[Antrax]

What is the accepted figure today for the amount of HCPs needed to make an NT/major-suit game? I was taught Goren originally said 26, but then declarer play improved so they lowered it to 25, and because of no shortage of entries, 12 opposite 12 should also suffice. Is that more or less sane? Is there a consensus?
(inspired by this: http://www.bridgebas...post__p__657322)

[nigel_k]

In notrumps, assuming balanced hands and average honour distributions and intermediate cards, you generally want to be in game with 25 and don't mind either way with 24. But sometimes if the bidding is at 1NT and you know the combined range is 23-25, you may just pass because 25 is less likely and you don't want to put your sure plus score at risk.

For major suit games, there will usually be other factors to consider beside high card points. But if you have two balanced hands with eight combined trumps and all other things are equal, probably about the same high card strength as above is required. Maybe a little less in a 4-4 fit.

[Cyberyeti]

The above is about right, old style Acol bids game with 15 opposite 9 but not 16 opposite 8. Old style precision does the reverse. I think 25 is the normal answer, but 24 is no disaster.

[Siegmund]

I might phrase it slightly differently: 25 is a reason to be in 3NT rather than 1NT. 24 is a reason to pass 1NT, but to be in 3NT rather than 2NT.

[JLOGIC]

I think that generally applies to imps.

In MP I'd often rather play 2N than 3N with 24 HCP if there was no remarkable feature.

In suit contracts degree of fit matters a lot, it is definitely more complicated than balanced NT games.

[frank0]

I had done a short DD simulation on how often 24HCP 3NT can be made, for your reference, I post the result here.

On each hand NS has exactly 24HCP total and both hands are 4432 or 4333. South is always the declarer of NT contract and the card play is analyzed by DD solver. There were 100 sample hands.

Result:
Tricks Frequency(# of hands)
11......1
10......8
9.......26
8.......39
7.......21
6.......5

If you're red in IMP 3NT has slightly advantage over 2NT(0.38IMP/board) but 1NT is the best if it's possible to stay there.

[Antrax]

Why no 5-3-3-2 hands, and why must both hands be balanced? As responder I raised to 3NT with semi-balanced hands even as a beginner.

[phil_20686]

 Antrax, on 2012-August-08, 03:42, said:

Why no 5-3-3-2 hands, and why must both hands be balanced? As responder I raised to 3NT with semi-balanced hands even as a beginner.

Not to mention that not all hands with the same points are equal. Lots of the hands that dont make will be when both players have ugly 4333 hands and would choose to be less aggressive in RL.

And double dummy generally favours the defence here over real life.

[Zelandakh]

A simple IMP rule for Walruses - with 24 or more bid game, with 24 or less do not bid game. Obviously the reality should include a little judgement.

[P_Marlowe]

 frank0, on 2012-August-08, 02:38, said:

I had done a short DD simulation on how often 24HCP 3NT can be made, for your reference, I post the result here.

On each hand NS has exactly 24HCP total and both hands are 4432 or 4333. South is always the declarer of NT contract and the card play is analyzed by DD solver. There were 100 sample hands.

Result:
Tricks Frequency(# of hands)
11......1
10......8
9.......26
8.......39
7.......21
6.......5

If you're red in IMP 3NT has slightly advantage over 2NT(0.38IMP/board) but 1NT is the best if it's possible to stay there.

Which means, if you have at most 24HCP dont invite, if you have at least 24HCP invite.
Which is done with standard agreements.

Ok, ..., you usually invite, if you have at least 23, to see, if 23 is the cut off, you would need to rerun the simulation
with 23.

With kind regards
Marlowe

[frank0]

 Antrax, on 2012-August-08, 03:42, said:

Why no 5-3-3-2 hands, and why must both hands be balanced? As responder I raised to 3NT with semi-balanced hands even as a beginner.

Because 4333/4432 are the shape people usually use HCP to evaluate without additional adjustment. With a 5+ suit or shortness people start to upgrade/downgrade the hand depends on the degree of fit.

 P_Marlowe, on 2012-August-08, 07:17, said:

Which means, if you have at most 24HCP dont invite, if you have at least 24HCP invite.
Which is done with standard agreements.

Ok, ..., you usually invite, if you have at least 23, to see, if 23 is the cut off, you would need to rerun the simulation
with 23.

With kind regards
Marlowe

Maybe also 25HCP, to see how much it cost to miss a 25HCP hand.

[P_Marlowe]

 frank0, on 2012-August-08, 08:11, said:

Because 4333/4432 are the shape people usually use HCP to evaluate without additional adjustment. With a 5+ suit or shortness people start to upgrade/downgrade the hand depends on the degree of fit.


Maybe also 25HCP, to see how much it cost to miss a 25HCP hand.

Yes, ... and when you are at it, 22 / 26, to see, if the conclusion
which action to take for 23 / 25, get heavier supported, stronger when
you look at 22 / 26.

With kind regards
Marlowe

[mycroft]

I've always said (not that I'm an Expert, or even an Eeeexpert) that I don't mind being in 24-HCP 3NTs if the alternative is missing 26-HCP 3NTs. I'm not going to go looking for a 24, though (well, maybe 12/12).

[mikeh]

I appreciate that this is the N/B forum, but I think that it is still appropriate to suggest that one cannot answer the OP question without noting that spot cards matter.

AJ109 K109 Q109 Q109 is a much more powerful hand (especially in notrump) than AJ32 K32 Q32 Q32, yet the 4321 point count implies that both hands are worth the same.

This sort of issue is rarely addressed by those who do double dummy simulations, for example. Indeed, my double-dummy generator/analyzer doesn't allow one to use constraints that use spots as adjustable criteria.

In addition, it seems to be generally accepted that the 4321 count is at best a rough approximation of playing value....if all other factors are equal, it seems to undervalue Aces and Kings (perhaps more so for suit contracts than for notrump) and overvalue Queens and Jacks.

This is why experienced bridge players describe hands with language such as 'I held a good 12 count' or 'I had a bad 14 count'....the adjectives reflect the often subconscious valuation of such features (along with a number of other adjustments that good players implement without recourse to arithmetic).

I think a good 12 opposite a good 12 is usually sufficient for game (3N or, with some shape and fit, 4M) at any form of scoring, but a bad 12 opposite a bad 12 should stick to 1N (or the 2 level in a suit) if possible

Learning that the 4321 guide needs to be mentally adjusted for various factors is an important step for aspiring bridge players, imo. Learning HOW to adjust is another topic altogether. Most are taught arithmetical adjustments....so many 'points' for extra length or for shortness, etc. However, we eventually stop using arithmetic and replace it with judgment, which is the result of experience.

[Antrax]

Mikeh, I thought statistical analysis showed that for balanced hands 4-3-2-1 is pretty close to optimum evaluation, no?

[mikeh]

 Antrax, on 2012-August-08, 11:40, said:

Mikeh, I thought statistical analysis showed that for balanced hands 4-3-2-1 is pretty close to optimum evaluation, no?

I haven't paid much attention to statistical analyses for many years. I do know that (years ago) the expert consensus was that 4321's main attraction was ease of use rather than precision. Besides, I don't think many experts really use point count very much: at least not in the sense that beginners are taught to use it.

I recently saw either Fantoni or Nunes open 1♠ (15+, forcing 1 round) on something that looked like K109xxxx xx void AKQx (I am approximating). I very much doubt that opener mentally 'added up' his points, with distributional tweaks, to come up to 15 or 16...I suspect he just looked at the hand and realized that this was a strong playing hand, too good for the limited fantunes 2♠ opener. I may be wrong in my assessment, but I doubt it.

Anyway, my point about spot cards should (if accurate) demonstrates that no matter whether the 4321 is 'better' than, say 6421 or 5321 or 6.324/4.873/2.983/1.783, when generalized over all possible balanced hand, any metric that gives weight only to the highest 4 cards in each suit has to be imprecise.

Any double dummy analysis of the hcp required for 2 balanced hands to bid 3N is going to be deeply flawed if it ignores spot cards.

[kenrexford]

 mikeh, on 2012-August-08, 13:00, said:

I haven't paid much attention to statistical analyses for many years. I do know that (years ago) the expert consensus was that 4321's main attraction was ease of use rather than precision. Besides, I don't think many experts really use point count very much: at least not in the sense that beginners are taught to use it.

I recently saw either Fantoni or Nunes open 1♠ (15+, forcing 1 round) on something that looked like K109xxxx xx void AKQx (I am approximating). I very much doubt that opener mentally 'added up' his points, with distributional tweaks, to come up to 15 or 16...I suspect he just looked at the hand and realized that this was a strong playing hand, too good for the limited fantunes 2♠ opener. I may be wrong in my assessment, but I doubt it.

Anyway, my point about spot cards should (if accurate) demonstrates that no matter whether the 4321 is 'better' than, say 6421 or 5321 or 6.324/4.873/2.983/1.783, when generalized over all possible balanced hand, any metric that gives weight only to the highest 4 cards in each suit has to be imprecise.

Any double dummy analysis of the hcp required for 2 balanced hands to bid 3N is going to be deeply flawed if it ignores spot cards.

Come on, Mike. I just watched a short lesson on TV from a pro golfer who said that he puts a club on the ground before he swings practice shots. The same basic idea applies in bridge. No one ever grabstheir and and does not count out the 4-3-2-1 count. This is routine. You go from there, but you ALWAYS know the HCP count.

[mikeh]

 kenrexford, on 2012-August-08, 15:12, said:

Come on, Mike. I just watched a short lesson on TV from a pro golfer who said that he puts a club on the ground before he swings practice shots. The same basic idea applies in bridge. No one ever grabstheir and and does not count out the 4-3-2-1 count. This is routine. You go from there, but you ALWAYS know the HCP count.

I didn't say what you think I said. Of course everyone counts hcp (at least, I think they do) but I don't know any good player who, for example, looks at a long suit and adds some number of points for each card over a certain length, or who looks at shortness and adds points for that. So I don't think a Fantoni or a Nunes would pick up the example hand and say: I have 12 hcp on and i can add 5 points (or any other number of points) for my length in spades and that makes 17, so I open with a call showing 15+. I think a player like that probably sees 12 hcp and immediately realizes that this hand is far too strong for an opening limited to 14 hcp, without worrying if the 'correct' number is 15, 16, 17 etc.

Anymore than a lesser player such as I would pick up AQJxxx Kxx xxx x and decide that this is a 12 or 13 point hand, or whatever total some distribution point count metric would suggest, and defend my 1♠ bid on that basis....anymore than a good player would (unless a Grannovetter clone) say this was a 10 count and therefore a weak 2 bid.

I was talking about how the 4321 count is an imprecise metric. I also commented about the fact that players are often taught to adjust the count on a hand by numeric values for distribution, but that in my experience no good player uses such a method....any more than good players bid suit slams by adding up points to 33 or so....good players bid suit slams because they determine that they can make 12 or 13 tricks.

[kenrexford]

 mikeh, on 2012-August-08, 18:22, said:

I didn't say what you think I said. Of course everyone counts hcp (at least, I think they do) but I don't know any good player who, for example, looks at a long suit and adds some number of points for each card over a certain length, or who looks at shortness and adds points for that. So I don't think a Fantoni or a Nunes would pick up the example hand and say: I have 12 hcp on and i can add 5 points (or any other number of points) for my length in spades and that makes 17, so I open with a call showing 15+. I think a player like that probably sees 12 hcp and immediately realizes that this hand is far too strong for an opening limited to 14 hcp, without worrying if the 'correct' number is 15, 16, 17 etc.

Anymore than a lesser player such as I would pick up AQJxxx Kxx xxx x and decide that this is a 12 or 13 point hand, or whatever total some distribution point count metric would suggest, and defend my 1♠ bid on that basis....anymore than a good player would (unless a Grannovetter clone) say this was a 10 count and therefore a weak 2 bid.

I was talking about how the 4321 count is an imprecise metric. I also commented about the fact that players are often taught to adjust the count on a hand by numeric values for distribution, but that in my experience no good player uses such a method....any more than good players bid suit slams by adding up points to 33 or so....good players bid suit slams because they determine that they can make 12 or 13 tricks.

I count fingers and toes. If i run out of fingers, i open. If i run out of toes, i open 2C. I bet thats what they do, too.

[mikeh]

 kenrexford, on 2012-August-08, 18:55, said:

I count fingers and toes. If i run out of fingers, i open. If i run out of toes, i open 2C. I bet thats what they do, too.

so you open all 9 counts and open all 19 counts with 2C? Hmmm.

A