Preempt over Polish club

[antonylee]

WJ2000 did this (it's even mentioned in the preface of the WJ2005 booklet that the change was mostly to simplify the definition for foreign players). The "panel standard" published by the "Brydz" magazine in 2006 does that as well. WJ2010 does this as well (according to a quick Google search). (And in case this matter, I do that as well, and I even put 5♦(332) hands into 1♣ in order to free up the 1♦-1M-1N for some artificial use (distinguish 5♦-4♣ and 4♦-5♣ hands over a 1♥ response, and more importantly show 5♦-4♥, less than reverse strength hands over 1♠). I'm actually interested in knowing whether this is a good idea or not.)

 

So please stop trolling indeed.

[the hog]

I did say in old fashioned versions on PC where 5 card D suit is still played, (such as Balicki's), 4 card D suits were opened with 1C. That is a very old concept and is largely notplayed anymore. Look at Matula's and Idzdebski's versions for example.

 

"WJ2010 does this as well"

As a matter of fact it does not. "Opening 1D = 12–17 5+ or 4441; Bal 12–15 4D, 4D + 5C"

 

"(And in case this matter, I do that as well, and I even put 5♦(332) hands into 1♣ in order to free up the 1♦-1M-1N for some artificial use" This is not PC, but rather some homegrown hybrid version

If you ask questions about your hybrid home made system, please do not call it PC. Your post is titled "Preempt over Polish Club".

[antonylee]

I would be interested in knowing where you found that piece of info. A quick search gives me "WJ2010 standard, modified for TOP individual 2010": 1♦ 12-17 with 5♦ or 4441, bal. 12-15 4♦-5♣. I would be surprised if the modification for some individual event (whatever TOP is) changed something as basic as the definition of the 1♦ opening.

As for my homegrown system, it is irrelevant to the discussion -- all still applies more or less identically in a WJ2010 (or 2000 if you really want) context.

 

Coming back to the main thread, I like straube's idea of using 4♦ as a strong ♥ raise -- but even more using Rubensohl at the 3-level as well.

[straube]

WJ2000 did this (it's even mentioned in the preface of the WJ2005 booklet that the change was mostly to simplify the definition for foreign players). The "panel standard" published by the "Brydz" magazine in 2006 does that as well. WJ2010 does this as well (according to a quick Google search). (And in case this matter, I do that as well, and I even put 5♦(332) hands into 1♣ in order to free up the 1♦-1M-1N for some artificial use (distinguish 5♦-4♣ and 4♦-5♣ hands over a 1♥ response, and more importantly show 5♦-4♥, less than reverse strength hands over 1♠). I'm actually interested in knowing whether this is a good idea or not.)

 

So please stop trolling indeed.

 

I'm only a little acquainted with Polish Club, but hopefully this is helpful anyway. I think that if your system allows 1D-1H, 1N to show a specific 5/4 holding (e.g. 4D/5C or 5D/4H) that something is wrong here. I'm thinking that it points to your 1C being very overloaded (Polish Club is anyway imo) and your 1D is very underloaded. Tally up your opening frequencies and see if you agree.

 

Are you familiar with relay systems? That's what I play mostly and when we're working on our structure I have in mind about how much pattern we ought to be showing at any given point. For instance, our 2C response to our strong club shows 5/4 in the minors either way (properly loaded). Our 1D-1H, 2C shows 5D/4C (underloaded). Our 1D-1S, 2C shows 1435, 1444, or various 5+/4+ patterns (overloaded).

 

So for you to be able to show the relative lengths of 2-suits by the point of 1N makes me think it is too much of a good thing. You have an opportunity cost here.

 

I guess my first thought would be to move some/most of your 4+ diamond balanced hands into 1D. Has this fallen into disfavor? I'm not sure why though. I would think you would have an easier time finding diamond fits and also major suit fits when responder is not constrained to respond 1D over your 1C opening...assuming responder still needs 7+ points to respond a major.

 

No doubt your 1D opening would be less directional and less descriptive if you did this (maybe that's why some roll their eyes when they see partner has something like 2344) , but it shouldn't be very descriptive being your second lowest opening....taking up as little space as it does. Just for contrast, our 1D doesn't promise any diamonds and we do very well with it. You'll be far ahead of us promising 4+.

 

Good luck anyway.

[bluecalm]

Wow, I didn't know Hog is capable of arguing such hopeless case where he could be proven wrong by one google search. Pretty amazing stuff.

 

I guess my first thought would be to move some/most of your 4+ diamond balanced hands into 1D. Has this fallen into disfavor? I'm not sure why though

 

Polish 1D is one of the strongest part of this system because:

a)1D opening means we can often compete in diamonds because either opener has 5 of them or is unbalanced, suit oriented hand (Hx - Hx - 4 -5 or something opens 1C usually)

b)you don't tell them which minor to lead on simple 3NT hands.

 

You'll be far ahead of us promising 4+.

 

Well, at leat your system has advantage b) so it's not that simple.

 

Jassem's experiment to try to make pc more like sayc didn't work as people didn't like his idea. Apparently he doesn't like it either because he doesn't play it and he retreated from it in wj2010.

[straube]

I'm sure the 1D opening opening does very well. I'm questioning whether one can afford to keep such specific meaning for such a low opening. I haven't studied PC so I'm open to being wrong on this point. Here's Rodwell discussing minor openings. It may or may not be relevant as he is primarily contrasting strong club vs standard....

 

 

BridgeMatters: Do you think, when opening a Polish 1C or a Precision nebulous 1D, that there is an advantage in that responder and the opponents do not know what opener’s better minor suit is?

 

Eric Rodwell: Let me answer this way. Eric Kokish, one of the chief proponents of weak notrumps, once said that weak notrumps bury your fit but they also bury their fit. I think, to a certain extent, that is true with artificial one-of-a-minor openings. They make it harder to find where your minor fit really exists, but you have advantages with the opponents having less information. For example, if you open a Precision 1D and rebid 1NT, and the opening leader is 4-4 in the minors, he has no idea which one to lead or whether to punt with a short suit lead. Playing standard, the leader has a much better idea.

 

BridgeMatters: How would you weigh that trade-off—responder not knowing what opener’s best minor is but also hiding it from the opponents?

 

Eric Rodwell: It is a loser for the strong club if that was all there was to it. In other words, in standard, I would not want to open 1C or 1D randomly. The combination of limiting the hand with a well-designed system—and there are a lot of ways to do it, and we are just talking about some of them—can more than compensate for that disadvantage. You are limited to certain value ranges, and there are fewer types of things you need to clarify later, especially in a competitive auction. So, depending on how well the system is designed, the strong club is my preference, though a lot of people would disagree with that, obviously.

[Zelandakh]

The trade-off in PC is that the 1♦ opening is very well defined (and a big winner) but that the 2♣ opening is poorly defined and a loser. YOu probably know by now that I use almost exactly the same opening bid structure as PC for my system except that the "weak" nt range in 1♣ is 15-17 for me rather than 12-14. The 1♣ opening is therefore much less common but has the same number of hand types which makes little difference where iot matters, in compeittion.

 

The trend for these systems is for the 2♣ opening to be 6+ without a major and to throw the 5♣4M hands into other openings. But somehow this just does not seem to work as well in a multi-way club context as with precision. So yes, PC can afford for its 1♦ to be so specific. Let me give you an analogy here which may or may not make sense. One of the trends (arguably) being seen in natural systems is for more and more balanced hands to open 1♣ and for 1♦ to be closer and closer to promising an unbalanced hand. Here the 1♦ opening is covering the same shape range of hands but a slightly increased points range. We only need to be able to find a way of handling the 18-21ish hcp hands via the 1♣ opening for these to be practically identical. Or, to put it another way, it is not so much that the 1♦ bid is dangerously underloaded in PC so much as it is (again arguably) overloaded, in comparison to 1♣, in 5 card major standard.

[straube]

I agree that 1D is overloaded in comparison to 1C in 5-cd major standard. 1C leaves so much more room so that if your only intent in opening 1D is to announce a weak NT hand looking for a major suit fit, then you would do better to open 1C and use transfer (or other) responses. It seems quite a bit different when you're adding strong meanings to 1C (as in PC). When you add strong meanings to 1C, then your response structure to 1C changes significantly.

 

So your 1C is 15-17 NT, 15+ clubs, 18+ any? That's basically a strong club? I think I like this much better than PC...but I'm a strong club fan.

 

I appreciate your trying to explain why PC's 1D can be so specific, but I don't get it. Perhaps you're saying in part that at the point of 1D-1S, 1N showing perhaps 4H/5D that you have a larger range (12-17 or so?) and that the bid therefore is not as underloaded as would be true of an 11-15 range.

 

I'd love to see a distribution breakdown for PC openings with 1C as all weak NT, 15+ clubs, 18+ other....

 

1C ought to be sky high in contrast to 1D and other openings.

[mgoetze]

perhaps you would like to contribute something useful, such as pointing to systems where balanced hands are opened with 4D are opened with 1C or is that beyond your intelligence?

 

Polish Club, as played by actual Poles. But this has already been pointed out, so I didn't consider it necessary.

[antonylee]

I'd love to see a distribution breakdown for PC openings with 1C as all weak NT, 15+ clubs, 18+ other....

I did a quick sim yesterday, keeping only hands that open something between 1♣ and 2♣ inclusive. The definitions are not very exact (e.g. where to draw the line between 1♣ (good 15) and 2♣ (bad 15)) but it was something like 1♣ 39%, 1 other suit 15% each, 1NT 11%, 2♣ 5% (assuming 1♣ also contains balanced 4♦ hands). If balanced 4♦ hands are put into 1♦ the percentages only shift by about 2-3%. Of course if switching to WNT the main flux in probablilities is between 1♣ and 1NT (more than 10% move from the former to the latter).

[hrothgar]

My copy of The Polish Club by Matula is at home (and rather dated)

 

As I recall, balanced hands with 4 Diamonds were opened 1♦ rather than 1♣...

[straube]

I did a quick sim yesterday, keeping only hands that open something between 1♣ and 2♣ inclusive. The definitions are not very exact (e.g. where to draw the line between 1♣ (good 15) and 2♣ (bad 15)) but it was something like 1♣ 39%, 1 other suit 15% each, 1NT 11%, 2♣ 5% (assuming 1♣ also contains balanced 4♦ hands). If balanced 4♦ hands are put into 1♦ the percentages only shift by about 2-3%. Of course if switching to WNT the main flux in probablilities is between 1♣ and 1NT (more than 10% move from the former to the latter).

 

Thanks for running it. I looked through 100 hands. I placed 5M332s into strong NT and balanced 1D into 1C.

 

1C-38

1D-14

1H-13

1S-16

1N-14

2C-5

 

My results were similar. I didn't count how many balanced hands had 4 diamonds.

[bluecalm]

It's not true that system should be designed in such a way that the lowest openings contain the most hands (on frequency basis). The case could be made for lowest opening containing the most hand types (even if their combined frequency isn't that great) but even that is not absolute truth.

Imo putting balanced hand with 4 diamonds is design mistakes similar to:

 

1S - 2D not promising 5 or:

 

1S - 1NT

2D - not promising four

 

You just need to pack balanced hands into one call (2C and 2C respectively) and have other bids to be more descriptive and useful.

[straube]

Our 1D is significantly more frequent than our strong 1C. Frequency is a rule of thumb and I understand this rule can be broken. We break this particular rule because 1C is strong and forcing and important and we need more room to sort these hands out. I still feel PC's 1C is overloaded compared to its 1D, but I haven't played the system. I know it's a successful system, and if PC experts find that placing balanced diamond hands into 1C is more playable, then I expect they are right.

[the hog]

My copy of The Polish Club by Matula is at home (and rather dated)

 

As I recall, balanced hands with 4 Diamonds were opened 1♦ rather than 1♣...

 

You are quite correct, Richard, as they are in Idzdebski's version. Even in Strefa, a Polish C variant, 1D showed 4 and could be balanced. However, let the boys play with themselves, as that is all they are capable of doing.

[benlessard]

unbalanced hands with D in the range 17-21 should be in 1D not in 1C IMO. Balanced hands with 5D can also be in 1D.

 

1C-38
1D-14
1H-13
1S-16
1N-14
2C-5

 

Those % looks slightly off. 1D should be higher than 1S or 1H, there is 4441,4144,1444,??45 and maybe even some ??46. While the 1M advantage is when its 5M+5D (5251,5152,5350,5053) and some 5H+6D. So in the end 1D should have a higher frequency. In PC its sure that 1D is underused and 1C is overused.

[straube]

unbalanced hands with D in the range 17-21 should be in 1D not in 1C IMO. Balanced hands with 5D can also be in 1D.

 

1C-38
1D-14
1H-13
1S-16
1N-14
2C-5

 

Those % looks slightly off. 1D should be higher than 1S or 1H, there is 4441,4144,1444,??45 and maybe even some ??46. While the 1M advantage is when its 5M+5D (5251,5152,5350,5053) and some 5H+6D. So in the end 1D should have a higher frequency. In PC its sure that 1D is underused and 1C is overused.

 

Well, I placed 12-14 5M332s into 1M and 12-14 5D332s into 1C. I tended to count 11 pt hands with shape (including diamond hands) and not 11 pt balanced hands. I only looked at 100 hands.

[Tomi2]

Jassems WJ05 System differs in the english book and the polish one, because he tried to make the system easier for foreigners. I talked to Jassem about his books and he says, that he a) does not like everything that is mentioned and b) has not invented everything there himself, his aim was to define some actual standard in Poland and make some easy book for non-polish readers

 

If you open 1Dia in Poland with a bal hand and 4 card only you surely play against the field! If you go to Poland and play an individual tournament, and your CHO is K. Jassem, you can expect him to open 1 club on 2344...

 

this 90% somebody has mentioned fits with my experience of 5000+ boards at polish bridge festivals. some 10% really open 1dia and sometimes gain tricks because silly oponents like me forget to ask about their openers and play on 5 cards on defense and declarer play...

[bluecalm]

I guess I like humiliating Hog in this discussion:

 

as they are in Idzdebski's version

 

Yeah, but Izdebski is one guy, who has his own vision of bridge. While he writes a lot he is a bit of controversial figure and but no way authority on what is standard or good agreement in pc as he usually pushed his own hybrid standard/polish system.

 

Even in Strefa, a Polish C variant, 1D showed 4 and could be balanced

 

This is just not true. Here are links to the cc's of best Polish pairs from recent years playing this variant of pc:

 

http://bridgefiles.net/pdf/Kwiecien-Pszczola.pdf

http://www.ecatsbridge.com/documents/files/ConventionCards/2008Pau-european/OpenTeams/Poland/jagniewski_kwiecien.pdf

http://www.ecatsbridge.com/documents/files/ConventionCards/2011Veldhoven/BermudaBowl/Poland/buras-narkiewicz.pdf

 

As you can see all of them play 1D as 5+ or 4-4-4-1.

Even in original Martens - Lesniewski system which began "strefa" system 1D was 5+ or 4-4-4-1.

It never changed in mainstream and I think there never was top Polish pair playing 1D as 4+.

 

So Hog... what about admitting you were wrong for once and move on ?

[the hog]

"Yeah, but Izdebski is one guy, who has his own vision of bridge."

In your mind only. Jealous? Really BC, you appear to think you are a legend in your own underpants. I think the only one humiliated is you.

I suggest you look up Jassem-Tuszynski writeup.

My last communication with you. Perhaps you know the old adage "Never wrestle with a pig. The pig enjoys it and you only get dirty.

Talking of humiliation, by the way, you still have not apologised in the other thread for stating someone was cheating!

[mgoetze]

The humiliation can't penetrate his reality distortion field.

[bluecalm]

I suggest you look up Jassem-Tuszynski writeup.

 

So you found two players who at one point in time played together and used that opening.

Guess what, they are now in new partnerships (with Martens and Nowosadzki) and they both don't play that anymore. As Tuszynski is definitely the system maker in his new partnership I guess he reconsidered.

[Zelandakh]

To me it matters less what percentage of hands are contained in a bid (although this is a useful metric) but rather how many hand types. Bids with multiple hand types often lack homogeneity which causes a problem in competition. The difference between 1♣ = 12-14 bal or 15-17 clubs or 18+ any and 1♣ = 15-17 bal or 15-17 clubs or 18+ any is miniscule in competition (the <18 hands are generally not acting in either method) and the Polish opening may well attract less competition (since you are mostly constructive in initial actions). Of course the Polish opening is much more common so if the bid is an overall loser then that effect is amplified; on the other hand if the bid is a winner that is also amplified. A big difference is in constructive auctions which are (imho) a little easier with the strong club than after a Polish 1♣. Whether this is enough to offset the additional preemption is an open question.

 

What I am trying to say here is that if you think the 1♣ opening in my system is not overloaded (as you say, it is not dissimilar from a strong club) then it is probable that the Polish 1♣ opening is also not overloaded despite what the percentages may say. The reason here is that there is a homogeneity between the weak and strong variants. Information that is good for one is usually goos for the other. This effect is seen even more so in Swedish Club methods. The nebulous 1♦ opening in modern Precision methods is an example of a bid that has mixed homogeneity - it is good from a strength point of view but bad in terms of hand pattern. Since the strength range is weak there is often not a huge amount of space to sort out the different hand types. The plus is that it is not always easy for the opponents to take advantage of this - overall though it is easy to see why this opening is a loser (in isolation).

 

As for the issue as to whether weak 4♦432 and 4♦333 hands should open 1♣ or 1♦ to be considered "official" Polish Club - well quite frankly who cares? It is clear that both styles are playable - opening 1♦ takes some hand types out of 1♣ (good) but considerably weakens the 1♦ opening. My preference would be to open these hands 1♣ and I would even give some thought/testing to opening weak 5♦332 hands with 1♣ too. But your mileage may vary.

[straube]

To me it matters less what percentage of hands are contained in a bid (although this is a useful metric) but rather how many hand types. Bids with multiple hand types often lack homogeneity which causes a problem in competition. The difference between 1♣ = 12-14 bal or 15-17 clubs or 18+ any and 1♣ = 15-17 bal or 15-17 clubs or 18+ any is miniscule in competition (the <18 hands are generally not acting in either method) and the Polish opening may well attract less competition (since you are mostly constructive in initial actions). Of course the Polish opening is much more common so if the bid is an overall loser then that effect is amplified; on the other hand if the bid is a winner that is also amplified. A big difference is in constructive auctions which are (imho) a little easier with the strong club than after a Polish 1♣. Whether this is enough to offset the additional preemption is an open question.

 

What I am trying to say here is that if you think the 1♣ opening in my system is not overloaded (as you say, it is not dissimilar from a strong club) then it is probable that the Polish 1♣ opening is also not overloaded despite what the percentages may say. The reason here is that there is a homogeneity between the weak and strong variants. Information that is good for one is usually goos for the other. This effect is seen even more so in Swedish Club methods. The nebulous 1♦ opening in modern Precision methods is an example of a bid that has mixed homogeneity - it is good from a strength point of view but bad in terms of hand pattern. Since the strength range is weak there is often not a huge amount of space to sort out the different hand types. The plus is that it is not always easy for the opponents to take advantage of this - overall though it is easy to see why this opening is a loser (in isolation).

 

As for the issue as to whether weak 4♦432 and 4♦333 hands should open 1♣ or 1♦ to be considered "official" Polish Club - well quite frankly who cares? It is clear that both styles are playable - opening 1♦ takes some hand types out of 1♣ (good) but considerably weakens the 1♦ opening. My preference would be to open these hands 1♣ and I would even give some thought/testing to opening weak 5♦332 hands with 1♣ too. But your mileage may vary.

 

When you say "hand types" are you referring to "hand patterns"? I'm not sure.

 

My preference is for strong club because while opener may have any hand pattern, the message that he has a strong hand has been delivered. For me, PC's 1C is used for disparate/inconsistent "hand types"...weak balanced, medium clubs, strong any. I frankly don't understand how the system is any good at all, but I also know that many many people do very well with it. So I can criticize it while admitting that it's quite possible or likely that I'm just not seeing the benefit. True, opponents can't compete with crap, but the continuations after 1C-1M, 2D (etc) seem very awkward to me. I also don't get 1C-1D, 1M nf. Seems a tremendous waste.

 

I understand your structure better, Zelandakh, because it's much more akin to strong club...it's just that your club is not quite strong enough to enter a GF easily. I'm not sure if you have overloaded 1C-1D sequences or how you manage this.

 

As far as our nebulous diamond...it's primarily a weakness when we might have opened 1D naturally and partner might have competed in diamonds. It is frequently a nuisance to opponents. They don't have a cue bid for example. The reason we settled on a nebulous 1D is because it actually handles fewer hand types than does Meckwell's 1D which promises two+ diamonds. That's because their 2D opening handles only 4 hand patterns while our 2D handles all the single-suited diamond hands and 6D/4C hands. Awm's 1D is even more limited in hand patterns than is ours.

 

Even so, we can relay our 1D hand patterns pretty easily. We also have room to relay our 1M, 2m as well. Our whole structure is set up with relays in mind. The funny thing though is that even when we don't relay, leaving appropriate amounts of room for relays seems to work out well for non-relay auctions.

[Zelandakh]

When you say "hand types" are you referring to "hand patterns"? I'm not sure.

 

Sort of. It kind of depends on context. In a competitive auction a weak NT is pretty much 2 hand types, either we want to compete or not. In PC the weak NT in competition is a single hand type since you are not allowed to compete. However, in an uncontested auction where we have the points the weak NT is multiple hand types depending on the level of detail required. The reason why Swedish club is so sensible in competition is that the hands across from a weak NT that want to compete provide exactly the same information as the strong hand needs to know. That means that Opener is usually well placed to make a sensible decision - this is not so good in PC because the 15-17 club hand complicates matters. But it is also not as bad as it first seems.

 

My preference is for strong club because while opener may have any hand pattern, the message that he has a strong hand has been delivered. For me, PC's 1C is used for disparate/inconsistent "hand types"...weak balanced, medium clubs, strong any. I frankly don't understand how the system is any good at all, but I also know that many many people do very well with it. So I can criticize it while admitting that it's quite possible or likely that I'm just not seeing the benefit. True, opponents can't compete with crap, but the continuations after 1C-1M, 2D (etc) seem very awkward to me. I also don't get 1C-1D, 1M nf. Seems a tremendous waste.

 

The incosistencies in the hand type disappear somewhat when you actually get into competition (as above) and without competition you have time to unwind. Since most of the time the 1♣ opening is a weak NT Responder essentially just assumes this and bids accordingly. If Opener has the strong hand then they can show this and the information from Responder's bid is helpful.

 

The reason why PC is a good system is essentially because the hand types are organised very efficiently. The 1M rebids are non-forcing but you generally only get dropped there when the hand belongs to the opponents, or when we have a sure fit and not enough for game. Playing in a 3-3 fit undoubled at the one level is not so bad when the opponents might have a game on!

 

Incidentally, strength-showing bids are not at all bad despite the many (distributional) hand types they contain. The extra power generally makes it somewhat safer to unwind them in competition and the general message is useful. There is a reason why I chose the 15-17 range for my 1♣ opening after all so it should not come as too big a surprise that I also like a strong club. I just do not like some strong club systems very much.

 

 

I understand your structure better, Zelandakh, because it's much more akin to strong club...it's just that your club is not quite strong enough to enter a GF easily. I'm not sure if you have overloaded 1C-1D sequences or how you manage this.

 

I manage this by burying the club suit. After 1♣ - 1♦, 1♥ is 18-20 any or 23+ balanced, 1♠ is 18+ 3-suited or any unbalanced game force, 1NT 15-17 balanced, and 2♣ 15-17 natural. That means I have to play the same system as over a 2♣ opening even for 15-17 point hands. This is surely not everyone's cup of tea!

 

 

As far as our nebulous diamond...it's primarily a weakness when we might have opened 1D naturally and partner might have competed in diamonds. It is frequently a nuisance to opponents. They don't have a cue bid for example. The reason we settled on a nebulous 1D is because it actually handles fewer hand types than does Meckwell's 1D which promises two+ diamonds. That's because their 2D opening handles only 4 hand patterns while our 2D handles all the single-suited diamond hands and 6D/4C hands. Awm's 1D is even more limited in hand patterns than is ours.

 

I have to admit that I am not a fan of the traditional Precision 2♦ opening and also believe there are more efficient ways of dividing the hand types up. I actually quite like a limited 1♦ opening that is diamonds or clubs but not both. I have considered that approach for my system too but there are certain reasons why it is less efficient there than in a 16+ context. Besides, you actually only need to get close to breaking even on these hands because you are winning on the 1M and 2m hands which are quite precise in your system. Therefore you can afford to lose some of these part-score battles.

 

 

Even so, we can relay our 1D hand patterns pretty easily. We also have room to relay our 1M, 2m as well. Our whole structure is set up with relays in mind. The funny thing though is that even when we don't relay, leaving appropriate amounts of room for relays seems to work out well for non-relay auctions.

 

There is an excellent reason for this! People tend to divide hand types into relay and non-relay. The relay is typically the next step but that leaves free every other bid for non-relaying. That means there is just as much space for non-relay auctions as relay ones. So the space is automatically close to being optimised. I suspect there is a pretty sound principle at work here that is useful even for non-relay systems.