Auction Termination - discussion, observations

[rbforster]

I thought we'd need a thread to discuss what we think we've learned from this practice. Perhaps I will also summarize the results after a while and everyone who cares has bid.

 

My first observation is that it sucks to hold a small doubleton as relayer, whether or not you're asking with QPs or RKC. Perhaps QP is slightly better, since with a total QPs you might be able to infer a little about the holding (but A vs KQ will never be possible, and often matters). Maybe this means you should try to reverse relay on such hands if that's possible in your system, or maybe its worth thinking about some sort of suit-exclusion method.

[rbforster]

Another observation is that my methods are pretty bad at inferring the presence or absence of stiff honors (except to the point QP count requires them to add up, which is rare). I guess it only matters when you've got a strong suit opposite a stiff, which might play meaningfully better with a stiff honor as filler. Proabably too rare to justify changing, but it might be an argument for a modified QP counting rule - I.e. count all honors, or don't count stiff honors, or only count stiff A (or AK not Q), etc.

[hrothgar]

Another observation is that my methods are pretty bad at inferring the presence or absence of stiff honors (except to the point QP count requires them to add up, which is rare). I guess it only matters when you've got a strong suit opposite a stiff, which might play meaningfully better with a stiff honor as filler. Proabably too rare to justify changing, but it might be an argument for a modified QP counting rule - I.e. count all honors, or don't count stiff honors, or only count stiff A (or AK not Q), etc.

 

I've noticed a number of hands where being able to place a stiff honor would have been nice

(This might just be a function of the particular deal)

 

None-the-less, I was toying with the idea of adding a one time scan that would come right before the first Jack scan that would announce whether or not RR held any honor (AKQ) in a singleton suit

[akhare]

Another observation is that my methods are pretty bad at inferring the presence or absence of stiff honors (except to the point QP count requires them to add up, which is rare). I guess it only matters when you've got a strong suit opposite a stiff, which might play meaningfully better with a stiff honor as filler. Proabably too rare to justify changing, but it might be an argument for a modified QP counting rule - I.e. count all honors, or don't count stiff honors, or only count stiff A (or AK not Q), etc.

 

That's a very interesting comment and the optimal approach might depend on the control scan method. For example, PCB (scan all suits *once*, including singletons) is arguably better than DCB at picking stiffs.

 

Also, there's a question the question on whether stiff Ks should be counted as 2 or 1 (I think almost everyone counts A=3 regardless and ignores stiff Qs).

[rbforster]

Another observation is that on misfit hands, a lot more strength is needed for a good NT slam than I thought. My rule of thumb with QPs was that 20 was always worth trying for slam, and 19 (or rarely 18) were worth considering in a suit when you had favorable distribution / no wastage like Axx vs stiff. I think my intuition is decent on suits, but for balanced or misfits in NT, maybe I need to up my standards or consider adding ways to inquire about extra strength from partner (like jacks). I remember TOSR had several routes to 4N that were various types of quant invites based on 2 point NT ranges (like direct 4N when 3N was available, 4♦ end signal and then 4N, ask for QPs then 4N, etc).

 

In some sense, this might be sample bias, since all these hands have opener with 12 QPs. With my rules that 20 QPs combined should investigate slam and 8 QPs by responder will respect a signoff, I need to ask with 12 but that's my minimum to ask. So maybe this is why I end up getting uncomfortable at the 5 level sometimes when I wanted to just be in game.

 

Also, there's a question the question on whether stiff Ks should be counted as 2 or 1 (I think almost everyone counts A=3 regardless and ignores stiff Qs).

I count my stiff Qs as 1. Maybe I shouldn't?

[benlessard]

Only hand 3 was tough for me. This is because I had to keycard in 4C.

 

♠ JT73

♥ K53

♦ AK743

♣ 4

 

Relay Asker

♠ KQ2

♥ AQT94

♦ 82

♣ AKJ

 

but 2 keycard and K of D make 6 close to 50%

 

I think opener was too strong to give any real problems. I think if pts are 15-15 its going to be a lot more challenging.

 

The other interesting hand was

 

♠ AJ964

♥ J53

♦ AJ54

♣ 2

 

Relay Asker

♠ KQ2

♥ AQT94

♦ 82

♣ AKJ

 

I know RR doesnt have KQd and KH. but hes showned the 2 aces and some extras. These cards can only be 3 jacks and stiff Q. Aslo by asking K&D of D and getting negative replies im pretty much sure they are going to lead D. So for me slam cannot be better than 50%.

 

Also I feel we were lucky that responder didnt have a stiff KorJ of H.

[Free]

My first observation is that it sucks to hold a small doubleton as relayer, whether or not you're asking with QPs or RKC. Perhaps QP is slightly better, since with a total QPs you might be able to infer a little about the holding (but A vs KQ will never be possible, and often matters).

That's not true, during the first scan you show 1 or 2 top honors, during your second scan you show how many exactly. When you know you have 3 QP's in that suit, you'll also know which honors partner holds based on the number of honors. If it's 1, then it's the Ace. If it's 2, then it's KQ.

[rbforster]

That's not true, during the first scan you show 1 or 2 top honors, during your second scan you show how many exactly. When you know you have 3 QP's in that suit, you'll also know which honors partner holds based on the number of honors. If it's 1, then it's the Ace. If it's 2, then it's KQ.

You're right of course. The problem I was running into in the first set of hands had to do with holding no honors as relay asker, combined with counted-but-unscaanned singletons that might have been honors. Then it wasn't possible to figure out if there was a singleton honor, or more honors opposite the empty suit.

[benlessard]

A very important thing ive noticed is that QP look like an asymmetrical way to scan.. something i didnt noticed before. In Sscan it doesnt really matter how the pts are spread (evenly is just slightly tougher)

 

EX

AKQxx

AKxx

AQx

X

 

and partner

 

xxxx

Qxx

JTx

Axx

 

3D--3H ..5431--(Srkc)

3S--3Nt (14-Qs ?)

4H--5C (y+Kh but no KD ,QD ?)

5H--6S (Y but no extras)

 

switch 6 pts

 

AQxxx

Axxx

AQx

x

 

 

Kxxx

KQx

JTx

Axx

 

3D--3H ..5431--(Srkc)

3Nt--4C (30-Qs ?)

4H--4Nt (Y but no Kh---Kof D ?)

5C--5H (no, QD ?)

5Nt--6S (Y but no extras)

 

Its not really a good example ... you lost a space because its 30 not 14 and you lost a space because there is no touching sequence like QS+KH

but if you switch pts back and forth its going to be essentially the same in term of space. So its almost irrelevant if it 10-20 or 20-10 or 15-15 as long as the hand with shape is the one that is RR. I dont think this is true about QP even if you are the correct base.

 

Another thing important thing about Sscan is that the tougher case are

 

135 (RR) ---246

 

ask for 1---yes but not 2

ask for 3---yes but not 4

ask for 5---yes but not 6

ask for 7---no

 

1-3-4-6-7-9-10-11

 

the easiest cases are

123-456

 

ask for 1 ---ive got 123 but not 4

ask for 7 ---no

 

1-5-8-9

 

vice versa

456-123

 

ask for 4---456 but not 7.

 

4-8

 

This mean that if you have or miss cards in sequence (Q trump & 1stK &2ndK) than Scanning is nice if you dont have sequence you pay a penalty. Its something QP players should think into account when they are torn between QP and scan.

 

There is of course some luck like who fall on the soff but in the long run space consumed will be the same something that i dont think is true with QP even if you have the correct base.

 

 

 

Also the other thing im wondering is how you handle a stiff A if you dont scan in the suit. Is it possible that you think ive got

 

Kxxxx

AQxx

KQx

x

 

but I hold

 

Kxxxx

Axxx

Qxx

A

 

its only when you scan the 2nd round of H that youll discover the layout ?

[straube]

I also hate holding xx.

 

I like how S2 end signal worked out on bd 8. I was able to stop in 3N. I like to have 21 QPs before investigating misfits.

 

We count the stiff K as 1 and a stiff Q as 0. We can only find a stiff Q when we RKC for that suit. I think we've a wide difference of practice on stiff honors. On the one side, Adam counts the stiff K as 2 and stiff Q as 1 (full value) and scans for the stiff honor. Those using DCB seem only able to infer a stiff honor and it seems frequently don't.

[awm]

I think the relayer hand is particularly suited to RKC because three of the four most important missing cards are keycards (♠A,♥K,♦A... the fourth being ♦K). A hand with more controls and missing more secondary cards (say ♠Axx ♥AQT9x ♦Ax ♣Axx) would be absolutely awful for RKC whereas QP asks would still seem to do okay...

 

I'm just concerned that by over-focusing on one relayer hand, we may be missing the boat here.

[hrothgar]

I think the relayer hand is particularly suited to RKC because three of the four most important missing cards are keycards (♠A,♥K,♦A... the fourth being ♦K). A hand with more controls and missing more secondary cards (say ♠Axx ♥AQT9x ♦Ax ♣Axx) would be absolutely awful for RKC whereas QP asks would still seem to do okay...

 

I'm just concerned that by over-focusing on one relayer hand, we may be missing the boat here.

 

The intent is to repeat this experiment with multiple hands...

[benlessard]

Is there a frequent Moscito auction that start with base 7 or 8 ? What is the standard base of a 11-14 opening ?

 

I think the relayer hand is particularly suited to RKC because three of the four most important missing cards are keycards (♠A,♥K,♦A... the fourth being ♦K). A hand with more controls and missing more secondary cards (say ♠Axx ♥AQT9x ♦Ax ♣Axx) would be absolutely awful for RKC whereas QP asks would still seem to do okay...

 

I'm just concerned that by over-focusing on one relayer hand, we may be missing the boat here.

 

The big weakness or RKC is when your missing one keycard you dont know wich one. When I hold 4 keycard or the total count is 5 the main problem is sequencing or keycarding too high. Here ill know the perfect location of Keycards and partner might be able to shoot a bunch of cards in sequence. If we assume that responder has at least 6QP.

 

In your example lets say responder hold a

 

5143.

 

3H (S rkc)--3Nt (0)

4C (Qs ?)-- ??

 

4D No- (this is deadliest response because you have to stop..but it may be a 70% slam. if you have JT9xx and no losers elsewhere)

4H - Y but no K of clubs (mean Qxxxx,x,KQx,Qxxx) easy stop (you just need too much secondary S honnor and J of clubs)

4S - Y+KC no KD (Qxxxx,x,Qxx,KQxx) pray you dont get a D lead or that hes got JofS.

4NT - Y+KC+KD-no QC (Qxxxx,x,Kxx,Kxxx)

 

Not perfect but not dramatic.On some of them you might already be too high but i think its decent.

 

But the same (Qxxxx,x,Kxx,Kxxx)

 

QP-3Nt (7)

4C-4NT 1-(12)-(12) here QP put you in a clueless spot because there is 3 combo KKQ

 

 

However in general having aces is good for DCB because A=3 K=2 mean that aces are space consummer when you QP.

[akhare]

Having reviewed the hands again, boards 1 and 9 are very interesting. In both cases, the 6 QPs in responder's hand create the tantalizing possibility of the perfect holding (2 Aces, A + KQx opposite the small doubleton).

 

Some posters just signed off in game (very conservative IMO) and others soldiered on, though there was no guarantee of a safety net.

 

While considering these and future boards, should we have a guideline regarding how these situations should be handled (maybe say that hands with a combined total of 20 QPs and reasonable 5-level safety should keep looking)?

[Zelandakh]

None-the-less, I was toying with the idea of adding a one time scan that would come right before the first Jack scan that would announce whether or not RR held any honor (AKQ) in a singleton suit

I pointed out to straube a while back a possible solution for this, using relay breaks of 4/5NT in a similar style to the one I use for control points but adjusted for a QP method.

[straube]

Having reviewed the hands again, boards 1 and 9 are very interesting. In both cases, the 6 QPs in responder's hand create the tantalizing possibility of the perfect holding (2 Aces, A + KQx opposite the small doubleton).

 

Some posters just signed off in game (very conservative IMO) and others soldiered on, though there was no guarantee of a safety net.

 

While considering these and future boards, should we have a guideline regarding how these situations should be handled (maybe say that hands with a combined total of 20 QPs and reasonable 5-level safety should keep looking)?

 

board 1 would at best be a 50% slam with A + KQx opposite small doubleton. It's on the heart finesse.

 

I don't see how folks are getting to slam on bd 9. For it to be worthwhile, one needs the HJ, but this is the last jack to be scanned and no one finds it until 5M has been passed.

[hrothgar]

board 1 would at best be a 50% slam with A + KQx opposite small doubleton. It's on the heart finesse.

 

I don't see how folks are getting to slam on bd 9. For it to be worthwhile, one needs the HJ, but this is the last jack to be scanned and no one finds it until 5M has been passed.

 

I managed to convince myself that it was worthwhile to bypass 5M for 5N in order to ask for the Jack

[straube]

I managed to convince myself that it was worthwhile to bypass 5M for 5N in order to ask for the Jack

 

Asking about the jack at that level has committed to the small slam anyway. I mean, you have chances that partner has the jack (odds against but possible) or that they don't lead a diamond and you manage to play hearts for no more than one loser, but the odds are subjective since they depend on the opening lead.

[hrothgar]

Asking about the jack at that level has committed to the small slam anyway. I mean, you have chances that partner has the jack (odds against but possible) or that they don't lead a diamond and you manage to play hearts for no more than one loser, but the odds are subjective since they depend on the opening lead.

 

Asking about the jack committed me to 5N...

(If partner doesn't ahve the jack, he'll bid 5N)

 

This might be moot (arguably, its better to play in 6M regardless of the position of the heart jack)

[Free]

On some hands it's important which hand will play which contract imo. Especially when the strong ♣ opener has a suit without any honors, Kx+ in partner's hand isn't always enough imo.