[jdonn]

awm, on Sep 17 2008, 01:25 PM, said:

jdonn said:

That is untrue, there is a huge reason. How much that reason weighs on you relative to other ones is of course up to you.

I would like to hear what this reason is.

1 - You can EASILY get to a 4-2 diamond fit, perhaps missing an 8 card heart fit and 7 card club fit along the way.
2 - If partner rebids 2NT after 1♦ 1♠ 2♣ you will miss any heart fit.
3 - You can easily get to a 4-3 diamond fit at the three level (see 1.)

There are more reasons (maybe your opponents make something in spades!) but I think you have to admit it's a pretty foolish statement that there is no reason to open 1NT.

[mikeh]

I play a weak notrump.

I have no problem with either a pass or an offshape 1N on this hand: I would not open 1♦. The stiff A is not worth the hcp strength that we usually associate with an Ace, since it has no kickers to which it adds strength, nor is at useful, should partner hold spade cards, as it would be if it were Ax. I hope this is self-evident, but consider a simple example such as A opposite KJxxx compared to Ax opposite KJxxx.

But we might be dealt a slightly stronger hand: say add the ♦J to the hand, while keeping it 1=4=4=4. Few would support Pass now.

Say we didn't like 1N: say we held x KQxx KJxx Axxx. We'd be less inclined to open 1N with a stiff x as opposed to a stiff A, and 1♦ still risks reaching a 4-2 fit.

I think that this is just a cost of doing business: show me a system with no built-in vulnerabilities, and you are showing me a system not yet invented.

Oh, I forgot: maybe we can all switch to mini-roman, and ask wtp? I knew that there had to be a hand on which mini-roman would shine.

[awm]

han, on Sep 17 2008, 03:24 PM, said:

In the Las Vegas swiss I ended up bidding 1NT 3 hands in a row. On the first hand I had rebid 1NT with a singleton and we had a good result. On the second hand I overcalled 1NT over 1H with a singleton diamond. The opponents were clearly surprised when I ruffed the second diamond trick.

On the third hand I had some 4333 16-count and opened 1NT. The opponents completely misdefended, I suspect that they were playing me for a stiff somewhere. This is an advantage of bidding 1NT with a stiff that I had never considered before.

There is a big difference between rebidding 1NT, overcalling 1NT, and opening 1NT. I know this concept is tough for a lot of people. But the general point is that a 1NT rebid denies length in partner's suit. You can't have four cards (would've raised) and certain hands with three cards there would've raised. Since your most common holding will be two cards in partner's suit, adding a few awkward hands which have only one card is no big deal. As for overcalling, there are many hands where you have no other sensible bid except pass, because your natural opening of 1m in a mediocre suit has been taken away by the opponents' opening bid.

jdonn said:

There are more reasons (maybe your opponents make something in spades!) but I think you have to admit it's a pretty foolish statement that there is no reason to open 1NT.

Nope. There are also "reasons" to open 2NT on this hand. After all, it may talk the opponents out of a cold game. It may make it hard for them to find their sacrifice in 4♠ over our 4♥. It may be among the only ways to reach a pushy but making game for our side. It may "right side" our eventual heart contract.

Nonetheless, I think most of us would agree that there is no reason to open 2NT on this hand. In other words, while there are some "reasons" of things that might happen which are favorable due to that action, the most likely effects are so overwhelmingly negative that we wouldn't do it. Maybe a more accurate statement is: there is no reason that makes sense, from the standpoint of a rational player trying to optimize his expected or most likely results on the hand in question. And I feel the same way about opening 1NT. What is to stop partner from doubling their cold spade contract after we open 1NT? Who is to say that we are not better off in two of a minor than 1NT with our weakish hand? And don't an awful lot of people play reverse flannery (in which case you cannot miss a heart fit on this hand anyway)?

[jdonn]

You are starting to make me lol. I guess I'm just an irrational player who does not try to optimize my expected results.

[han]

han, on Sep 17 2008, 03:31 PM, said:

awm, on Sep 17 2008, 01:25 PM, said:

I would be willing to bet that most partner hands with 33(34)/(23)44 shape and Kxx or weaker spades will play better in 5m than 3nt.

OK I'm taking you up on this bet, will post results later.

I regret to say that I lost the bet.

I ran a double dummy analysis of 200 deals. In those deals south had the opening hand with the stiff ace of spades and partner had any 33(43) or (23)44 hand with 12-14 HCP and at most Kxx in spades. The results:

3NT makes on 69 hand.
The better 5m makes on 77 hands.

Whether this has anything to do with bridge is not clear. For example, most of my partners would not bid 2NT holding xx Axx Axxx AQxx, yet that was one of the 200 hands. Also, opponents don't always defend perfectly and typically they defend and lead better when you give them more information. I noticed that the lead was much more relevant against 3NT than against 5m, in 3NT the opponents almost always had to lead a spade to get the optimal result. Finally, even awm might not always get to the best 5m when partner is 4-4 in the minors, yet on those hands 5C and 5D didn't always take the same number of tricks.

As an aside: when partner has exactly Kxx in spades then 3NT makes 63% of the time and the best 5m makes 20% of the time.

So I remain convinced that showing spade shortness on the hand in question is very poor bridge.

[CSGibson]

han, on Sep 18 2008, 10:51 PM, said:

han, on Sep 17 2008, 03:31 PM, said:

awm, on Sep 17 2008, 01:25 PM, said:

I would be willing to bet that most partner hands with 33(34)/(23)44 shape and Kxx or weaker spades will play better in 5m than 3nt.

OK I'm taking you up on this bet, will post results later.

...

A shortened version of the previous post:

I wrote and ran a simulation.

I was wrong.

The simulation was flawed.

I am right.

[awm]

I will also note that partner's actual hand looks something like the hand Han says is "not a 2NT bid."

You could certainly bid:

1♦ - 2♣
3♣ - 3♦
3♥ - 4♣
5♣

on these hands, and it has the advantage that if responder really did have ♠Kxx there is a good chance he would've bid 3NT over 3♥ or maybe just responded 2NT to begin with.

[jdonn]

awm, on Sep 19 2008, 01:29 AM, said:

I will also note that partner's actual hand looks something like the hand Han says is "not a 2NT bid."

xxx AQx ATx KJxx

"looks a lot like"

xx Axx Axxx AQxx

???

Come on, it's plainly obvious the actual hand has better holdings to bid notrump in all four suits!!! (one may be debatable, three are clear). You are just trying to score cheap political-type points. Thats no less ridiculous than if he said "I would not open 1NT with Ax xx KQxxx AQJxx" and you reply "I will note that hand looks something like KJx xx KQTx AQJx"

[jdonn]

CSGibson, on Sep 19 2008, 01:10 AM, said:

A shortened version of the previous post:

I wrote and ran a simulation.

I was wrong.

The simulation was flawed.

I am right.

He revived a dead thread two days later, when no one probably even remembered what he had said, simply to admit he had been wrong in his bet. He then stated it hadn't changed his mind about the correct bid, and clearly laid out the reasons why. What exactly do you object to?? Is it so hard to believe that having an uninformative auction to 3NT is enough of an advantage over a very informative auction to 5♣/♦ to negate a 4% double-dummy advantage? Or that he was lost his bet based on weighing 3NT against the COMBINATION of 5♣ and 5♦, giving those contracts the advantage when only one was better, even if you would (randomly?) have to reach the better of the two fits?

[CSGibson]

jdonn, on Sep 18 2008, 11:45 PM, said:

CSGibson, on Sep 19 2008, 01:10 AM, said:

A shortened version of the previous post:

I wrote and ran a simulation.

I was wrong.

The simulation was flawed.

I am right.

He revived a dead thread two days later, when no one probably even remembered what he had said, simply to admit he had been wrong in his bet. He then stated it hadn't changed his mind about the correct bid, and clearly laid out the reasons why. What exactly do you object to?? Is it so hard to believe that having an uninformative auction to 3NT is enough of an advantage over a very informative auction to 5♣/♦ to negate a 4% double-dummy advantage? Or that he was lost his bet based on weighing 3NT against the COMBINATION of 5♣ and 5♦, giving those contracts the advantage when only one was better, even if you would (randomly?) have to reach the better of the two fits?

I don't find the post particularly objectionable, just slightly ironic. I probably should have just enjoyed it myself instead of enjoying the irony publicly.

[han]

Gibson, why don't you focus a little more on the substance and a little less on the form.

When Adam posted his bet, he made it very clear what the conditions were and I accepted the bet understanding those conditions. So I was wrong, not right. Now, when a simulation is as flawed as mine was I think it is right to point out the flaws, especially when I feel the simulation results suggest a call that I think is very bad in the real world.

Adam, are you actually saying that xx Axx Axxx AQxx is a 2NT bid? It seems that way since you object to my objection and state that the hand is similar to a hand you suggested 2NT with. I have difficulty believing it. Will you say anything, no matter how foolish, to come out of a discussion as a winner?

[TimG]

han, on Sep 19 2008, 12:51 AM, said:

han, on Sep 17 2008, 03:31 PM, said:

awm, on Sep 17 2008, 01:25 PM, said:

I would be willing to bet that most partner hands with 33(34)/(23)44 shape and Kxx or weaker spades will play better in 5m than 3nt.

OK I'm taking you up on this bet, will post results later.

I regret to say that I lost the bet.

I gave south the hand that started the discussion, ♠A ♥KJ75 ♦Q852 ♣QT63, and gave north a 33(34)/(23)44 12-14 count. I produced 1000 hands to fit the criteria. I found that game in NT (34.2%) and game in a minor (33.0%) were nearly identical propositions. (I think the difference will be significant once you add in the real world factor of a relatively blind opening lead after 1N-3N compared to the revealing auction to 5m.)

Some other findings:

• Game in NT made 24.3% of the time when responder had two spades and 35.9% of the time when responder had three spades.
  
  
• When responder held Kx(x) or QJx in spades, 3NT made 51.6% of the time; when responder had worse spades, 3NT made 27.4% of the time.
  
  
• When responder held Kx(x) or QJx in spades, 5m made 13.9% of the time; when responder had worse spades, 5m made 40.5% of the time.
  
  
• When responder had 12 HCP, 3NT made 22.7% of the time; when responder had 13 HCP, 3NT made 32.7% of the time; when responder had 14 HCP, 3NT made 53.5% of the time.

The middle two items suggest to me that this is a situation where an exchange of information will help the partnership land in the right spot -- the difference between making 3NT 51.6% of the time responder has spade help and 27.4% of the time responder has no spade help is not going to be overcome by putting the defenders to a blind opening lead (and defense). Of course, this is only a subset of the possible hands facing a 1444 opening 1NT, so there may be lots of responding hands where an exchange of information won't help our side and will help the defenders.

Below is the input file I used with mdealer for this simulations.

 [space] [space]predeal south SA, HKJ75, DQ852, CQT63

 [space] [space]shape_ok=shape(north, 3343 + 3334 + 2344 + 3244)
 [space] [space]hcp_ok=hcp(north)>11 && hcp(north)<15
 [space] [space]spades_ok = hcp(north, spades)<4
 [space] [space]
 [space] [space]north_ok = shape_ok && hcp_ok && spades_ok

 [space] [space]tn=tricks(north,notrump)
 [space] [space]td=tricks(south,diamonds)
 [space] [space]tc=tricks(south,clubs)
 [space] [space]
 [space] [space]diamond_fit = diamonds(north)==4
 [space] [space]club_fit = clubs(north)==4
 [space] [space]
 [space] [space]kx = hascard(north, KS)
 [space] [space]qjx = hascard(north, QS) && hascard(north, JS) && spades(north)==3
 [space] [space]bad_spades = !kx && !qjx
 [space] [space]
 [space] [space]game_nt = tn>8
 [space] [space]game_m = td>10 || tc>10


 [space] [space]condition north_ok
 [space] [space]produce 1000

 [space] [space]action
 [space] [space] [space] [space]average "Tricks in NT" (tn),
 [space] [space] [space] [space]average "Game in NT" (game_nt),
 [space] [space] [space] [space]average "Game in minor" (game_m),

 [space] [space] [space] [space]frequency "Tricks in clubs by club length" (clubs(north), 3, 4, tc, 9, 10),
 [space] [space] [space] [space]frequency "Tricks in diamonds by diamond length" (diamonds(north), 3, 4, td, 9, 10),

 [space] [space] [space] [space]frequency "Tricks in NT by spade length" (spades(north), 2, 3, tn, 7, 8),
 [space] [space] [space] [space]frequency "Tricks in NT with bad spades" (bad_spades, 0, 1, tn, 7, 8),
 [space] [space] [space] [space]frequency "Game in minor with bad spades" (bad_spades, 0, 1, game_m, 0, 1),
 [space] [space] [space] [space]
 [space] [space] [space] [space]frequency "Tricks at NT by HCP" (hcp(north), 12, 14, tn, 7, 8),

[Free]

1NT - 3NT