[Sigi_BC84]

Hi folks,

I'm playing a standard 5-card major system and Walsh-♦ (i.e. bypass ♦ in favor of major on first response unless GF+ values are there). We have some sophisticated agreements to handle the 1NT rebid after both a 1♦ and 1M response to the opening (relay/puppet structures). The memory load is quite high already, especially since some sequences are quite rare. So at the moment I'm toying with the idea of playing Keri or (preferably) Heeman over the 1NT rebid (and of course over the 1NT opener ;-). At first sight this looks workable to me (and I know that the idea isn't new). That would lower the mnemonic load as one would have to learn only one (tried-and-true) structure.

Now for my question: Is anybody here playing something like this and how well does it work? In case it didn't work that well, what were the issues?

Any comments appreciated.

--Sigi

[Free]

Don't understand why you'd use 4 card transfers (Heeman) when you already showed a 4 card suit... What's the point in an auction like 1♣-1♥;1NT-2♦!;2♥-3♦! showing an invite with 4♥ and 5+♦?

[awm]

Keri works pretty well over 1NT rebids, with the obvious modification that showing an "invite with the original major" should show five cards rather than four. Josh Sher and I play this (in part because we didn't want to take the time to discuss anything more complex).

[joshs]

If you want to play KERI over the 1N rebid, there are a few sequences which are idle, and others which are not all that useful and should be modified.

Obviously,
1C-1H-1N-3D can't be heart shortage. But thats not that big a deal.

The bigger deal, is that
1m-1M-1N-2C-2D-2N doesn't help much playing standard keri. I recomend either:
a. Playing that as a sound invite with a 5 card major (my usual treatment)
or
b. Game forcing waiting bid
or
c. just change the responses

Also,
1C-1H-1N-2S-2N-3new as rkc doesn't make much sense.

My preference is x-fer checkback instead of KERI (see, for instance, Eric Leong's article in TBW from june 91). This lets responder completely pattern out in game forcing auctions.

Josh

[Sigi_BC84]

joshs, on Feb 13 2006, 06:01 PM, said:

If you want to play KERI over the 1N rebid, there are a few sequences which are idle, and others which are not all that useful and should be modified.
[...]
My preference is x-fer checkback instead of KERI (see, for instance, Eric Leong's article in TBW from june 91). This lets responder completely pattern out in game forcing auctions.

Actually we do have a comprehensive checkback structure already that covers most cases. I was hoping somebody has already come up with a modification of Heeman or Keri that would nicely apply to the 1NT rebid -- apparently it's not that easy, and if the deviation from the standard sequences are getting too high one will end up with two different structures again, which would eliminate the main benefit of the approach (reduction of memory load).

About playing Heeman over the rebid: apparently that was not such a great idea of mine to begin with, I start to see that it's simply not effective (Free pointed out one pretty obvious example already).

Josh, could you maybe summarize the x-fer checkback approach? I'd like to compare that to what we are currently playing. Chances that I'll get hold of TBW from June 1991 are pretty close to zero...

--Sigi

[joshs]

Note: The following assumes natural 1m openers. If you are playing a precision 1D there are some modifications that you need.

After 1m-1S-1N
2C forces 2D and then:
2H or 2S are natural and INV (Optional: Play this as a shapely Invite)
2N Many options for this. I like: shows 5 in M and a sound invite
3m INV, only 4 in M
2D shows 4+H, either weak or game forcing. Opener bids 2H or preferences 2S with equal length.
2H shows 5+S, either weak or game forcing
2S is a game forcing raise in openers minor (in Leong's Article) Or you can play it as Game force with interest in a minor (as I currently play it)
2N Natural
3C to play
Others Natural and slammish with good suits (or INV with good suits if you prefer).

Follow ups from Leong:
x-fer then 2N = 4 card support for minor, you can then play natural continuations or relay for shape. I can't remember what was in leong's article
x-fer then 3m = unbalanced with 3 card support (hence typically 5134)

ALternatively Play:
2N Game forcing waiting
3m=4 card support

3'rd Alternative: x-fer then 2N as Keri, other's 5 card suits

After opener shows both majors, his 3'rd bid just patterns out.

Example 1:
Axx Qxx xxx AQxx
KQxxx AKJx x Kxx
1C-1S-1N(12-14)-2D(4+H)-2S(preference)-3C(GFing, 5413 or similar)-3S(Agrees spades for the moment)-?
Now by usual rules we don't play non-serious 3N here (NT already bid naturally) so responder will have to make 1 cue bid with this mild slam try and then opener will co-operate. Note that in these auctions, opener should only take over after a cue-bid with a perfeco since responder didn't have non-serious available.

Example 2:
Axx Qxx xxx AQxx
KQxx AKJ KJxxx x
1C-1S-1N-2S(Game force in some minor)-2N-3D-3N or
1D-1S-1N-2S-2N-3D-3S(weak hearts)-3N

Instead:
Example 3:
Axx Qxx AQxx xxx
KQxx AKJ KJxxx x

1D-1S-1N-2S-2N-3D-3H(weak clubs)-4C(Cbid)-4D-4H(rkc)
Probably 4H directly over 3H should be an offer to play opposite 4 good hearts.
Alternatively opener might just raise 3D to 4D directly. I don't think opener can ever sit for 3N with 10/12 points in responders suits.

Example 4: (If you want to play this in xyz auctions also)
KQxx Axx x Axxxx
xx KQxxx Ax KQxx
1C-1H-1S-2D(usually 6 H or a GF in this auction since opener doesn't have to have 2 hearts )-2H (or 3H with this super-max!)-2N (or 3C)-etc and a slam will be reached. Playing imps 6C is the better spot....

Example 5:
KQxx Ax xx Axxxx
xx KQxxx Ax KQxx

Here the Keri -auction will miss the great club slam:
1C-1H-1N-2D-2H-2N-3C(no fit, shortage?)-3N(None, less than a sound slam try)-P

A natural continuation works well here:
1C-1H-1N-2D-2H-2N(wating)-3C(5 clubs or 4 strong clubs)-4C
or
1C-1H-1N-2D-2H-3C(5 clubs or 4 strong clubs)-? (my vote is 3S then 4C but 4C directly is also fine)

Example 6:
KQx Ax xxx Axxxx
Xx KJxxx Qxx xxx
1C-1H-1N-2D-2H-P

In General, all you really need is:
a. INV sequences start with 2C (or a direct bid of 2N)
b. x-fers are weak or strong
c. in a game forcing auction, all 2N bids are just waiting
d. x-fer to 2N is GFing waiting with interest in partner's minor
e. In a game forcing auction, opener does not have to complete a x-fer, but can bid naturally instead:
Kx Axx xxx AKxxx
AQxx Kxx Ax Qxxx
1C-1S-1N-2S(Game forcing waiting)-3C(strong 5 card suit)-4C-etc.
In this auction responder bid 2S just in case opener would rebid 3C, but was planning on just bidding 3N otherwise.
f. Playing this structure, I reccomend using 3 level jumps as slammish (new suits are good 5 card suits) except for 3C which has to be to play. Alternatively, give up the 2N natural and play 4 suit x-fers and use 2C then 2N as the only NT invite.

There are many possible varients here...

Josh

[whereagles]

I've been playing "revised Keri" over 1 & 2NT rebids with the following modifications:

- Transfer to suit bid is 5 cards
- Invite in a major is 5 cards
- Transfer to spades after 1♥ response is 44 majors.
- Splinter into a suit bid is natural, 55 game-forcing. (With 64 you transfer and follow up with 2NT.)

The auction 1x-1y-1NT-2♠-2NT-3♣ is RKCB to suit y.

[joshs]

Sorry in my example 5 I meant to have
KQx Ax xxx Axxxx

[awm]

Playing Keri-Garrod methods over 1NT rebid:

Example 1:
Axx Qxx xxx AQxx
KQxxx AKJx x Kxx

1♣ - 1♠ - 1NT - 2♦(1) - 2♠(2) - 3♣(3) - 3♠(4)...

(1) Transfer to ♥; in context this will be always 5+♠ and 4+♥
(2) Preference; no reason to accept transfer when spade fit known to be at least as good.
(3) Patterning out; game forcing since a weak hand would pass.
(4) Sets ♠, good hand for slam in context.

Example 2:
Axx Qxx xxx AQxx
KQxx AKJ KJxxx x

1♣ - 1♠ - 1NT - 3♦(1) - 3NT(2)

(1) Shows short clubs; this rearrangement is a change between Keri and K-G.
(2) Clubs well-stopped, no interest in any other contract.

Example 3:
Axx Qxx AQxx xxx
KQxx AKJ KJxxx x

1♣ - 1♠ - 1NT - 3♦(1) - 4♦(2) - 6♦(3)

(1) Shows short clubs
(2) Clubs not suitably stopped; suggests a ♦ contract but denies a "perfect maximum."
(3) Partner should have 10-12 working points, and 27-29 outside the shortness suffices. Of course, it's also possible to go slow here.

Example 5:
KQx Ax xxx Axxxx
xx KQxxx Ax KQxx

1♣-1♥-1NT-2♦(1)-2♥-2NT(2)-3♣(3)-3NT(4)-4♣(5)...

(1) Transfer to ♥, starting a normal K-G notrump sequence.
(2) This sequence is one of the weaknesses in Keri, because holding a (52)(42) pattern responder's minor is never shown (instead shortness is shown). This detail is corrected in K-G methods. 2NT shows a game force with 4+♣.
(3) Relay.
(4) 2-5-2-4 shape. This guarantees either slam interest, or weak doubletons. Holding strong doubletons and no further ambitions we would use the "balanced game force" sequence.
(5) With three small diamonds and a nine-card club fit, it must be better to play in clubs. Some combination of cuebids and/or keycard should lead us to slam.

Example 6:
KQx Ax xxx Axxxx
Xx KJxxx Qxx xxx

1♣-1♥-1NT-2♦(1)-2♥-Pass

(1) Transfer to ♥.

-------------

Basically I'm just using K-G notrump systems here. The only changes:

(1) Bidding spades and then transferring to hearts implies 5+♠ and 4+♥. Opener should bid 2♠ instead of accepting the transfer if holding at least as many spades as hearts. Follow-ups by responder in this particular auction don't carry their normal K-G meanings, because of the guarantee of spades (which would not be known after 1NT-2♦-2♥ normally). Instead of the usual follow-ups responder simply makes natural calls (patterns out with a GF, passes with a weak hand).

(2) Similarly, bidding spades and then going through 2♣...2♥ implies 5+♠ and 4+♥ (invitational this time). Opener should bid accordingly.

These things are pretty intuitive actually, since inferences are available from the previous bidding without learning any "additional structures."

[joshs]

Adam, In example 5 how would you bid the 5422 with 4 in the other minor? You want to play 4M opposite a 4333 hand. Hence I didn't think that the sequence you have actually showed 4 card support for partner, at least not without discussion. Now there are some idle sequences that can be used to solve this problem (2C-2D-3N offers a choice, x-fer to the major and then 3N orders partner to show support) although there are still issues with mild slam tries (requires a max and a fit for slam).

BTW, tell us more about the re-arangement of the x-fer splinters?

[joshs]

BTW, here is another possible re-arangement in KERI:

1N-2C-2D-2H

In Standard Keri:
2S 4/5 Spades, 2H Min
2N 2-3 Spades, 2H Min
3C 4 Hearts, Good Min
3D any 4333, max
3H 3 trumps, not 4333, max
3S 4/5 spades, 2H max
3N 2-3 Spades, 2H Max

A re-arrangement:
2S 4/5 Spades, 2H Min
2N 2-3 Spades, 2H Min
3C Max Either 3H Or 5S or Both
3D any 4333, max
3H 4 hearts, good min
3S 4 Spades, 2H max
3N 2-3 Spades, 2H max

After the 3C bid:
3D asks:
3H 3 hearts (maybe 4 spades) then 3S asks for 4 spades and 3N says no.
3S 5S, 2H
3N 5S, 3H


You can also use 3H or 3S to ask for other things directly- for instance if you want to play in spades without revealing if partner has 2 or 3 hearts.

Anyway, I think being able to distinguish between 4 and 5 spades is probably more useful than being able to make a counter-try over the 4 cards support but good min
bid.


Josh

[joshs]

Here is a hand from the Albuquerque regional:
Partner: Kx xxx Axxxx AQx
Me: AQJxx AJx KQT xx

Our Auction:
1D(Precision)-1S-1N(Good 13-15)-2D(ART GF)-3D(Reasonable 5 card suit)-4D(I like that suit)-4S(Hx, non-forcing)-6N(A club lead might hurt us, or it might give us our 12'th trick, let them guess)
Partner was min, but both suits came in and we picked up a good swing. Slam was a very good contract.

Playing x-fer checkback then waiting 2N:
1D-1S-1N-2H-2S-2N-3D(4 strong diamonds or 5 decent diamonds)-4D-4S-6N(I think he has 5 Diamonds)

Playing KERI:
1D-1S-1N-2H-2S-4N-P
There is not even a good bid with this hand over a natural 1D opener, and thats a problem.

[joshs]

joshs, on Feb 13 2006, 06:02 PM, said:

Here is a hand from the Albuquerque regional:
Partner: Kx xxx Axxxx AQx
Me: AQJxx AJx KQT xx

Our Auction:
1D(Precision)-1S-1N(Good 13-15)-2D(ART GF)-3D(Reasonable 5 card suit)-4D(I like that suit)-4S(Hx, non-forcing)-6N(A club lead might hurt us, or it might give us our 12'th trick, let them guess)
Partner was min, but both suits came in and we picked up a good swing. Slam was a very good contract.

Playing x-fer checkback then waiting 2N:
1D-1S-1N-2H-2S-2N-3D(4 strong diamonds or 5 decent diamonds)-4D-4S-6N(I think he has 5 Diamonds)

Playing KERI:
1D-1S-1N-2H-2S-4N-P
There is not even a good bid with this hand over a natural 1D opener, and thats a problem.

Maybe play x-fer then partner's suit = 4+ cards
x-fer then new suit= 5 cards.

And x-fer then 2N = 3 cards in partner's suit or 4 cards in a new suit, with slightly modified steps.

Or instead we can turn the 2C-2D-2N sequence into something useful. Its possible that that is the way to bid these hands:
Play 1m-1M-1N-2C-2D-2N:
3C= 5 card minor then 3D asks for 3 card support and 3M asks for Hx in support
Over 3C-3D: 3M=3 card support, 3OM=5-4 in minors, 3N=others
3D = 4333 hand
3M = 3 card support, 4432 type hand (RARE since didn't raise)
3OM= 4432 hand with Hx in major
3N=4432 without Hx (or possibly 4441 with a stiff spade)

Anyway, just some random thoughts.

[awm]

joshs, on Feb 13 2006, 06:02 PM, said:

Here is a hand from the Albuquerque regional:
Partner: Kx xxx Axxxx AQx
Me: AQJxx AJx KQT xx

Our Auction:
1D(Precision)-1S-1N(Good 13-15)-2D(ART GF)-3D(Reasonable 5 card suit)-4D(I like that suit)-4S(Hx, non-forcing)-6N(A club lead might hurt us, or it might give us our 12'th trick, let them guess)
Partner was min, but both suits came in and we picked up a good swing. Slam was a very good contract.

Playing x-fer checkback then waiting 2N:
1D-1S-1N-2H-2S-2N-3D(4 strong diamonds or 5 decent diamonds)-4D-4S-6N(I think he has 5 Diamonds)

Playing KERI:
1D-1S-1N-2H-2S-4N-P
There is not even a good bid with this hand over a natural 1D opener, and thats a problem.

This hand will be hard under any methods that rely on responder describing instead of opener describing. Obviously a "2♦ GF relay" is ideal here. However, my feeling is that it's usually better for the shapely hand to describe, and that's not opener.

Perhaps the "best" notrump system would always have available a way for responder to:

(1) Describe his distribution to opener.
(2) Relay out opener's distribution.

Then responder could pick (1) whenever he has a shapely hand and (2) when he is balanced. However, there don't seem to be enough available calls to really implement both of these and still have a sufficient set of signoff and invitational sequences. Given the choice I'll go with (1).

However, it's possible to fix this partially using K-G checkback:

1♦ - 1♠ - 1NT - 2♣(1) - 2♦(2) - 2NT(3)

(1) Puppet to 2♦
(2) Forced
(3) Balanced game force.

Now according to the Keri notes (and K-G agrees here):

3♣ = no 5cM, not 4333
3♦ = any 4333
3♥ = 5-card hearts
3♠ = 5-card spades and 3 hearts
3NT = 5-card spades and 2 hearts

However, given the 1♦ opening I don't think it makes much sense for opener to have a five card major. An easy change:

3♣ = some 4432 shape, without 4-card support for partner's major
3♦ = 4333 (without 4-card support for partner's major)
3♥ = 5 clubs (3♠ is checkback for 3-card major support)
3♠ = 5 diamonds and 3-card support for partner
3N = 5 diamonds and 2-card support for partner

This gives away a lot of information, so if responder just wants a straight "choice of games" with a 5332 type pattern he can transfer followed by 3NT.

With this modification in place:

1♦(1) - 1♠ - 1NT(2) - 2♣(3) - 2♦(4) - 2NT(5) - 3N(6) - 4♦(7)-4♠(8)...

(1) Precision, could be short.
(2) 13-15 balanced, without 4 spades
(3) Puppet to 2♦.
(4) Forced.
(5) Balanced game force. Probably some sort of slam interest (else 2♥ transfer...3NT)
(6) 2-3-5-3 precisely
(7) Suggest a diamond slam.
(8) Cuebid of cheapest control. Now we are off to the races.

[awm]

joshs, on Feb 13 2006, 05:12 PM, said:

Adam, In example 5 how would you bid the 5422 with 4 in the other minor? You want to play 4M opposite a 4333 hand. Hence I didn't think that the sequence you have actually showed 4 card support for partner, at least not without discussion. Now there are some idle sequences that can be used to solve this problem (2C-2D-3N offers a choice, x-fer to the major and then 3N orders partner to show support) although there are still issues with mild slam tries (requires a max and a fit for slam).

BTW, tell us more about the re-arangement of the x-fer splinters?

With 4 diamonds, transfer followed by 3♣ (showing exactly 4♦ and 5+ in the major). Now opener's 3♦ is a relay for shortage (high/low/equal NF/equal substantial extras). Note that these "relays" are not mandatory, and in fact opener will usually refuse to relay when holding a good fit for the major suit transferred to. In these auctions there's no way opener will try bidding 4-minor on three cards opposite a known 4-card suit rather than try 4-major in a known 5-3 fit. If the minor suit slam on a 4-3 plays better than the 5-3 major fit, well, so be it.

As to the re-arrangement of splinters, the idea is that the minors should be symmetric with each other. In "normal" Keri you can make a diamond splinter (via 3♣) with all sorts of hand patterns such as 3-4-1-5, 4-3-1-5, and even 3-3-1-6 in addition to the "normal" 4-4-1-4. However, the club splinter (3♠) needs to be very precise shape because you have no room to explore for major suit fits. K-G removes this asymmetry by having splinter bids always go through bidding the "other suit of same rank" so 3♣ is short diamonds and 3♦ is short clubs. In both cases you can explore for a fit in a major without bypassing 3NT. The major suit splinters always promise exactly four cards in the major bid; hands with 3-1-4-5 shape (for example) go via a transfer to the five card minor followed by bidding the major suit fragment (this shows 3-1 or 3-0 in the majors with three cards in the major bid and 5+ cards in the minor transferred to).

Another K-G difference is, as you mentioned, the followups after a major suit invite. After 1NT-2♣-2♦-2♥ we play:

Pass = min with 3-4♥, if 4♥ min will be pretty bad
2♠ = 4+♠, if minimum will have doubleton heart only, if max may have 3♥ (but not 4)
2NT = min with 2♥ and 2-3♠
3♣ = max with 2♥ and 2-3♠; this is technically non-forcing.
3♦ = max with 3♥ and 2-3♠; now 3♠ is "are you 4333"
3♥ = good minimum with 4(+)♥
3NT = 3433 exactly maximum values (choice of games)

When compared to your followups, we lose the ability to stop in 2♠ when opener has 4+, but we gain the ability to bail in 3♣/♦ with a weakish 5-5 invite after we discover opener has a maximum but with no major suit fit.

[Sigi_BC84]

Wow, thanks for all the comprehensive info, I didn't expect that much. Really appreciated.

FYI, this is what we are playing after 1x-1M-1NT at the moment:

2♣ puppet to 2♦, then:
..pass..weak with longer ♦
..2M....inv+ with 5M, no side suit
..2♥....inv. with 5♠/4♥ (if first response was ♠)
..2NT..nat, balanced inv with 4M
..3x....inv with 5M/5x
..3M...inv with 6M
..3NT..5M, choice of games
..4x....6M, GF+, chicane splinter

So most invitational sequences go via 2♣ puppet.

2NT puppet to 3♣, then:
..pass..weak with longer ♣
..3♦..GF+ with supp for opened m (4M)
..3♥..GF+ with 5♠/4♥ (if first response was ♠)

The direct bids:
..2♦....exactly inv. with longer ♦ (because of Walsh treatment)
..2M..any weak hand with 5+M
..2♥....nat (if first response was ♠)
..3x....5/5 distribution, GF+
..3M....6+M, no shortness, slam interest
..4x....singleton splinter, 6+M, GF+

I hope I got it all right.

--Sigi

[joshs]

Adam, I wasn't claiming that KERI was a bad base for a method. In fact its a very good base (I agree that in general, its best to have the NT bidder who is balanced find out responder's shape and then evaluate his honors). I was just commenting that you need to make various modifications for it to be really effecient, since there are a lot of idle sequences and some difficult hand types without them. I wasn't giving my example's initially to show the superiority of one x-fer checkback scheme vs another, just to give examples of how x-fer checkback schemes work in general, since I was asked about x-fer checkback schemes (other than KERI). The examples were just cut and pasted out of one of my sets of system notes (I switched some 1D precision openings to the correct minor).

I also think that:
a. natural minors then 1N
b. natural minors then 1N but will bypass 1S even over 1H (e.g. most weak NT schemes)
c. non-natural minors
etc.


have some subtle differences. For instance, Eric Leong's orginal suggestion of 2S over the 1N rebid as a 4+ card raise of opener's minor doesn't make much sense playing c, but is pretty good playing a or b (but has a problem with my example 5332 responding hand). I think in Leong's method that hand can x-fer and then bid 3D showing a 3 card raise. That sequence is usually 5431 with 4 in the other minor, but perhaps can be this hand type also. Also, x-fer to 2M and then bid 3 OM can be giving an ART meaning. Lots of possibilities....

But the key elements are always the same:
responder should be able to:
a. show a singleton (or imply one by bidding around it if opener's doesn't show that suit)
b. generally show shape, so opener can evaluate his honors
c. have some way of bidding strong balanced hands with mild slam interest and 3/4 cards in partner's suit.