[gnasher]

inquiry, on 2013-June-21, 14:00, said:

I would think the same rules apply to hand in question after the 2NT opening

I wouldn't. The main reason for choosing to play 4♥ in the 2NT auction is that we can use a ruff as an entry to dummy to cash the hearts. In the 1NT auction, dummy will have more high cards, so this is less of a consideration.

[inquiry]

ok i tried a simulation using the south hand. The constraints i placed on the north hand was 5-10 hcp, 2-3 spades, 5 hearts, 2+ in each minor. i ffigured no 4 spades due to lack of stayma followed by smolen. i look at 1000 unique deals and aanalyzed them by double play using deal masterpro.

The results were close but a "big" advantage seemed to be clear at matchponts or imps to the pass.

FIRST OF THE 1000 HANDS, 75% (751) made four hearts, while 80% (803) made 3NT. There is no worrying about overtricks in those numbers, clearly your game would be more likely to make in notrump than in hearts.

At matchpoints, it is a closer decision. Here you would have to take into acount the number of times you took one or more tricks in hearts than in nortump.. for example, while 80% of the hands took 9 tricks or more at notrump, 94% took nine tricks (or more) in hearts. Let's just compare the number of hand that took exactly 10 tricks in hearts to the number of hands that took exactly 10 tricks in notrump. This iis 751 in hearts to 526 in notrump. Clearly hearts took more tricks, on average, than notrump. The problem in figuring out which is better at matchpoint is that when there is a tie, notrump wins. So one might assume (this is not 100% right) that the same hands taht took 10 tricks in notrump took at least 10 tricks in hearts. So I compared all 1000 notrump contracts to 1000 heart contracts, and found that notrump took the same number or more tricks than hearts on 565 of the 1000 hands. That means notrump was a better spot 56.5% of the time. There were times when hearts took two or three tricks but at matchpoint it is the frequency.

I admit my constraints are not ideal perhaps. I should not allow an opponent to have 8 spades for instance. But these numbers support fairly losely the numbers i quoted from bridgebrowser (This was way more deasls, and used double dummy play and defense). In the bridgebrowser study, 3NT won by about an imp and a quarter over 4H. Somehow that seems a little high for a 5% difference in making. The matchpoint difference was 54% here is is 56.5% "better". Anyway, others can try thiis hand and simulations to see what they find. In the simulations i did, there was some huge swings between the number of tricks taken in notrump and hearts on the hands.

[neilkaz]

gnasher, on 2013-June-21, 16:39, said:

I wouldn't. The main reason for choosing to play 4♥ in the 2NT auction is that we can use a ruff as an entry to dummy to cash the hearts. In the 1NT auction, dummy will have more high cards, so this is less of a consideration.

Agreed completely and I clearly think there's a big difference between a 1NT opening and 2NT opening with this hand.

[gnasher]

inquiry, on 2013-June-21, 21:47, said:

The constraints i placed on the north hand was 5-10 hcp, 2-3 spades, 5 hearts, 2+ in each minor.

Suppose that partner opened 2NT and you had one of these hands:
x Qxxxx xxx Kxxx
Kxx Jxxxx xxx xx
x Jxxxx xx QJxxx

What would you do?

[MickyB]

Thanks Ben. Looks to me like 4H would come out on top once hcp+shape requirements were relaxed as per Andy's post.

[PhilKing]

MickyB, on 2013-June-22, 04:02, said:

Thanks Ben. Looks to me like 4H would come out on top once hcp+shape requirements were relaxed as per Andy's post.

I doubt it. Partner won't have a stiff spade that often.

Also I think the upper limit should be lower, which will accentuate the difference (4-8 hcp shapes: 1-3, 5, 0-5, 0-5).

I don't think an automated computer sim is best - you have to look through the results by hand, since when partner is impleted, he should often just insist on hearts if he is 5-5 (QJTxx JTxxx) and sometimes even when he is 5-4.

[gnasher]

PhilKing, on 2013-June-22, 05:38, said:

I don't think an automated computer sim is best - you have to look through the results by hand, since when partner is impleted, he should often just insist on hearts if he is 5-5 (QJTxx JTxxx) and sometimes even when he is 5-4.

Even with xx QJ10xx xxx Jxx I'd insist on hearts.

[y66]

Impleted? As in filled, replete (with something)? The last recorded use of this word in the Oxford Dictionary is from 1694:

J. T. in Philos. Trans. (Royal Soc.) 18 111, I found..its Vesicles impleat with a grumous Blood.

Etymology: < Latin implētus, past participle of implēre to fill up, < im- (im- prefix1) + *plēre to fill.

[PhilKing]

.

[PhilKing]

y66, on 2013-June-22, 13:02, said:

Impleted? As in filled, replete (with something)? The last recorded use of this word in the Oxford Dictionary is from 1694:

J. T. in Philos. Trans. (Royal Soc.) 18 111, I found..its Vesicles impleat with a grumous Blood.

Etymology: < Latin implētus, past participle of implēre to fill up, < im- (im- prefix1) + *plēre to fill.

You are correct - filled up with good spot cards. Andrew Robson popularised it as a bridge adjective to differentiate ♠T9♥QT987♦JT98♣T9 from, say,♠32♥Q5432♦J432♣32. The former hand has impletion.

There are not many obvious modern words that really fit the bill.

[han]

I did a double dummy simulation somewhat similar to Ben's, but allowing partner to hold a singleton and taking note of the fact that RHO is a passed hand. I merely counted the number of IMPs and MPs won by passing compared to bidding 4H, I did not look at how many times contracts were made.

The simulation says that bidding 4H will be a small winner at IMPs, with 0.377 IMPs per hand. For 4000 hands the standard deviation is less than 0.1. At MPs it is too close to tell (the difference is less than one standard deviation after running 4000 hands).

I think that PhilKing's objection is, although valid, not an objection to automated simulations but to lazily done automated simulations, such as mine. I also think that the effect of disallowing highly impleted hands would be quite small.

[Winstonm]

Perhaps the question is wrong - perhaps the hand that needs to change bids is responder. If responder holds no apparent entry to his long major suit, perhaps the correct thing to do is transfer and then raise to game, ignoring 3NT?

I have no idea what is right or wrong in this debate - I would pass 3NT but only because that was what I was taught years and years ago to do. I am open to a better methods supported by evidence from simulations.

When the facts change, I change my mind.

[nige1]

Winstonm, on 2013-June-24, 08:05, said:

Perhaps the question is wrong - perhaps the hand that needs to change bids is responder. If responder holds no apparent entry to his long major suit, perhaps the correct thing to do is transfer and then raise to game, ignoring 3NT?

If you impulsively transfer at the three-level and importunately raise, you may create the impression of a better hand. Perhaps this implies, for reasons imputed to Gnasher, that, with a PhilKing impleted hand, at imp scoring, it's imperative to transfer at the four level?
e.g. over 2N bid 4♦ with ♠ J x ♥ Q J T 9 x ♦ J T x ♣ J T x ?

[Zelandakh]

PhilKing, on 2013-June-22, 13:53, said:

There are not many obvious modern words that really fit the bill.

Texture?

[helene_t]

nige1, on 2013-June-25, 13:20, said:

If you impulsively transfer at the three-level and importunately raise,[...] Perhaps this implies, for reasons imputed to Gnasher, that, with a PhilKing impleted hand, at imp scoring, it's imperative to [...]

You should quote your source: http://www.elsewhere.org/pomo/

[dake50]

I would pass. Let me propose an alternative question.
I am not up to running a simulation today, but I did do a very quick bridgebrowser check for 2NT-P-3D-P-3NT where the 2NT bidder had a 4432 with specifically AK doubleton in hearts and responder was balanced or semi-balanced with exactly five hearts. Passing 3NT was the clear winner at both matchpoints and imps. I did not examine the hands, just average results.. no double dummy play was studied... just how people faired. The sample size was small, 77 deals -- each deal played 16 times (a total of 1136 different auctions). In this study, the scores for 3NT and 4♥ are shown. Of course, other people reached different contracts on the same hands... slams, partscores, even other suits. 3NT was a clear winner, averaging in this LIMITED study +1.46 imps, and 54.2% matchpoints. 4H averaged +0.26 imps and 44.7% matchpoints. The sample size is too small, plus, with no constraints but the auction, a lot of people way overbid either not stopping lower when they should or going too high. In fact, less than 1/2 the auctions ended in either 3NT or 4♥, so it is not clear how useful this is. I did look at 2N-3h-3S-bid game and it followed the same trend and the same preference for 3NT to be right.

**** Did you correct for SA+CA+HAK+DKQJx = 8 controls?
The very point of correcting to 4H is the tops/controls.
Why sim POOR(3,4) controls???? Instead or 8 controls??

[Winstonm]

PhilKing, on 2013-June-22, 05:38, said:

I doubt it. Partner won't have a stiff spade that often.

Also I think the upper limit should be lower, which will accentuate the difference (4-8 hcp shapes: 1-3, 5, 0-5, 0-5).

I don't think an automated computer sim is best - you have to look through the results by hand, since when partner is impleted, he should often just insist on hearts if he is 5-5 (QJTxx JTxxx) and sometimes even when he is 5-4.

I think the sim needs to be quite specific: QJxxx as hearts and another random Q, but not more than random Q and J, not in the same suit. The issue is how often is 4H superior when dummy has no apparent entry to the long suit.

[wank]

there are other refinements you can make to the sim that would favour returning to hearts. would you transfer with 5 low and a balanced 6 count? i know i wouldn't. i'd content myself with stayman. there is a bias towards partner's suit being chunky just because he bothered to show them.

[jdeegan]

The game is not all that hard. You just bid 3NT. Playing 20-21 HCP 2NT open, you might just possibly consider an upgrade and open 2♣. If you got a bad result and it worries you, then maybe a strong club system is in your future? What did partner have?