2/1 or Big Club

[awm]

For perhaps a simpler example, take a look at the Ambra notes on Dan Neill's website. This is a pretty complex "2/1" based system, with notes that are extremely dense and to which a lot could be added for a serious partnership adopting the methods. You can play a complicated 2/1 system, it's just that a lot of people don't.

[glen]

Understand that when I say "I never heard of a more complex version of 2/1 than Mike's", that I wrote the small-font 259 page ETM Gold (has mini-multi structures etc.), and the shorter ETM Victory that has the "Dan-complexity level" warning label. Btw I have a new plug n play system up, CANDY, that combines two-way club, canape majors and Fantunes 2M openings - so a system that Soloway will never play.

[mikeh]

At the risk of boring those with no interest in the topic, the 2/1 to which I and glen refer included:

 

1) variable notrump, with different relay structure responses to the different ranges, and specified defences to virtually all defence methods in use

 

2) Artificial 1♦ response to 1♣, with relay sequences thereafter

 

3) transfer jump shifts over 1♣

 

4) 2♣ artificial gf relay responses to 1♦/♥/♠ openings (2/1 2♦ or 2♥ bids were invitational plus, so it was still a form of 2/1)

 

5) gf relay initiated by 4SF in otherwise non-relay sequences. All relays (other than over weak notrump) allowed for:

 

- approximate shape

- precise shape

- number of controls

- sometimes number of queens

- location of controls

- location of secondary honours, including, in extreme cases, Jacks

- breaking of the relay at any time

- in some cases, general size ask early

 

Those who subscribe to the BW will have seen some of our ideas used by L'Ecuyer - Marcinski in the Challenge the Champs.. altho I am not really saying they used 'our ideas'.. since the concepts have been widely available for many years, and the L'Ecuyer-Marcinski method is quite different from ours. But they use some of the compression approaches we used, in particular the denial cue-bid structure. This is an example of convergent evolution, not plagarism :)

 

6) denial cue-bidding for controls in relays after shape/size asks

 

7) spiral keycard

 

8) simple relays after 2♦ opening

 

9) transfer advances after their takeout double, including transfer flower bids

 

10) transfer advances after our overcalls

 

11) complex major suit raise agreements, based on the lack of need for a forcing raise (2♣ included all forcing raises).

 

12) complex system of responding to our 2N opening (initially a relay method, later changed to a more natural, but still very complex approach)

 

13) a large number of specific agreements re competitive and constructive bidding beyond those above... a significant number of agreements actually never came up in 5 years of serious competition.... but we had the agreements!

 

Including compressed flow-charts, in which all relay developements could be charted in a one-half page chart, the notes exceeded 175 pages, and were in most cases condensed, altho a number of pages were devoted to examples.

 

It was, and is, a wonderful system provided that you have a taste for extreme detail and both the time and the memory required to internalize the relay 'engines', which were designed utilizing a Fibonacci analysis (to ensure the most common responses were shown cheaply) tweaked to maximize the chance that relayer became declarer. We had many auctions in which we'd arrive at slam and all dummy could/should say in response to any question about the auction was: "Given what he knows about my hand, he wants to play this contract".

 

Moreover, while the CNTCs are not the Vanderbilt or the US Team Trials or the European Open, etc, the method did very well at the Canadian level... and got us to a couple of WCs.

 

Having said all that, my then-partner, who did 99.9% of the system design, always pondered the idea of developing a more complex approach based on a big club method :D

[kenrexford]

See! Proof is provided. 2/1 GF can be VERY complex, if you so desire.

[han]

I doubt I would hire Soloway once I won the money. And you know what, I would probably ask the pro for advice about system.

[helene_t]

I doubt I would hire Soloway once I won the money. And you know what, I would probably ask the pro for advice about system.

Sounds like a good idea. But you won't have to win the lottery since you and Arend will get rich defeating my 3-card major system.

[Echognome]

It was, and is, a wonderful system provided that you have a taste for extreme detail and both the time and the memory required to internalize the relay 'engines', which were designed utilizing a Fibonacci analysis (to ensure the most common responses were shown cheaply) tweaked to maximize the chance that relayer became declarer. We had many auctions in which we'd arrive at slam and all dummy could/should say in response to any question about the auction was: "Given what he knows about my hand, he wants to play this contract".

A few points here.

 

1. I don't think your disclosure there is complete as stated. Given that you say above that relayer can break at any time, perhaps you might mention that relayer is also denying certain hands types by failing to break the relay. I'm not saying that you should describe all the possible hands that are denied, but maybe if you could categorize them as "and he has denied certain minimum hands that would have broken the relay," it might be better. I know it's not your intention in the slightest to hide anything, but against a slam, it's probably good to discuss both the bids that were chosen (relays) and those that were not chosen (breaks).

 

2. "Fibonacci analysis" seems like a strange term to me. Maybe it is appropriate. But, to me, the Fibonacci aspect of relays is just that the number of hand types that can be shown by a given level up to another given level is determined by the Fibonacci sequence. Usually as a target, the goal is to be able to describe as many hands as possible at or below 3NT. The hands that are described above 3NT typically have extreme shape. Furthermore a system designer will have several goals when designing relays. One such goal is to make Relay Asker the declarer. (Note how TOSR focuses on this aspect.) The reason for that is to make it harder for the defense to know declarer's assets. Another important goal in many designs is to reduce the memory load. That goal will often conflict with the efficiency of the relay (more common hands bidding at a lower level). Also, I found an important goal is to keep the more similar hand types together in case relay asker just wants to jump to game (in particular if the strain has already been taken by relay responder).

[mikeh]

It was, and is, a wonderful system provided that you have a taste for extreme detail and both the time and the memory required to internalize the relay 'engines', which were designed utilizing a Fibonacci analysis (to ensure the most common responses were shown cheaply) tweaked to maximize the chance that relayer became declarer. We had many auctions in which we'd arrive at slam and all dummy could/should say in response to any question about the auction was: "Given what he knows about my hand, he wants to play this contract".

A few points here.

 

1. I don't think your disclosure there is complete as stated. Given that you say above that relayer can break at any time, perhaps you might mention that relayer is also denying certain hands types by failing to break the relay. I'm not saying that you should describe all the possible hands that are denied, but maybe if you could categorize them as "and he has denied certain minimum hands that would have broken the relay," it might be better. I know it's not your intention in the slightest to hide anything, but against a slam, it's probably good to discuss both the bids that were chosen (relays) and those that were not chosen (breaks).

 

2. "Fibonacci analysis" seems like a strange term to me. Maybe it is appropriate. But, to me, the Fibonacci aspect of relays is just that the number of hand types that can be shown by a given level up to another given level is determined by the Fibonacci sequence. Usually as a target, the goal is to be able to describe as many hands as possible at or below 3NT. The hands that are described above 3NT typically have extreme shape. Furthermore a system designer will have several goals when designing relays. One such goal is to make Relay Asker the declarer. (Note how TOSR focuses on this aspect.) The reason for that is to make it harder for the defense to know declarer's assets. Another important goal in many designs is to reduce the memory load. That goal will often conflict with the efficiency of the relay (more common hands bidding at a lower level). Also, I found an important goal is to keep the more similar hand types together in case relay asker just wants to jump to game (in particular if the strain has already been taken by relay responder).

1. The answer I gave assumed that relayer had relayed until he placed the contract. Obviously, if relayer broke the relay, and the auction continued, as it often did, we would explain the inferences from the relay break. A common break was after a cheap response to 2♣, a bid of 3♣ announced a gf hand with clubs and no slam interest. Furthermore, our agreements were that the more one relayed, the stronger one's hand, so that if we relayed several times, and then broke the relay, opener had rights. And, my explanation in this forum was not intended to be a complete description of what we'd announce at the table: thus responder would give a detailed description of opener's hand...once in a while we could claim before the opening lead and we did, at least once, expose declarer's hand to the opps before the opening lead. That was before we realized that sometimes we'd screw up the relay :P

 

2. I agree with your description of the goals, and excuse me for a possibly poor use of the term fibonacci analysis: my partner was the math major/computer whiz who designed the system. We focussed on maximizing concealment of declarer's hand and nesting hand-shapes to maximize bidding space below 3N: we had the only responses to 1M/2♣ beyond 3N as 5=4=2=2, with the major and clubs as the long suits and significant extra values including at least 5 controls... responder always held either a gf in clubs, a gf raise of the major or big notrump hand or some other hand with strong slam ambitions opposite a minimum hand, so that in all cases we had 4 level safety. We were not so interested in minimizing memory work, since we had agreed that we had a specific goal in mind: to play in the BB.. which was a lofty goal given that neither of us had previously won a CNTC when we got together, and we were prepared to WORK on the method.

 

I certainly would not recommend our method for anyone other than a work-alcoholic. I don't regret the work at all: it was a lot of fun, and when we were 'on' we were a pretty formidable partnership. But I doubt that I would ever do it again... I was only 42 when we started, and I'm not sure that I could or would do it again, hence my choice of 2/1 in response to the OP: I didn't mean a relay 2/1 either :P

[Echognome]

<snip>

2. I agree with your description of the goals, and excuse me for a possibly poor use of the term fibonacci analysis: my partner was the math major/computer whiz who designed the system. <snip>

 

I certainly would not recommend our method for anyone other than a work-alcoholic. I don't regret the work at all: it was a lot of fun, and when we were 'on' we were a pretty formidable partnership. <snip>

Mike,

 

I probably did post this with the correct language. My point was not to find fault with your terminology, but rather to say that the concept isn't as complicated as it is often made out to sound. I know *you* understand the concept, but I'm going to post the logic behind it for those that do not.

 

Imagine that we have a set of possible hand types. Typically these types are in the form of hand shapes. For example, one hand shape might be 4=1=5=3 and another 4=3=1=5. In order to show each of these hand types, we need at least two bids (one for each shape) if we want to show this exactly. We often call this a "full" relay. If we only want to show that the pattern is, say 4=x=5=y, where x and y could be any length, then we have a "partial" relay. Either way, the number of types of hands we can show depends on how much bidding room we have left below a given level (which I'll set at 3NT as it's the most common one).

 

So imagine that we have described our hand up until now and the next asking bid (relay) from partner is going to be 3NT. Obviously, since partner asks with 3NT we have no room to describe any hand types, so we assume only the 1 hand type is shown. If partner asks with 3♠, we have a similar problem. We can only bid 3NT (and stay at or below 3NT). But suppose now that partner asks with 3♥, then we can bid either 3♠ or 3NT. That is to say, we can now show 2 hand types. If partner asks with 3♦, then we can show 3 hand types: one each for 3♥, 3♠, and 3NT. However, the complexity goes up when partner asks with 3♣. Why? Well we can group together more than one hand in our 3♦ response, since partner can make a further ask with 3♥. Thus we can put 2 hand types in the 3♦ response and 1 each in the 3♥, 3♠, and 3NT response, for a total of 5 hand types. If we carry this on, we get:

 

Ask #Hand Types

3NT---1

3♠---1

3♥---2

3♦---3

3♣---5

2NT---8

2♠---13

2♥---21

2♦---34

2♣---55

1NT---89

 

This sequence you will not is one where the next value in the sequence is the sum of the two previous values. (3 = 2 + 1, 13 = 8 + 5, etc) That sequence is known as the Fibonacci sequence. It is useful for planning how much information you can extra at a given level. Thus for Mike and his partner, if they use 2♣ as their relay after 1M, they can show 55 types of hands at or below 3NT. If they used a 1NT relay instead, they could show 89 hand types. It actually gets more useful than that. Suppose they do use 2♣ as their relay. They can go further with the classification. As those 55 hand types can be broken down as follows:

 

2♦ - Can hold 21 hand types

2♥ - Can hold 13 hand types

2♠ - Can hold 8 hand types

2NT - Can hold 5 hand types

3♣ - Can hold 3 hand types

3♦ - Can hold 2 hand types

3♥/♠/NT - Can hold 1 hand type each

 

For a total of 55 hand types. So the subsequent sequences are also Fibonacci. I measure how efficient my relay systems are by seeing how much of a load each of the "buckets" holds. Maybe I'll put 20 hand types in my 2♦ bid, 11 in my 2♥ bid, 8 in my 2♠ bid, etc. That tells us how much use we are getting out of each bid.

 

I guess my point is that designing the relay can be fun and makes you think of how much information you can share. Maybe your goal is not to divide everything by full shape, but rather partial shape and also some measure of strength or controls. At least you can figure out how much you can describe.

[Badmonster]

Big club sure looks interesting. But 2/1 is accessible.