[Fluffy]

How do you evaluate stiff honnors after strong club opening?

The difference between standard bidding and 1♣ opening is, if you are giving a positive response, you are close to slam investigation, and if you have some 4441 or 5431, partner will know your distribution and your range below 3NT, and will place your honnors in the other suits more likelly. Also you have to put your hand on a close range (8-10 or 11+ for me), before knowing anything about partner except that he is storng.

So what do you think a stiff queen, King or Ace is worth? (The jack is worth zero for sure).

[olien]

Well, my partner and I use modified relay points to account for stiff honors. The method has been used by Fredin-Bertheau and is called "ZZ points" and they work like this:

A = 3 points (always)
K = 2 points, unless singleton, then just 1 point
Q = 1 point, unless singleton, then 0 points

Has worked well so far. Has kept us out of some poor slams because of the systemic downgrade. Only once did we miss a slam because the singleton honor turned out to be a working card for partner (stiff Q opposite KJTxx which provided necessary discards to make 6).

[Mbodell]

I don't claim it is best, but I don't find a problem with counting them for full value. I expect partner to be able to find out what he needs to know about my hand through asking me. My responses are usually shape based, rather than point based, so as long as it is a positive it shouldn't matter. And then either I'm relaying and partner can find specific cards or else he has specific control asking bids. Often the stiff honors are working as they will help fill in partner's suit.

[shevek]

olien, on 2011-November-16, 18:50, said:

Well, my partner and I use modified relay points to account for stiff honors. The method has been used by Fredin-Bertheau and is called "ZZ points" and they work like this:

A = 3 points (always)
K = 2 points, unless singleton, then just 1 point
Q = 1 point, unless singleton, then 0 points

Has worked well so far. Has kept us out of some poor slams because of the systemic downgrade. Only once did we miss a slam because the singleton honor turned out to be a working card for partner (stiff Q opposite KJTxx which provided necessary discards to make 6).

Well, as devised by Dave Cliff and others (who also invented DCB) in 1968.
Above is his usage and ouras for QPts. When looking at kontrols, K=1 even if stiff.

[Cascade]

This isn't a full answer but hopefully it is interesting.

I did some simulations with a 4441 hand with 8-10 hcp opposite a 16-20 hcp hand with no major fit. I also restricted the strong hand to not have too many diamonds (less than six) or to have good or long clubs so that 3NT would often be the typical contract.

I counted the tricks won in 3NT. The tricks were determined double dummy.

I repeated the exercise with the 4441 hand having any singleton, stiff ace, stiff king, stiff queen, stiff jack and small singleton.

The results were as follows:

Any singleton average 9.305 tricks

Stiff ace average 9.189 tricks

Stiff king average 9.194 tricks

Stiff queen average 9.139 tricks

Stiff jack average 9.228 tricks

Stiff small average 9.381 tricks

As a crude first approximation assuming 40 hcp and 13 tricks we have 40/13 points per trick.

A stiff ace costs 0.116 tricks which is the equivalent of about 0.36 points

A stiff king costs 0.111 tricks which is the equivalent of about 0.34 points

A stiff queen cost 0.166 tricks which is the equivalent of about 0.49 points

A stiff jack cost 0.077 tricks which is the equivalent of about 0.24 points

A small singleton gains 0.076 tricks which is the equivalent of about 0.23 points.

All of these adjustments are under half a point. Alternatively using the small singleton case as the reference they are all well under a one point adjustment.

If I get organized I might do some bigger samples overnight and confirm or not the relative sizes of these approximations.

[Zelandakh]

I personally subtract a point for a stiff K, Q or J. This is probably an over-adjustment as Wayne's analysis shows but I think it alleviates some of the issue you are referring to of partner misplacing honours. In the later relays I treat a stiff king almost like a stiff queen, that is it is not included in control totals and is given as a positive for queen ask purposes.

[whereagles]

Gonzalo, when I used to play relay systems we would count those as normal hcp (and control points). Then, after relays, pard would usually manage to dig out by exclusion that the singleton was an high card.

We never got around to improve on this, but the idea would only fail systematically on the last 20 hands of the "1000 hands bidding challenge"

In case you're interested, it was stuff like J vs AKQxxx being the key to 7.

[Fluffy]

We had a hand like this on Tuesday, I had something like:

♠A10x
♥KJ
♦AJxx
♣KQ98

I opened 1 club and partner showed 11-12 4414, I relayed to know partner has ♣A, ♥A but no ♥Q, didn´t have the space to ask anything about spades or diamonds before going over 5♣, so I assumed he'd have ♠K or ♠QJ at least, but couldn't be sure.


Wayne, I was mostly worried about slam investigation in suit contracts, I supose the result will be similar anyway.

[shevek]

Fluffy, on 2011-November-17, 04:10, said:

We had a hand like this on Tuesday, I had something like:

♠A10x
♥KJ
♦AJxx
♣KQ98

I opened 1 club and partner showed 11-12 4414, I relayed to know partner has ♣A, ♥A but no ♥Q, didn´t have the space to ask anything about spades or diamonds before going over 5♣, so I assumed he'd have ♠K or ♠QJ at least, but couldn't be sure.


Wayne, I was mostly worried about slam investigation in suit contracts, I supose the result will be similar anyway.

Better to forget HCP after GF established. For slams, much better to find out QPs (or SPs/RPs as sometimes known)
If pard showed 8 QPs, 6♣ is okay. Since stiff ♦K counts as 1SP in most symmetric systems, perhaps you could settle in 4NT/5♣ opposite that.

[Fluffy]

does that find slam when partner has ♠QJ? Looks like QJ counts the same as ♦K.

[shevek]

Fluffy, on 2011-November-17, 05:04, said:

does that find slam when partner has ♠QJ? Looks like QJ counts the same as ♦K.

I'll try in our method. Here goes ...

1♣ - 1♥ =8+ with ♠s
1♠ - 1N = ♥s
2♣ - 2♦ = 3-suited
2♥ - 2N = 4-4-1-4
3♣ - 3♠ = 7 SPs
4♣ - 5♣ ♥ = y (but not AK/AQ), ♣ = y, ♠ = y, ♦ = no (positive cue) 5♣ = legal limit

East shows ♠Qxx ♥Axxx ♦x ♣Axxx, though could be
♠KQx ♥Qxxx ♦x ♣Axxx
since that is also 7SPs with same DCB responses.

With
♠xxxx ♥Axxx ♦K ♣Axxx

East bids 4♠ over 4♣ to deny a spade honour

[Cascade]

Fluffy, on 2011-November-17, 04:10, said:

Wayne, I was mostly worried about slam investigation in suit contracts, I supose the result will be similar anyway.

Not sure. I wondered about suit contracts. I just started off with no trumps. I will try a slightly stronger range, say 18-22 and balanced and add in a 4=4 fit.

[Fluffy]

Cascade, on 2011-November-17, 20:27, said:

Not sure. I wondered about suit contracts. I just started off with no trumps. I will try a slightly stronger range, say 18-22 and balanced and add in a 4=4 fit.

Thank you

[Zelandakh]

shevek, on 2011-November-17, 20:06, said:

I'll try in our method. Here goes ...

1♣ - 1♥ =8+ with ♠s
1♠ - 1N = ♥s
2♣ - 2♦ = 3-suited
2♥ - 2N = 4-4-1-4
3♣ - 3♠ = 7 SPs
4♣ - 5♣ ♥ = y (but not AK/AQ), ♣ = y, ♠ = y, ♦ = no (positive cue) 5♣ = legal limit

East shows ♠Qxxx ♥Axxx ♦x ♣Axxx, though could be
♠KQxx ♥Qxxx ♦x ♣Axxx
since that is also 7SPs with same DCB responses.

With
♠xxxx ♥Axxx ♦K ♣Axxx

East bids 4♠ over 4♣ to deny a spade honour

How about Kxxx/Qxxx/K/Axxx? It seems to me you have simply fixed the problem because of matters relating to this particular hand. Switch a few cards and you are back to the same issue.
Axxx/Kxxx/K/Qxxx versus Txx/AJ/AJxx/AK98
and
AQJx/Kxxx/x/Qxxx versus Txx/AJ/AJxx/AK98

both seem to match the given auction, for example. Perhaps for this specific issue it could be solved by using control points combined with queen-based cues, but this is generally a poor method. The 'ideal' solution is probably to use the multi-way cues that periodically get mentioned on these forums. But they are much more complicated than most are willing to work with and do not always work either.

For what it is worth Fluffy I suspect that this is something that no system solves 100% without giving up something more important elsewhere. Having a step for control points ot queen points would certainly help (it eliminates the SK hand for a start) but is not a total solution. Of course on the given hand, even if partner has the worst possible you still have a play for 12 but I understand you are speaking in general.

[Cascade]

Slightly different methodology.

Total 10000 hands Precisely 4441 8-10 hcp

Opposite 18-22 Balanced with four spades.

Average Double Dummy Tricks

11.10632876 Small singleton 7063 occurances

10.87076537 Singleton Jack 797 occurances

10.78606357 Singleton Queen 818 occurances

10.72826087 Singleton King 736 occurances

10.94539249 Singleton Ace 586 occurances

11.0241 Any singleton 10000 occurances

Compared with average (for this distribution 4441)

x 0.082228755 0.253011555
J -0.15333463 -0.471798861
Q -0.23803643 -0.732419786
K -0.29583913 -0.910274247
A -0.078707509 -0.242176949

The first column is the difference in average tricks.

The second column is a first approximation of converting this to a hcp adjustment (tricks * 40/13)

If you prefer you can compare the hands with the small singleton case

J -0.235563385 -0.724810416
Q -0.320265186 -0.985431341
K -0.378067886 -1.163285803
A -0.160936264 -0.495188505
-0.082228755 -0.253011555

The final row is for an average hand.

This is equivalent to thinking that partner is going to play you for a small singleton. How much worse is my hand given I have a singleton honour.

So a jack is nearly 3/4 hcp worse, a queen nearly 1 hcp worse, a king over 1 hcp worse, an ace around 1/2 hcp worse.

Of course these are all averages.

[Zelandakh]

Cascade, on 2011-November-18, 03:10, said:

So a jack is nearly 3/4 hcp worse, a queen nearly 1 hcp worse, a king over 1 hcp worse, an ace around 1/2 hcp worse.

It seems like the rough approximation of subtracting a point for a singleton K, Q or J is not so bad after all!

[shevek]

Zelandakh, on 2011-November-18, 02:41, said:

How about Kxxx/Qxxx/K/Axxx? It seems to me you have simply fixed the problem because of matters relating to this particular hand. Switch a few cards and you are back to the same issue.
Axxx/Kxxx/K/Qxxx versus Txx/AJ/AJxx/AK98
and
AQJx/Kxxx/x/Qxxx versus Txx/AJ/AJxx/AK98

both seem to match the given auction, for example. Perhaps for this specific issue it could be solved by using control points combined with queen-based cues, but this is generally a poor method. The 'ideal' solution is probably to use the multi-way cues that periodically get mentioned on these forums. But they are much more complicated than most are willing to work with and do not always work either.

For what it is worth Fluffy I suspect that this is something that no system solves 100% without giving up something more important elsewhere. Having a step for control points ot queen points would certainly help (it eliminates the SK hand for a start) but is not a total solution. Of course on the given hand, even if partner has the worst possible you still have a play for 12 but I understand you are speaking in general.

♠Axxx ♥Qxxx ♦K ♣Axxx

bids 4NT over clubs as ♠/♥/♣ = y, ♦ = stiff K/A (postive cueing with singleton)

Maybe lucky that simple DCB works fine here. Perhaps not so good with different West hands.
There are other solutions with different openers. like RKC in any suit after shape known. We don't use that.
We have 4♦ as a kind of trump ask, meaning "bid your bad trump suits up the line". Would be useful if opener has ♠Txx.