[RSClyde]

PhilKing, on 2013-November-27, 05:59, said:

The simulations are not bridge:

1. Real opponents do not unerringly lead your doubleton - the lead is less likely to be critical when we are, say, 4333. In actual play, the stats show that "declarer advantage" occurs mainly on the opening lead.

2. Real declarers don't get 100% of 2-way finesses right or pick the correct suit to establish when they have a 4333 opposite a balanced dummy. In actual play, decent defenders drop fewer tricks after the lead than decent declarers when compared to simulated perfection, and this is particularly true on hands that require good guessing.

3. Real opponents occasionally throw the wrong thing when we run our five-card suit - these are the kinds of hands where declarer advantage still applies after the lead.

In response to the second question, without wishing to state the bleeding obvious, the fifth card is a potential extra trick.

These things may cause the simulator to be a little off center, but it's not clear that this would be a significant error. The things you mention can be true when you have a 5 card suit as well just to a slightly lesser extent.

I'm open to the possibility that upgrading is right, I'd just like a clear reason to believe this beyond something like "all the good player's know..." or "It's always worked for me" or "It's obvious that this hand is worth more than...". I get that the logistics of testing such a thing would be difficult, so like I say I opened minded on this. I do however find it suspicious that the one test I can run doesn't bear out the that upgrading is right.

[rhm]

PhilKing, on 2013-November-27, 05:59, said:

The simulations are not bridge:

1. Real opponents do not unerringly lead your doubleton - the lead is less likely to be critical when we are, say, 4333. In actual play, the stats show that "declarer advantage" occurs mainly on the opening lead.

2. Real declarers don't get 100% of 2-way finesses right or pick the correct suit to establish when they have a 4333 opposite a balanced dummy. In actual play, decent defenders drop fewer tricks after the lead than decent declarers when compared to simulated perfection, and this is particularly true on hands that require good guessing.

3. Real opponents occasionally throw the wrong thing when we run our five-card suit - these are the kinds of hands where declarer advantage still applies after the lead.

In response to the second question, without wishing to state the bleeding obvious, the fifth card is a potential extra trick.

RSClyde, on 2013-December-04, 16:17, said:

These things may cause the simulator to be a little off center, but it's not clear that this would be a significant error. The things you mention can be true when you have a 5 card suit as well just to a slightly lesser extent.

I'm open to the possibility that upgrading is right, I'd just like a clear reason to believe this beyond something like "all the good player's know..." or "It's always worked for me" or "It's obvious that this hand is worth more than...". I get that the logistics of testing such a thing would be difficult, so like I say I opened minded on this. I do however find it suspicious that the one test I can run doesn't bear out the that upgrading is right.

Philkings objections highlights the differences between simulation and bridge.
However, these objections do not explain why simulation should be further off base with a five card suit present in declarer's hand.
All simulations I know attach little value to the fifth card.
However, you could interpret Richard Pavliceks analysis at http://www.rpbridge.net/9x42.htm from real play as an indication that long suits are a positive sign and worth close to a full HCP..

Rainer Herrmann

[PhilKing]

rhm, on 2013-December-05, 05:08, said:

Philkings objections highlights the differences between simulation and bridge.
However, these objections do not explain why simulation should be further off base with a five card suit present in declarer's hand.
All simulations I know attach little value to the fifth card.
However, you could interpret Richard Pavliceks analysis at http://www.rpbridge.net/9x42.htm from real play as an indication that long suits are a positive sign and worth close to a full HCP..

Rainer Herrmann

I think the data overstates the case. For instance, the statisitic for bidding 3NT with a combined 23 and 24 count with one five-card suit are not representative - they are just cases where one team decide to bid up with those values, usually for a good reason. If you look through the hands, generally the hand without the five-card suit was bidding up on the basis of good intermediates. Board 1 in the 24 point group and hand 13 in the 23 point data are good examples.

As to the difference, I suggest that the opening lead is the main difference, since it is the first blow in what is essentially a sprint race. When we are 4333, the lead is less critical.

[NickRW]

Zelandakh, on 2013-November-27, 08:42, said:

A long time ago, some French guy did some analysis and came to the conclusion that the 5 card suit is worth around 0.4hcp.

That's about right (for NT purposes, according to the SIMs I've done).

However it shouldn't be forgotten that 4432 hands compared to 4333 hands are also worth about 0.2hcp (also for NT purposes). Further, 4333 compared to both 4432 and 5332, for play suit purposes, is worth at least a full point less (and you don't always end up in NT just because you open NT).

So I think people have it the wrong way round: Instead of mentally upgrading the 5332 hands, downgrade the 4333 ones.

Nick

[RSClyde]

NickRW, on 2013-December-06, 07:00, said:

That's about right (for NT purposes, according to the SIMs I've done).

However it shouldn't be forgotten that 4432 hands compared to 4333 hands are also worth about 0.2hcp (also for NT purposes). Further, 4333 compared to both 4432 and 5332, for play suit purposes, is worth at least a full point less (and you don't always end up in NT just because you open NT).

So I think people have it the wrong way round: Instead of mentally upgrading the 5332 hands, downgrade the 4333 ones.

Nick

I think there is a fallacy that occurs here however. Which hand is better if partner has a long suit to play in? 4333 or 4432? It's not automatically the latter, because that suit is frequently your short one. Yes the 4432 shape is better opposite the 5(+) card spade suit (than the 4333) if spades isn't the short suit, otherwise the 4333 is better.

[aguahombre]

blackshoe, on 2013-November-27, 12:25, said:

The problem, if there is one, is of course that it not only gives more information to opponents, it also gives more information to partner.

If we agree it is a basic condition of contest that pairs are expected to know what their 1-bids mean, I can't imagine how an alert or announcement by partner will be a problem. Responder is telling the table what Opener is showing; Opener knows what he is showing.

[NickRW]

RSClyde, on 2013-December-06, 08:19, said:

I think there is a fallacy that occurs...

Yes it does vary according to what exact shape is opposite what other exact shape. The numbers I was quoting were averages over large sample sets. At the point you open the bidding (which is what the OP was talking about) you have no idea what partner has - only what you have - so can only go on averages - or your impression of them anyway.

N.

[benlessard]

IMO you should rely on the majors events stats by R.Pavlicek.

http://www.rpbridge.net/rpme.htm

---------------------------------------
Holding 25 pts and no 5 card suits, 140 imps for 3NT-- 119 imps for partscores.

25 pts with 5 card suit = 418 imps for 3nt, 121 imps for partscores.

---------------------------------
24 pts without 5 card suits, = 323 imps for 3NT, 295 for partscores = 52.27%

24 pts with 5 card suit = 712 imps for 3nt, 433 for partscores = 62.18%


IMO its clear that THE FIRST 5th card suit is worth close to 1 pts when your aiming for 3nt since 3NT with 24 pts and a 5 card has a higher rate of succes than with 25 pts without any 5 card suits.

This doesnt take into account when you have two 5 cards suits however, the effect probably doenst add since you will rarely have time to established both suits.

[PhilKing]

benlessard, on 2013-December-06, 15:41, said:

IMO you should rely on the majors events stats by R.Pavlicek.

http://www.rpbridge.net/rpme.htm

---------------------------------------
Holding 25 pts and no 5 card suits, 140 imps for 3NT-- 119 imps for partscores.

25 pts with 5 card suit = 418 imps for 3nt, 121 imps for partscores.

---------------------------------
24 pts without 5 card suits, = 323 imps for 3NT, 295 for partscores = 52.27%

24 pts with 5 card suit = 712 imps for 3nt, 433 for partscores = 62.18%


IMO its clear that THE FIRST 5th card suit is worth close to 1 pts when your aiming for 3nt since 3NT with 24 pts and a 5 card has a higher rate of succes than with 25 pts without any 5 card suits.

This doesnt take into account when you have two 5 cards suits however, the effect probably doenst add since you will rarely have time to established both suits.

I think you are making a statistical error by treating hands where (for example) players chose to bid game with 24 points and a five card suit as representative of hands with 24 points and a five card suit as a whole.

Similarly, for the 25 points no long suit data, downgrading often came in to play - look at Woodridge's remarkable decision on board 6 to treat a 3433 18 count with two tens as a 15-17 NT. If the data included hands where both sides bid 3nt with no long suit, the results might be very different.

[benlessard]

Unless there is a study on hands where opener VS where its responder that has a 5 cards nothing we can do about it.

But these are hands where one table bid 3NT and the other didnt.

Hands where both tables bid the same contracts are not calculated. So it kind of neatly show that the FIRST 5th card suit is worth a lot because on close hand (assuming that the score is moslty due to bidding and not to declarer play/def) being agressive with a 5th card seems to reward.

Of course its not a lot of hands but its all top-level.

[jogs]

blackshoe, on 2013-November-25, 20:00, said:

Hand evaluation is more art than science.

Point count was designed to measure the effects of honors. It doesn't do well estimating the value of length or shortness. What's the fifth card in a suit worth? 65432. Probably nothing. AKQJT. Five full tricks. AKQJ2. Probably still five tricks.

[Lovera]

sfi, on 2013-November-27, 08:08, said:

On the face of it this statement seems wrong. Otherwise a combined 18 count would expect to give you play for game most of the time. I was going with the simple 40/13 = 3 points/trick calculation, so where does this figure hold?

This simply consideration done has started (toghether other ones) my valutation of the hand based on a (light) variation of that made by S. Stayman (maximum points in 4-3-3-3 is 37 for 13 tricks in NT ..)

[johnu]

Lovera, on 2014-September-13, 03:41, said:

This simply consideration done has started (toghether other ones) my valutation of the hand based on a (light) variation of that made by S. Stayman (maximum points in 4-3-3-3 is 37 for 13 tricks in NT ..)

???

[PrecisionL]

blackshoe said:

1385576707[/url]' post='766785']
And IMO at least disclosure is complicated by the ACBL's "state the range" with examples that only say "X to Y" with no caveats. So is "good 14 to flat 17" proper under this regulation, not withstanding that it gives more information than "14 to 17"? The problem, if there is one, is of course that it not only gives more information to opponents, it also gives more information to partner.

14+ - 17 I have seen to show a possible upgrade.