Danlo

Thanks for the information! Eerie, the similarity between the opening bids. And probably originating from very different lines of inquiry (not much double dummy analysis in those days, I figure). I concede the 4H4S point. Not a good idea. I should have thought about it earlier. I have a problem with the NF 1NT bid, I don't know on which hands to bid it - probably 11-12 balanced with a doubleton spade, but those are relatively rare. A wider range means I'll miss games, a weaker range means I'll go down too many too often. I'll take my chances passing 1S, let the opponents figure things out. They don't know what I have. I'm far from convinced that playing 1NT is generally better than playing 1S. In a MP setting it may be, in an IMP setting I'm reasonably sure that on average it isn't. I've been contemplating the canapé issue, but I'm not convinced that it gains. Because minor suit openings are no longer 4333/4432 with spades, a 1S rebid after a minor suit opening already promises a 5 card in the minor and 4 spades. So you have already gained a perfect way of bidding those hands, why overload the spade opening? It may be a good strategy to dislodge the stronger, unbalanced hands with a five card spade opening. Not the 5332 or 5422 hands. But only if there is a nice way to place them in another part of the system. The 1S bid is loaded, but not overloaded now.

But then you use a canapé type bid with four spades and possibly a longer second suit? I want to make a primary distinction between 4/5 card suits and 6+ card suits (possibly combining 6322 and 7222 with the 4/5 card suits. These hands are not as strong as they seem). And secondary balanced vs unbalanced. And after that bal4 vs bal5 or unbal4 vs unbal5.

I'm not afraid of losing the suit. But if you don't bid it, you give the opponents more room to push you to the three level. Having the boss suit is something to capitalize on, so you make it more difficult for the opponents to push you, because they must take more risks to enter the bidding. And they have less room to investigate strength and fit. The goal is to buy the contract at the two level more often, and to be able to punish them for balancing more often. There is another gain (that can also be a loss): you unburden your minor suit opening bids buy taking out the weakish balanced hands with spades (I dislike the 1C-1H-1S-1NT sequence in natural systems). However, the opponents get a little edge, because they can overcall spades just a little more easily. The concept can also be useful to unburden the 1D opening bid in precisionish systems and gain even more competitive edge (Especially in combination with the 2 level one suited opening bids). You may even be able to drop the range of the 1S opening bid a bit. Even though I don't think opening balanced hands with less than 10 HCP is a long term winner. It also combines well with a conventional 1NT opening bid (as opposed to the standard weak/strong 1NT opening bid). Something I'm quite fond of (Hence this post in non-natural systems). If you use a 'fuzzy club' opening that can also hold 13-15 balanced hands without spades, combine it with a conventional 1NT (for some of the more difficult strong variants, possibly some two- or three suited hands) and one suited 2 level bids, you end up with a whole new system concept. But now we're getting way out of the scope of my original post.

I meant which ends up being better: 4333 with 4 spades, or (say) 4432 with 2 spades and 4 of each minor? Does having the boss suit beat out having terrible shape? I'd have to look into that, it will take me some time though. The analysis is not yet rigged for this specific question.

Do you have more info on the 2 level openings? One suited two level openings is something I've been experimenting with and that I find attractive. I'd like to about the responses you use.

It shows that 4333 is worse than 4432. But my analysis gets a little thin there, the number of deals I use is simply not large enough. The difference seems to be about a point (about as expected) in the competitive region. ut may be up to two points in the game region, which is a little more than I expected. 5332 seems to be about equal to 4432, while 5422 seems to be worth a little more in the game region. But all with a healthy amount of scepticism, because of the very small (<100 to <<100 deals) samples. The reason I do give you the results is because they are reasonably consistent over the whole area of analysis. A 1NT opening that promises exactly 4 spades has two disadvantages over my 1S opening: 1. it is more risky for partner to pass (not only because of double trouble, but also because of possible bad results compared to 1S passed) 2. there is no obvious forcing bid that doesn't have the drawback of limiting partners nonforcing options (like 1NT after 1S) Even so, it may be that the 1NT response (forcing and unlimited, but not in principle GF) is not allowed in certain ACBL levels of competition.

I've done some simulations on competitive bidding, that suggest to me it is preferable to open a five card hearts and a four card spades suit. The most standard layout: 1♥ = 5+ hearts, 12-19 HCP 1♠ = 4+ spades, 10-15 HCP The 1S opening will contain all 4333 and 4432 hands with 4 spades and may even contain some 5422 hands with a five card minor. Though the effect of that decision may is unclear, it may be too much of a hinder in subsequent bidding. The reason for this opening scheme is the (age-old) observation that it pays to have the boss suit. How much it pays, I only found out recently in my competitive simulations. In deals where there are no extreme distributions and no large (9+ card) fits, there almost always is a four point region (say 18-21 combined HCP), where the boss suit wins - even when the boss suit people have less trumps than the others. This region is far larger than I expected. There is a distinct and in this scheme maybe somewhat exagerated difference in the 1♥ and 1♠ opening: the 1♥ opening is more constructive (stronger) and shows more hearts, so you can more easily compete at the two/three level. The spade opening is more competitive (weaker) and concentrates on blocking the opponents path quite early in the bidding. The spade opening has to use a relaybid to be feasable (note that it now contains all kinds of balanced hands that in standard systems reside in 1C and/or 1NT), my preference is to use 1NT for this purpose. The rest of the bids are nonforcing. Does anyone even like the idea, or does anyone have experience with this kind of system?

jtfanclub wrote I do think 8-12 or 11-15 openers make a difference. The balance of power shifts, the frequencies shift, therefore the priorities shift. After an 11-15 opening bid, game is much more likely and therefore has much more priority in the response structure. It may be (but I'm exagerating here) that after an 8-12 opening bid it pays to give up on most of the game investigation because you need all non forcing bids you can get to get into a reasonable spot. As I said, opening more does not have to be better. It may be that your the fastest car around the first corner, but get caught on the second. And it may be that you make life for opponents easier because they have more bids or rounds of bidding available and more information to go on in bidding and play.

HRothgar wrote You're right, my statement was sloppy. What I'm talking about is that the primary goal of bidding is not always conveying clear information. Even though I realise that any bridge action (assuming that you want to maximize your expected score) does convey information, it may not be possible to state this information in a simple, declarative sentence (like: holds 8-12 HCP and has at least four spades). If you look at relaytype systems, the goal is often to narrow down the hcp range and give clear distributional information. This is what I meant when I talked about information theory as a model, but that was rather unclear. Especially early in the bidding, I have doubts whether this is the right strategy. I feel for a 'course coding' type of approach, where you do not give specific information. For example, you may have a bid that says 'spades is likely to be the right strain for this hand'. That may contain hands with long spades, but also 4432 hands with 44 in the minors. Because par analysis shows that for these 4432 hands spades is as likely a par contract as both minors.

I've read some post of the recent forcing pass thread and have some thoughts I'd like to share. First of all, I don't think that getting out of the starting blocks as quickly as possible is the best way to start the bidding. Starting the bidding is all about getting your partner into a good position to judge the continuation. I think bidding is a bit like (formula 1) racing. If there is only one corner to take, you want to take it as fast as possible. But if you have to take another corner right after the first, it may be best to slow down a bit in the first and get into a good position to take the second corner. It's all about controlled effort, not all out effort at every corner. This bugs me in the 8-12 opening bid range. It seems like a good idea looking at the opening bids only, but does it really get partner into the best possible position on the second bid? I know the 13+ pass probably doesn't, and I doubt whether the 8-12 opening bids really works that way. Secondly, the pass bid is the only bid that opponents cannot double, and passing after a (second seat or later) pass does not guarantee that partner gets another turn. So the preemptive effect of passing may be greater than the preemptive effect of opening 1C. So where a first seat forcing pass may or may not have merit, the second seat forcing pass may have a distinct disadvantage. Thirdly, the question is whether information theory is the best model for bidding theory. Bidding is not only about conveying information, it is also about finding the right contract to play. So passing should be an option for partner. I like to keep the number forcing bids to a minimum and I want to locate them where they balance loss (by not being able to play that contract) and gain (by being able to investigate other contracts). But balancing loss and gain is not only a question of passing on information. It is a question of grouping situations that require alike handling by partner and it is a question of frequency. A frequent situation takes more weight in the score than an infrequent situation. And if bid the same with two hand types where partner should pass if you have one type and bid if you have the other, partner has a true problem. So even if you give partner the best information possible, it may lead to your partner knowing that he is making the wrong decision, but not being able to help it. Since 8-12 opening bids put more strain on the subsequent bidding and invite opponents interference, they may suffer more from the above problems than standard opening bids.

You may want to check my site for the B Sharp system, a system I am developing to exhibit the system design features I'm exploring. The site is Bidding by Design and on the first page is a link to the B Sharp section. There is a B Sharp system summary, showing the current state of the system. Note that I do not believe the opening bids are leading in system design, they are mostly a consequence of the methods you want to apply later on. Furthermore, my primary interest is not in designing a bidding system, but in the principles underlying the design. I want to chart how certain design rules (like MAF) impact on a system, and how certain systems are related but use (slightly) different design rules priorities. I'm currently working on an article clarifying some of my design notions, but that is still some way from being finished. There are some snippets on the site, but they are not definitive.

Apart from some reservations I have on this concept (for one: why use hand shape as a criterium, rather then for example game forcing character of the bidding or trump suit establishment? Not that these are no better alternatives. But it seems that a system gets punished for conveying other than shape information and that doesn't seem fair), I have a remark about the calcution. 4432 has a frequency of about 21.5%. 5422 has a frequency of about 10.5%. 6322 has a frequency of about 5.6%. Calculate: 4432 with 3C = .215 * 16 = 3.44 5422 with 3D = .105 * 17 = 1.79 6322 with 3H = .056 * 18 = 1.01 Total 6.24 (are the factors correct? Seems to me 3C should be 12, but that is not really important) Compare this with: 4432 with 2NT = .215*15 = 3.23 5422 with 3C = .105*16 = 1.68 6322 with 4H = .056*23 = 1.29 Totalling 6.2 So it pays to ignore quite frequent shapes (I think 5.6% is quite frequent) in favor of even more frequent shapes. If we concentrate our bidding on 4333,4432,5332,5422 and 5431 shapes we can gain quite a lot, even though we shift all other shapes to (unreasonably) high bids. I don't think this is the right strategy. Of course, this is not a design tool or design criterium, but an evaluation criterium. And it should be used as that, to measure after the fact. But still, some unwanted quirks in a system may weigh positively in the result and that should be eliminated as much as possible. Is it possible to enhance the metric? I'd like to propose the following: instead of the step number, use the Fibonacci number of the bid. Starting with 1 for pass, 1 for 1C, 2 for 1D, 3 for 1H, 5 for 1S,..., 89 for 2NT, 144 for 3C, et cetera. The measure now becomes: Calculate: 4432 with 3C = .215 * 144 = 31 5422 with 3D = .105 * 233 = 24 6322 with 3H = .056 * 377 = 21 Total 76 Compare this with: 4432 with 2NT = .215*89 = 19 5422 with 3C = .105*144 = 15 6322 with 4H = .056*4181 = 234 Total 268 But compare this with: 4432 with 2NT = .215*89 = 19 5422 with 3C = .105*144 = 15 6322 with 3S = .056*610 = 34 Total 68 So the last variant is the best and bringing the 6322 to a higher level will no longer pay of. At least, this is closer to my intuition of what I would prefer to have in a bidding system. But please let me know if you don't agree. ============ Bert Beentjes Nijkerk (Gld.) The Netherlands

Thank you all for your comments. I've been thinking a lot about symmetry the last few days. The definition of Richard looks right to me, for a certain kind of symmetry. If I could coin a term, it would be 'absolute symmetry'. Referring to the fact that the symmetry stems from using the exact same bid for all instances of a shape. In my musings I have identified a second kind of symmetry, that I call 'relative symmetry'. Not very surprising of course, after introducing absolute symmetry. Relative symmetry means that the same bids have the same meaning, relative to X. Where X can be several things: the opening bid, the primary suit, the agreed trump suit and probably some other things. Relative symmetry is used quite a lot in natural systems, and somewhat less in relay systems, I suppose. Examples: simple raises of the opening suit. 1H-2H is symmetrical with 1S-2S, since the bid and the meaning are the same, relative to the opening suit. 2X means support in X and 6-9 HCP, where X can be subsituted by the opening suit. Relative symmetry is a little more complex then absolute symmetry, because you have to identify the kind of X needed and the actual X. Both kinds of symmetry have advantages and disadvantages. The most noteable disadvantages of absolute symmetry: if 3S shows 7330, the distance to 4H and 4S differs. This influences the possibilities for slam investigation. Also, depending on the opening bid, the distance from the opening bid to 3S differs. This will make it difficult to devise (relative?) symmetrical sequences ending in 3S. Furthermore, the ability to rightside the contract is probably depending on the seven card suit. If you have seven spades, the contract will probably be in the hands of the relay puppet, whereas a seven card heart suit will probably be played by the relay captain. Relative symmetry does not have these disadvantages. Because it is for example relative to the longest suit, it is easier to design a system where the contract is rightsided automatically in important cases. Relative symmetry also makes it easier to incorporate nonforcing bids. Relative symmetry has problems, however, when the envisioned trump suit is not the same as the suit where the symmetry is based upon, or when the prospective contract is 3NT. When you have relative sequences, 3NT will pop up in all kinds of places of all kinds of sequences, making it more difficult to devise a possibility for a last minute 3NT signoff. And if the long suit is diamonds, but you want to end in hearts, you may just be too high to do any kind of slam investigation. There are areas of the bidding where the advantages of each kind of symmetry are noticeable: relative symmetry is probably good in the start of the bidding, mostly because it keeps bidding straightforward and effective in non forcing situations. Relative symmetry is probably also effective in one suited scenarios. Since a six card suit is a likely trump suit candidate, centering the bidding around this suit seems a reasonable strategy. Absolute symmetry lends itself to situations where there are several trump suit candidates, or the result may be a no trump contract. But if you want to use these forms of symmetry in the same system (as of course we all do one way or the other), you'll have to fix the transition problem. To get from different opening bids into an absolute symmetrical structure, most systems use the 'big bang' option: the same (hence absolute symmetrical) bid is used to start the absolute symmetrical relay scheme. For example use 2C over any 1 level opening bid as a game forcing relay. Another strategy is to load the opening bids differently: put a little more into the 1H bid then the 1S bid. After 1H-1S(relay) the little more is bid with 1NT, the 2C and higher bids are used in the same way as after 1S-1NT(relay). This way you slide from different opening bids into an abolute symmetrical structure. I would call this something like 'slicing', since the difference in meaning between the next higher or next lower bid is the addition or removal of a slice. Relative symmetrical structures do not have the need foor big bang or slicing, they flow quite naturally, but only to a certain final contract. One of the problems I encounter in system design is that when reusing structures, the result is often not as expected. In one situation the structure works, but in another it doesn't. The problems I identified above are some of the reasons. At the moment (but insight change over time) I feel that using relative symmetry is the best approach when a five card or longer major suit is identified. Absolute symmetry is the best in most other situations, especially for (semi)balanced hands, but even with the most extreme minor suit oriented hand 3NT stays a good option. But in this approach full absolute symmetry is discarded for one suited hands and some two suited hands, because it is replaced by a hybrid approach (absolute when minor suit oriented, relative when major suit oriented). Anyone out there still with me ;) ? ============ Bert Beentjes Nijkerk (Gld.) The Netherlands "Give your partner easy decisions, not difficult puzzles."

Hi all, Symmetry is quite often used in the description of bidding systems, especially relay oriented systems. To be able to discuss symmetry, I'd like to hear some definitions (if possible: generally accepted definitions) of the symmetry concept. This will probably result in multiple definitions for related uses of symmetry, but that's OK with me. Thanks in advance!