GF....39%
SP.....42%
DN....19%

On this run, the current meaning of 1S (2 or fewer QPs) would actually occur 30.5% of the time! Our club is stronger than Moscito, but it seems like 1S must be a frequent (if less so) response for them as well.

I think this might be better for us. It allows more semipositives to be relayed (zooming with 4 QPs instead of the so infrequent 5 QPs). It avoids preempting hands that have some points; many of these hands are better than what is allowed by our current structure (which permits a lone ace as a semipositive)

[straube]

wclass___, on 2011-April-20, 11:01, said:

According to your bridge logic if there were only 3 shapes..e.g.: 4333; 6322;6430. With corresponding probabilities: 60%; 32%;8% and permutations: 4;12;24. It would be poor bridge to play anything but 1♣-1♦ as 4333; 1♥ as 6322 and 1♠ as 6430...

I disagree. We put all of our balanced (5332, 4432, and 4333) and all of our single-suited major and all of our 5/4s with a major into the GF 1D response so that we have room enough to unwind these patterns at an average of +0 steps compared to standard symmetric. This explains why the 1D response is so common (and yet not as common as would be a 0-7 response). Our 1N up responses unwind hand patterns at +0 steps compared to standard symmetric.

wclass___, on 2011-April-20, 11:01, said:

As i understand none of your 1♦/1♥/1♠ responses tells anything about suits...
In uncontested 1♣-1♦(negative) i expect way better results. Only mayor advantage should come from competitive bidding, but i can't imagine that knowing points (not knowing suits) would help there much. If we want to play something we need to get in suits anyway.

You're right about the 1D, 1H, and 1S responses. Hard to say what works best in a competitive auction. I feel sure (and I've read support for the idea here) that establishing a GF is very important for coping with further competition which 1C-1S as DN or GF balanced does not do. I figure that something like 25% of the time, responder will have a non-balanced hand. We start to unwind probably 45% of these patterns immediately (1N on up), another 45% on the subsequent round (again 1N on up), and perhaps 10% on the third round.

wclass___, on 2011-April-20, 11:01, said:

Also if GF is set i strongly prefer for strong hand to be captain of bidding (at least initially), it saves spaces and hides information from opponents.

But this looks like you are using SP/DN so that 1♣ opener would be captain of these auctions. That i find strange, especially as opener will probably be first to show his shape anyway.

For gf, usually opener is captain. He's able to reverse relay only a fraction of the time. Responder's balanced hands are placed within the 1D response to allow opener to have room to show distributional hands, but most of the time he has a balanced hand and a lot of the rest of the time, responder has started to spin off shapes first. Here's how it goes....


1C-1D (bal or major)
..........1H-spades or balanced (or any hand that thinks better to retain captaincy)
...............1S-balanced or hearts and clubs
....................1N-balanced or any hand that thinks better to retain captaincy
.........................2C-hearts and clubs
.........................2D+ balanced hand patterns
....................2C-spades and clubs
....................2D-6 spades
....................2H+ spades and diamonds
...............1N-hearts
....................2C-relays
.........................2D-6 hearts
.........................2H+ hearts and diamonds
...............2C-spades and clubs
...............2D-6 spades
...............2H+ spades and diamonds
..........1S-hearts
...............1N-relays
....................2C-hearts and clubs
....................2D-6 hearts
....................2H+ hearts and diamonds
..........1N-majors or 3-suited short a minor
..........2C-minors
..........2D-6 clubs
..........2H-3-suited, short a major
..........2S+ 6 diamonds
.....1N-majors or 3-suited short a minor
.....2C-minors
.....2D-6 clubs
.....2H-3-suited, short a major
.....2S+ 6 diamonds

As far as 1C-1H goes, opener can retain captaincy with 1S (relaying responder's shape) or can make a natural nf and descriptive bid.

1C-1S looks like co-captaincy is best with opener showing his shape in a way that permits a rebid for him and frequently for responder.

[wclass___]

straube, on 2011-April-20, 15:16, said:

I disagree. We put all of our balanced (5332, 4432, and 4333) and all of our single-suited major and all of our 5/4s with a major into the GF 1D response so that we have room enough to unwind these patterns at an average of +0 steps compared to standard symmetric. This explains why the 1D response is so common (and yet not as common as would be a 0-7 response). Our 1N up responses unwind hand patterns at +0 steps compared to standard symmetric.

Disagree with what? I will try to explain in very simple words what i tried to say in case you misunderstood me.

Although 4333 will come up 9 times more than 6430, 6430 includes 6 times more exact shapes and it has way more game/slam potential.. so it needs more bidding room. Same thing for range 5-7 or 8+ they might come up in same frequencies, but you will probably need more space for 8+ to describe them (to limit, honor location etc). So i think it is not correct just to look on probabilities.

Also i am afraid that you are ''-'' steps compared to standard methods. Opener has to bid 1♥ to start relays while, using 1♥+GF, it happens automatically.. as if opener had bid 1♦ to inquire. Your 1N+responses obviously helps here, but not enough to get till ''0''.

straube, on 2011-April-20, 15:16, said:

I feel sure (and I've read support for the idea here) that establishing a GF is very important for coping with further competition which 1C-1S as DN or GF balanced does not do.

I like these bridge basics: limit yourself, bid your longest suits first, don't under lead aces.. But then one should forget all rules and start thinking about each situation exclusively.

I think there might be hands that you can add for DN/negative that don't fear competition and can be implemented without noticeable effects.

straube, on 2011-April-20, 15:16, said:

Hard to say what works best in a competitive auction.

I see that you prefer points by far, but i believe that it is better that you know suits rather than points. But of course it is even better f you know both.

[straube]

wclass___, on 2011-April-20, 17:54, said:

Disagree with what? I will try to explain in very simple words what i tried to say in case you misunderstood me.

Although 4333 will come up 9 times more than 6430, 6430 includes 6 times more exact shapes and it has way more game/slam potential.. so it needs more bidding room. Same thing for range 5-7 or 8+ they might come up in same frequencies, but you will probably need more space for 8+ to describe them (to limit, honor location etc). So i think it is not correct just to look on probabilities.

Also i am afraid that you are ''-'' steps compared to standard methods. Opener has to bid 1♥ to start relays while, using 1♥+GF, it happens automatically.. as if opener had bid 1♦ to inquire. Your 1N+responses obviously helps here, but not enough to get till ''0''.

I disagree that my logic would assign the three patterns as you suggested. I look at frequency yes but I also look at how much information has to be contained in any single bid. That's why our club is less frequent than our 1D opening and that's why our 1C-1D sequence is less frequent than our 1C-1H sequence. Our opening structure is designed with relays in mind.

We have left extra room for shapely hands at the expense of balanced hands. We are + compared to standard for balanced hands, we are +0 compared to standard for the 2-suited patterns, and we are -compared to standard for the single-suited (6+) hands. When it's balanced opposite balanced then we have lost room, but when responder is balanced and opener is unbalanced, we are +0 or - compared to standard for an opportunity to reverse relay that most structures don't have. Moscito has that opportunity but I believe it is +1 for most shapes.

[Free]

wclass___, on 2011-April-20, 11:01, said:

According to your bridge logic if there were only 3 shapes..e.g.: 4333; 6322;6430. With corresponding probabilities: 60%; 32%;8% and permutations: 4;12;24. It would be poor bridge to play anything but 1♣-1♦ as 4333; 1♥ as 6322 and 1♠ as 6430...

On occasion Richard has posted a way to evaluate the efficiency of relay systems. In short, you multiply the frequency of the hand with the level required to bid the hand, and do this for every possible hand. The lower the total score, the better the relaysystem. Imo this is a theoretically sound and objective method.

I disagree with your statement that it would be poor bridge to ... (see quote), it's much better to use 1NT and higher to immediately show certain hand types, but that would get us too far.
However, if we were only allowed to respond 1♦/♥/♠, imo you should order your bids according to the information they convey, unless the difference in frequency is extremely high. On general principles I'd rather play the other way around (1♦ = 6430 (0 cards known), 1♥ = 6322 (8 cards known), 1♠ = 4333 (12 cards known)), although the frequencies might confirm your method is better here.

The difference with your theoretical case and the situation after strong ♣ is that all distributions that can be SP can also be GF (except the very very very rare extreme shapes like 13-0-0-0). If you want to relay both GF and SP hands, then frequency comes into play. With the standard approach of 1♣-1♦(neg)-1♥(extra)-1NT+ = SP and 1♣-1♥+ = GF, you'll see that GF hands are bid 2 steps lower than SP hands. If SP is more frequent, then you can improve the relay system by lowering the SP's at the cost of leveling up the GF hands. That's exactly what happens with the approach I use (based on Paul Marston's work). We bring most SP 1 step lower (sometimes 2, sometimes no difference), and we bring most GF hands 1 step higher. Since SP are considerably more frequent than GF in that system, we improve our relay system.

[wclass___]

Well, yeah, obviously i meant 1♦/1♥/1♠ to be only bids available.

That quote wasn't targeting other methods, but rather for straube's comment as i perceived it as something like ''it is no good to play 1X as Y, because it comes up too much.''

Btw, who is Richard?

[wclass___]

Free, on 2011-April-21, 02:18, said:

On occasion Richard has posted a way to evaluate the efficiency of relay systems. In short, you multiply the frequency of the hand with the level required to bid the hand, and do this for every possible hand. The lower the total score, the better the relaysystem. Imo this is a theoretically sound and objective method.

Don't you think that hands with playing strength potential should be allowed to bid more slowly to explore slams?
e.g. 4333 comes up like 3 times more than 4441, they have same level required to bid the hand, but as 4441 has more playing strength potential for slams you won't have 3 times more slams with 4333 comparing with 4441. Although 4333 will probably need more place to be described as if one is using something like spiral scan, 4441 due to shortness will probably need less space to be described.

[hrothgar]

wclass___, on 2011-April-21, 06:00, said:

Don't you think that hands with playing strength potential should be allowed to bid more slowly to explore slams?
e.g. 4333 comes up like 3 times more than 4441, they have same level required to bid the hand, but as 4441 has more playing strength potential for slams you won't have 3 times more slams with 4333 comparing with 4441.

This suggestion is a reasonable one, though it might not be practical to implement.

Assume for the moment that we had a metric that accurately described expected "safety level" for a hand as a function of shape. It would seem reasonable add this information into the metric.

From my perspective, there are two major flaws with this scheme

1. As far as I know, said metric doesn't exist
2. "Game before slam". Slams are exciting, sexy, and rare. Relay systems already do quite well of slam hands. From my perspective, their weak spot is making a well informed decision regarding strain (especially whether to play 3NT versus 4M on a Moysian versus 5m).

[akhare]

hrothgar, on 2011-April-21, 06:47, said:

From my perspective, their weak spot is making a well informed decision regarding strain (especially whether to play 3NT versus 4M on a Moysian versus 5m).

Absolutely -- do you think that a practical solution for it might be revert to natural bidding with two min bal hands opposite each other? The question of course is whether it's possible to make that determination at a low enough level (say 2♠ or so)...

[hrothgar]

akhare, on 2011-April-21, 13:47, said:

Absolutely -- do you think that a practical solution for it might be revert to natural bidding with two min bal hands opposite each other? The question of course is whether it's possible to make that determination at a low enough level (say 2♠ or so)...

Consider an aucion like the following:

1♣ - 1♦
1N - 2♦

Where

1♣ = Strong, art forcing
1♦ = Game forcing
1N = Reverse Relay #1 (minimum opening, prefers to ask rather than show) Balanced or 4441
2♦ = Reverse Relay #2 (denies slam interest, prefers to ask rather than show), balanced with 4+ Hearts

At this point in time

Both hands are known to be limited
Both hands are known to be balanced

A 2♥ response should still be a relay ask, however, it seems non-sensical to use a relay break as anything other than a mechanism to explore for strain...

This is obviously a contrived example, however, I'd argue that there are plenty of analagous cases.

From my perspective, failure to document / describe relay breaks is one of the big weaknesses of literature on relay methods (my own dabblings included)

[akhare]

hrothgar, on 2011-April-21, 13:56, said:

Consider an aucion like the following:

1♣ - 1♦
1N - 2♦

Where

1♣ = Strong, art forcing
1♦ = Game forcing
1N = Reverse Relay #1 (minimum opening, prefers to ask rather than show) Balanced or 4441

I find the first reverse relay very intriguing because it reflects my own opinion that they should be used to describe minimum hands with no slam interest. However, there are other schools of thought which maintain that it's better the balanced hand to do the asking (regardless of strength).

What is your general perpsective on the use of reverse relays?

[Free]

akhare, on 2011-April-21, 17:23, said:

I find the first reverse relay very intriguing because it reflects my own opinion that they should be used to describe minimum hands with no slam interest. However, there are other schools of thought which maintain that it's better the balanced hand to do the asking (regardless of strength).

What is your general perpsective on the use of reverse relays?

I think we're going too far off topic, so I started a new one:
http://www.bridgebas...reverse-relays/

[PrecisionL]

Gentlemen, an interesting discussion, however ...

The following is from a book by Benito Garozzo in discussing systems in general.

Bridge is a game of suits

“The cards don't always play the same way. On offense, a queen is somewhat useless in with two low cards, however it is a certain trick with an ace and king. On defense a suit headed by three major honors might not come to a trick, whereas a hand with a double fit, each to concentrated honors in a long suit, on offense, produce a quantity of tricks absolutely disproportionate in respect to their HCP total. Two mirroring 4333's require a huge number of honors to develop tricks, but the presence of a singleton or void reduces drastically the strength necessary for game or slam in a suit.

"They are, these, all familiar principles to any player agonista, but they seem to be forgotten when other used systems are examined.

"Also, certainly the incorrect application of the law of tricks is a clear sign of ignoring the fact that bridge is a game of suits. The breakdown of high cards and the presence of unusual distributions carry such an importance, in truth, on the play of the hand, such that renders nearly insignificant the use of the law of total tricks. Larry Cohen, in his books, places such and so many conditions for a correct application of his laws, to render obvious how imprecise the Law is in the absence of the distributional information.

"To be able to take correct competitive action, in fact, it is necessary to know the suits and the distribution of partner’s high cards and to inform partner likewise of one's own high cards and suits. Otherwise the appraisal of the offensive and defensive potentialities of the deal could be based alone on absolutely generic considerations, and so often lead to wrong competitive choices.

"Not always, obviously, does the auction allow the exchange of all necessary information. A good system, consequently, tries to anticipate the problem through a structure of openings and responses that permits a rapid transmission of the essential elements of the hand in terms of suits, distribution, strength, and honor concentration."

[Edited out the discussion of Splinters.]

from the Pit Bulls site in Canada: http://www.pitbulls....20Splinters.htm

However, I understand your problems are more complicated playing a strong club of 15+ hcp. I do play a version of the Millennium Club with transfer negatives at the 1-level with 1♠ = 0-8 hcp.