[straube]

I'm going to run a few hands and see if anyone is interested in looking at PCB options with me. I'm generating off of BBO and open to suggestions about better ways of generating/restricting hands. Corrections or other DCB/PCB methods welcome.


South is captain. Let's say North shows pattern and 7 QPs at 3N.


KT93 J32 AK64 84
AQ2 A7654 J AKQT

 DinDIP, on 2017-November-05, 05:33, said:

JVCB
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-5♥ (DK+SK+DA, no DJ)

Ulf's parity slam bidding
(This assumes 1♣ was 17+ and teller showed shape and 10-12HCP.)
4♣-4♥ (odd A + even K, could have Q as QP count not known)
4♠-4N (even Q, but slam still worth bidding if SK+HK and therefore another relay reasonable but riskier than using JVCB)
5♣-5♥ (odd parity in S, even parity in D)

Kleine Fugue (the less complex of the complex Swedish methods)
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-4N (4-6 ranking points [to be explained in main thread when I get to scan and convert notes])
5♣-5♦ (5 ranking points so HA+SK+CK OR DA+SK+DK)

Grande Fugue (the more complex of the complex Swedish methods)
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-4N (One of a. DA + ((SK+DK) OR (HK+CK)); b. SA + ((DK+CK) OR (SK+HK)); OR c. (HA or CA) + ((DK+HK) OR (SK+CK)), so known to be DA+DK+SK

And, for the sake of completeness, parity K ask followed by DCB
Standard DCB
4♣-4N (even Ks, S, D no H honour)
Shevek's DCB (where stop = 0 or 4+QP)
4♣-4♠ (even Ks, 1-3QP in S, 0 or AKQ/AK/AQ in D)

David

 awm, on 2017-October-20, 04:01, said:

4♣ - 4♥ (odd spades, even diamonds)

Now relayer knows that there are the following honor combinations:

♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK

Clearly the "cards are not placed." But none of these makes for a really good slam; perhaps if partner has the jack AND ten of hearts you are okay, but this is unlikely and also not something you can find out at a reasonable level. So I'd just pass 4♥.

IMprecision to card placement would be...
4C 4H odd spades, even diamonds
4S 4N even heart

 sieong, on 2017-October-21, 23:02, said:

EPCB.
4♣ - 4n (odd ♠, even all others)

Honor structure resolved.

 yunling, on 2017-October-18, 21:30, said:

4♦-4♥(ask 1&3, 1/4)
P

cards are not placed but you are sure to pass 4♥.

1) With rules of short to long (and stop with odd for singleton or doubleton), king parity last

4C-4H (even club, even heart)
4S-5C (odd spade, even diamond)
Cards placed

2) With rules of short to long (and stop with odd for singleton or doubleton), king parity first

4C-4S (even Ks, even club, even heart)
Cards placed.

3) With rules of long to short (and stop for singleton only), K parity last

4C-4H (odd spade, even diamond)
4S-4N (even heart)
Cards placed.

4) With rules of long to short (and stop for singleton only), K parity first

4C-4S (even Ks, odd spade, even diamond)
Cards placed

[foobar]

Think that with 20+ combined QPs, most methods will place cards easily. It's the 18-19 QP (slam with a 9+ card trump fit and fitting honors) that probably merit most exploration. I would guess to limit opener to 12 QPs (but ensure a shapely 9-card trump fit) to make things worthwhile.

Unfortunately, I can't think of an easy way to simulate card placement with PCB/DCB off the top of my head -- hrothgar -- any dealer script suggestions?

[Zelandakh]

 foobar, on 2017-October-17, 00:31, said:

Think that with 20+ combined QPs, most methods will place cards easily.

Agree with this completely. For example, replacing 7QPs with 4CPs and using simple DCBs gives:-

4♣ - 4♠ = spade control, diamond control, no heart control - cards placed

If the basic DCB method finds everything in one response, the hand is probably too simple to be useful. In the end what you want to measure is which method provides the required information below the safety threshold most often, which is not necessarily the same thing as the method that comes up with the solution at the lowest level on average. A method that is always good is therefore probably better than one that is usually super but sometimes very bad.

[straube]

 Zelandakh, on 2017-October-18, 00:09, said:

In the end what you want to measure is which method provides the required information below the safety threshold most often, which is not necessarily the same thing as the method that comes up with the solution at the lowest level on average. A method that is always good is therefore probably better than one that is usually super but sometimes very bad.

I agree but isn't this also a bit subjective? I mean one may strongly have a suspicion that partner has the right sort of hand with one method early but not know for sure until too late while another method might give you no idea early, usually finish before the other but sometimes be too late.

I invited you on another post to test your own methods. Another way you could participate is to assign grades (however you want to do that) for each method.

[foobar]

 straube, on 2017-October-16, 21:01, said:

I'm going to run a few hands and see if anyone is interested in looking at PCB options with me. I'm generating off of BBO and open to suggestions about better ways of generating/restricting hands. Corrections or other DCB/PCB methods welcome.

KT93 J32 AK64 84
AQ2 A7654 J AKQT

With IMP (long to short, skip with odd-parity unless singleton, K-parity in suit with a single honour):

4C...4H (relay; odd spade, even diamond); S(K) is obvious at this point, D(AK) or D(AQ) + H(Q)
4S...4N (relay; even heart, so perforce D(AK))

K-parity is perfunctory at this point, so we can pretty much place the final contract (which is probably pass?).

FWIW, RKC-H might be a better method on this hand because the only slam worth exploring is 6H.

[yunling]

I'm testing a quite different method here.
So after QP is shown:
step 1 asks (combined)number of honors in suits rank 1&2
step 2 asks (combined)number of honors in suits rank 1&3

and the response are:
+1 1/4 honors
+2 0/3/6 honors
+3 2/5 honors, 1/4 honors in suits 2&3
+4 2/5 honors, 0/3/6 honors in suits 2&3
+5 2/5 honors, 2/5 honors in suits 2&3, odd in suit 4
+6 2/5 honors, 2/5 honors in suits 2&3, even in suit 4, odd K
etc.

So theoratically cards may not be 100% solvable under this method, but it can quickly give you an idea of the concentration of high cards on a lower level(so can sign off early when slam is not possible)

There is a lot to improve in this method but I don't have a clear idea so I'll run through all these hands to find the problems.

For this hand it goes
4♦-4♥(ask 1&3, 1/4)
P

cards are not placed but you are sure to pass 4♥.

[awm]

I'm not sure you're really evaluating this right.

Supposing long-to-short, you get:

4♣ - 4♥ (odd spades, even diamonds)

Now relayer knows that there are the following honor combinations:

♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK

Clearly the "cards are not placed." But none of these makes for a really good slam; perhaps if partner has the jack AND ten of hearts you are okay, but this is unlikely and also not something you can find out at a reasonable level. So I'd just pass 4♥.

This seems like a big win over short-to-long, where you have:

4♣ - 4♥ (even clubs, even hearts):

♠K + ♥KQ + ♦K
♠K + ♥K + ♦A
♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK
♥KQ + ♦AQ
♥K + ♦AK
♥Q + ♦AKQ

This is a LOT more possibilities, and opposite some of them (notably the first and second examples) slam is pretty good regardless of jacks.

So you're basically forced to relay again and get to the five level, which can fail on a bad day.

[Zelandakh]

 awm, on 2017-October-20, 04:01, said:

4♣ - 4♥ (even clubs, even hearts):

♠K + ♥KQ + ♦K
♠K + ♥K + ♦A
♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK
♥KQ + ♦AQ
♥K + ♦AK
♥Q + ♦AKQ

You obviously understand the PCB method a lot better than me but I am a little confused here. I would have thought that 2, 3, 4, 7 and 8 were not possible. What does "even hearts" mean here if not restricting us to jxx or KQx?

♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK

versus

♠K + ♥KQ + ♦K
♠K + ♦AK
♥KQ + ♦AQ

...does not seem so bad, particularly as this case (having AKQ in a suit) is about as bad as it gets.

[awm]

 Zelandakh, on 2017-October-20, 07:43, said:

You obviously understand the PCB method a lot better than me but I am a little confused here. I would have thought that 2, 3, 4, 7 and 8 were not possible. What does "even hearts" mean here if not restricting us to jxx or KQx?

♠K + ♥K + ♦KQ
♠K + ♥Q + ♦AQ
♠K + ♦AK

versus

♠K + ♥KQ + ♦K
♠K + ♦AK
♥KQ + ♦AQ

...does not seem so bad, particularly as this case (having AKQ in a suit) is about as bad as it gets.

You're right about the number of hands here, but being unsure about the heart holding here seems problematic (you need two heart honors or ♥K + ♦A to have really good chances).

The general point in favor of "long-to-short" is that there are two times the honor holding becomes really important:

1. Honors that will be in the trump suit. These tend to be more important than side-suit honors because you can't prevent losers by pitching in the trump suit. Obviously any suit can be trumps, but describer's LONGER suits are more likely choices.
2. Honors that are opposite partner's shortness. These are less important than other honors for the obvious reason. Again, any suit can be partner's shortness, but describer's LONGER suits are more likely choices.

If the trump suit is solid and relayer has no shortness, than usually the number of QP is enough to make a slam decision. So locating cards is not necessarily that important.

[sieong]

EPCB.
4♣ - 4n (odd ♠, even all others)

Honor structure resolved.

[nullve]

 straube, on 2017-October-16, 21:01, said:

KT93 J32 AK64 84
AQ2 A7654 J AKQT

Auction, with some ambiguity left out:

(...)
3N-4♣ (11-13 hcp (say); relay)
4♦-4♥ (even # of kings; relay)
4♠-4N (even # of queens; relay)
5♥-? (♠K, ♦K, no ♠J)

This post has been edited by nullve: 2017-November-02, 07:47

[DinDIP]

 straube, on 2017-October-16, 21:01, said:

KT93 J32 AK64 84
AQ2 A7654 J AKQT

JVCB
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-5♥ (DK+SK+DA, no DJ)

Ulf's parity slam bidding
(This assumes 1♣ was 17+ and teller showed shape and 10-12HCP.)
4♣-4♥ (odd A + even K, could have Q as QP count not known)
4♠-4N (even Q, but slam still worth bidding if SK+HK and therefore another relay reasonable but riskier than using JVCB)
5♣-5♥ (odd parity in S, even parity in D)

Kleine Fugue (the less complex of the complex Swedish methods)
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-4N (4-6 ranking points [to be explained in main thread when I get to scan and convert notes])
5♣-5♦ (5 ranking points so HA+SK+CK OR DA+SK+DK)

Grande Fugue (the more complex of the complex Swedish methods)
4♣-4♥ (odd A + even K therefore known to be DA and KK and no Q, making slam worth bidding if SK+HK and therefore another relay reasonable)
4♠-4N (One of a. DA + ((SK+DK) OR (HK+CK)); b. SA + ((DK+CK) OR (SK+HK)); OR c. (HA or CA) + ((DK+HK) OR (SK+CK)), so known to be DA+DK+SK

And, for the sake of completeness, parity K ask followed by DCB
Standard DCB
4♣-4N (even Ks, S, D no H honour)
Shevek's DCB (where stop = 0 or 4+QP)
4♣-4♠ (even Ks, 1-3QP in S, 0 or AKQ/AK/AQ in D)

David

[straube]

Thanks a lot for posting this. I've copied and pasted these auctions as well as others offered by contributors here at the top of the thread. If you have time to bid any more of the hands, I'll do the same.

[DinDIP]

 straube, on 2017-November-05, 10:54, said:

Thanks a lot for posting this. I've copied and pasted these auctions as well as others offered by contributors here at the top of the thread. If you have time to bid any more of the hands, I'll do the same.

I will get around to the others: might take a little while as time limited at present. Also note that I edited the last two auctions: using so many different suit orders for the complex stuff confused me when it came to the methods I'd used for years!