BBO Discussion Forums: A more objective analysis - BBO Discussion Forums

Jump to content

  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

A more objective analysis

#21 User is online   DavidKok 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 629
  • Joined: 2020-March-30
  • Gender:Male
  • Location:Netherlands

Posted 2021-July-08, 06:27

View Postsmerriman, on 2021-July-07, 23:10, said:

You can tell South has the club from the lead, but this line only works if you know South has 4 hearts, since you need to save the heart as an entry.. it seems better to me to hope the queen drops in 3 rounds with AK + a ruff. That only takes 11 tricks on this layout, unless someone can think of a better line? If not, then 'should have' is incorrect.
It's probably worse, but I would have gone low on the first club. If North slams the ace we have two club tricks and are home, and if not I intend to play a heart to the ace, spade back, heart to the king and ruff a heart, then concede a club. I think this caters to hearts being 3-3, hearts 4-2 with a 4-card suit in South, and some unlikely but possible shapes with spades splitting 4-1. But I might lose the 11th trick if the hearts split 5-1 or worse with the 5-card suit being held by South and two club ruffs working because of a favourable spade split.

View PostAL78, on 2021-July-08, 06:12, said:

How do you practice a convention that only comes up a few times a year at most?
In all honesty I would attempt to scrap the convention instead, if those are your only options. But there might not be a good solution.
0

#22 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-08, 08:16

View PostAL78, on 2021-July-08, 05:47, said:

This version of Bergen is one she picked up from a book, going through 3 is 8-10 HCP with 3 card support, going through 3 is 7-11 HCP with 4 card support, leaving a jump raise as pre-emnptive. I don't think this is the standard definition of Bergen.


Can you find out the author of this book and the title so that I can make a mental note to never bother to check out any titles by this author?

Whoever wrote this has zero idea about bidding theory. Bergen raises are premised on law of total tricks, you want 9 trumps to force to 3 level, the idea being that it will usually be good advance sac if it goes down 1, and that the opps were going to get in. (More dubious when you have spades, sometimes 2s would have bought it and you regret being down 1 in 3). They also have some ancillary benefits, like opener can make pushier game tries knowing of 4 cd support, and also *not* make pushier game tries after 1M-2M, knowing 3 cd support only. The main downside is loss of the bid(s) for other purposes, and self-pushing when the opps had the wrong hands to get in and were going to let you win it at 2M. Plus some fairly rare fatal lead-directing doubles. Also deprives you of more accurate game tries describing the exact suit you need help in or being able to use say two-way game tries showing a shortness optionally. Plus occasionally if the opps were going to push you to the 3 level their bidding would have let you play more accurately than if you self-pushed.

So self-forcing to 3 level on 3 cd support and 8 to 10 is nuts. These are hands where both the opps are less likely to compete (the smaller your fit is, the less shapely they tend to be, higher likelihood they have no 8 cd fit at all), AND if they do you usually prefer defending their 3 level to declaring at 3 level. Plus, "7-11" for 3d is far too wide a range to accurately gauge whether to bid game or not, and especially with hearts there is no room for an ask. The normal ranges are ~7-9 & 10-12, for 3c/3d respectively, with 4 cd support (some people reverse the steps) and flattish hands; with added shape you'd potentially upgrade hands into stronger categories.

I think it actually far more likely that your partner misread the book and is misremembering things, than that somebody actually wrote a book advocating this. Perhaps she is conflating some steps with "BROMAD", Bergen raises over a major doubled, which applies over 1M-dbl-? and has artificial step showing constructive raise with 3 cd support, but at the TWO level, not 3.

And even if so, if you are using 3c as 10-12 3 cd support which is different from the book, then what are you using 3d as? 10-12 with 4 cd support?
0

#23 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-08, 08:32

View PostAL78, on 2021-July-08, 06:12, said:

The problem is Jacoby comes up so infrequently that we don't get anywhere near enough practice with it to commit the modified version to memory. How do you practice a convention that only comes up a few times a year at most?


You set up a bidding practice table, and write a dealer script so that you get dealt only hands where the convention would apply.


Even though I'm well aware of the problems with std Jacoby, I end up playing it with the vast majority of partners because of this issue. No one wants to learn some non-std Jacoby because there is no single non-std Jacoby and people play with a variety of partners so it's easier just to use std, it rarely comes up and when it does the std version functions adequately on most of them anyway, so gains from the mod are extremely rare.

0

#24 User is online   AL78 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 896
  • Joined: 2019-October-13

Posted 2021-July-08, 08:44

View PostStephen Tu, on 2021-July-08, 08:16, said:

Can you find out the author of this book and the title so that I can make a mental note to never bother to check out any titles by this author?

Whoever wrote this has zero idea about bidding theory. Bergen raises are premised on law of total tricks, you want 9 trumps to force to 3 level, the idea being that it will usually be good advance sac if it goes down 1, and that the opps were going to get in. (More dubious when you have spades, sometimes 2s would have bought it and you regret being down 1 in 3). They also have some ancillary benefits, like opener can make pushier game tries knowing of 4 cd support, and also *not* make pushier game tries after 1M-2M, knowing 3 cd support only. The main downside is loss of the bid(s) for other purposes, and self-pushing when the opps had the wrong hands to get in and were going to let you win it at 2M. Plus some fairly rare fatal lead-directing doubles. Also deprives you of more accurate game tries describing the exact suit you need help in or being able to use say two-way game tries showing a shortness optionally. Plus occasionally if the opps were going to push you to the 3 level their bidding would have let you play more accurately than if you self-pushed.

So self-forcing to 3 level on 3 cd support and 8 to 10 is nuts. These are hands where both the opps are less likely to compete (the smaller your fit is, the less shapely they tend to be, higher likelihood they have no 8 cd fit at all), AND if they do you usually prefer defending their 3 level to declaring at 3 level. Plus, "7-11" for 3d is far too wide a range to accurately gauge whether to bid game or not, and especially with hearts there is no room for an ask. The normal ranges are ~7-9 & 10-12, for 3c/3d respectively, with 4 cd support (some people reverse the steps) and flattish hands; with added shape you'd potentially upgrade hands into stronger categories.

I think it actually far more likely that your partner misread the book and is misremembering things, than that somebody actually wrote a book advocating this. Perhaps she is conflating some steps with "BROMAD", Bergen raises over a major doubled, which applies over 1M-dbl-? and has artificial step showing constructive raise with 3 cd support, but at the TWO level, not 3.

And even if so, if you are using 3c as 10-12 3 cd support which is different from the book, then what are you using 3d as? 10-12 with 4 cd support?


To quote from previous e-mail correspondance when I asked her this question:

"I have taken the Bergen responses out of Sally Brock's Book 'Easy Guide to 5 card majors'. This is what she says:

* A 3C response shows 10-12 pts with THREE card support
* A 3D response shows 8-11 pts with FOUR card support
* A jump raise shows 0-7 with at least FOUR card support, one useful High card and a doubleton.
I judge this bid on loser count and vulnerability.

She continues: This makes it easy to distinguish between 3 and 4 card support and helps to judge slam
expectations and whether to bid on if opponents intervene."

Sorry, it was 8-11, not 7-11 for the 3 response.
0

#25 User is online   DavidKok 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 629
  • Joined: 2020-March-30
  • Gender:Male
  • Location:Netherlands

Posted 2021-July-08, 09:51

View PostAL78, on 2021-July-08, 08:44, said:

* A 3C response shows 10-12 pts with THREE card support
* A 3D response shows 8-11 pts with FOUR card support
* A jump raise shows 0-7 with at least FOUR card support, one useful High card and a doubleton.
I judge this bid on loser count and vulnerability.
This makes a lot more sense. The standard limit raise (~8-11 with 3 card support) is a soft spot in modern standard (2/1) bidding, where typically you first go through 1NT, semiforcing or forcing, and then jump to 3 on the second round. This is somewhat of a 'worst of all worlds' - sometimes partner passes 1NT, if the opponents interfere you haven't shown your support, and you end up at the 3-level even if it is bad. Compared to that bidding 3 directly is not worse bridge.
However, it is completely non-standard. Stephen pointed out the downside of bidding to the 3-level one down while you hold the balance of strength, and with 3-card support this is quite an issue. Personally I do not like the split of the ranges at all - there are so many hands with a good 10 or 11 and 4-card support where I'd have to call 3 and watch partner sign off expecting 8-9, or force the issue with a Jacoby NT. That's why the original Bergen raises had such a fine-grained split:
  • 0-5 with 3-card support: pass, then support if you get to bid later.
  • 6-9 with 3-card support: 2M
  • 10-11 with 3-card support: 1NT, then 3M
  • 12+ with 3-card support: 2m (usually 2), then 2/3M depending on style
  • 0-5 with 4-card support: 3M
  • 6-9 ('bad 9') with 4-card support: 3 Bergen Raise
  • 9-11 ('good 9') with 4-card support: 3 Bergen Raise
  • 12+ with 4-card support: Jacoby 2NT

This way at least if you bid your somewhat OK 11-count with 4-card support with 3 at least partner would put you within a point of your actual playing strength. With the wider ranges you describe you may well run into issues here.
0

#26 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-08, 10:12

View PostAL78, on 2021-July-08, 08:44, said:

To quote from previous e-mail correspondance when I asked her this question:

"I have taken the Bergen responses out of Sally Brock's Book 'Easy Guide to 5 card majors'. This is what she says:

* A 3C response shows 10-12 pts with THREE card support
* A 3D response shows 8-11 pts with FOUR card support
* A jump raise shows 0-7 with at least FOUR card support, one useful High card and a doubleton.
I judge this bid on loser count and vulnerability.

Sorry, it was 8-11, not 7-11 for the 3 response.

OK, this makes a ton more sense. You originally wrote 3c = 8-10 with 3 cd support which doesn't make any bridge sense.

So this structure is only partially Bergen raises. 3D/3M are similar to Bergen, but 3c is not. I guess this is "Brock raises". She's swapped out the constructive Bergen raise (~7-9 with 4 cd support, not enough shape to upgrade to an invitational raise) for the 3 cd limit raise hands. This is playable, it's only forcing to the 3 level on 3 trumps on game invite hands which are likely to take 9 tricks opposite a decline anyway, and that more common methods also force to the 3 level (after 1nt forcing). Perhaps she is advocating this in a book catering to Acol 5-cd M converts still using a weaker NF 1M-1nt response (9/10 max), or perhaps still likes a bid to show this hand playing a semi-forcing 1nt (12 max but still NF opposite bal 12-13).

Hopefully on the 3d = 8-11, she is talking solely about hcp range and only advocating the lower part of the range with added points for shape. Like 2 aces and a stiff is reasonable for a 4 cd limit raise, but 2 aces in a flat hand is too aggressive to invite with, and the latter should just bid 1M-2M. The range might be 8-11 in terms of raw HCP, but it should be ~10-12 in "support points".
0

#27 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-08, 10:18

View PostDavidKok, on 2021-July-08, 09:51, said:

  • 0-5 with 3-card support: pass, then support if you get to bid later.
  • 6-9 with 3-card support: 2M


Note that a fair number of 5CM players will use a forcing NT with ~4-6 and 3cd support, intending to rebid 2M, and have 2M be a semi-constructive raise of 7-bad 10 or so. This is the structure in the "Washington Standard" book by Steve Robinson, for example. The idea is to discourage opps from competing without encouraging partner too much with the weak raise hands.
0

#28 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-08, 10:59

View PostAL78, on 2021-July-08, 06:23, said:

It is sometimes not clear to me whether it is best at MPs to take a safe line which guarentees the contract, or take a risk which will result in an extra trick or two if it works, but goes down if it doesn't. An example would be declaring 3NT, you have an easy nine tricks, do you take a finesse for an overtrick if you go down when the finesse loses?


This is often field dependent. In a really poor field, the more you should play safely to make your contract, because of people who have bidding accidents and miss game, or who butcher the play completely and go down. The better the field the more risks you should be taking.

The things to ask yourself are:
1. Is a significant chunk of the field going to miss bidding the game, or is everyone in it? If people are missing game, tend to play safe.
2. Do I think I have received an unusually favorable lead not given at other tables, or has the defense already blundered a trick later in the play? If you are already one trick up on the field, going for a second trick usually gains nothing or not much, while giving it back is disastrous, so play safe.
3. On the flip side, have the opps appeared to make an unusually good lead for them not found at other tables, and you think you are a trick behind the field, and the hook (which field isn't taking, on a different lead) would bring you back to even?
4. Am I in a higher scoring game than the field is in? Am I in 3nt when the field will be in 4M, or vice versa? Is the field going to take this hook or not, and how many tricks are they going to end up with? Can I do something so that my score will surpass the likely field result if my contract is unusual, or is the die already cast?
5. What are my actual odds of success? If it's a straight 50/50 finesse, I am basically never going to take it if it risks the contract, because of stragglers who somehow aren't in game. But reasonably often there is a restricted choice situation (e.g 1/2 finesses with xx opposite AJT), or some opponent has been counted out to have 2x number of cards in the suit in question (if LHO has 4 cds in suit, RHO is 2, in a vacuum of other relevant info LHO is 2 to 1 to have the critical card). Or maybe an opponent has preempted and the hook against his partner in some suit is a lot more likely to succeed based on vacant spaces. In these situations, if everyone's in the same game, you should be risking your contract for these type of probable overtrick.

There's a good chapter on MP play in Bill Root's "How to Play a Bridge Hand", a great book overall, worth getting.
0

#29 User is online   mw64ahw 

  • PipPipPipPip
  • Group: Full Members
  • Posts: 251
  • Joined: 2021-February-13
  • Gender:Not Telling
  • Interests:Bidding & play optimisation via simulation.

Posted 2021-July-08, 11:14

View PostStephen Tu, on 2021-July-08, 04:08, said:

More like 100% of players who are any good, not merely a majority. It makes it impossible to play in any other suit so you want to have 4+; reasonably often if there is another 4-4 or 9/10 cd fit in addition to just a 5-3 cd fit in the opened major, the other fit can take more tricks and slam is only possible or much better percentage in the other strain. Also 4+ is important for hand evaluation purposes.

Sounds a bit harsh given the old version of Jacoby 2NT suggested 3 with 2 of the top 3 honours was OK. I thought I'd read on one of the forums that 3+ was used regularly in some areas, but I guess none of these were good players!

View PostAL78, on 2021-July-08, 06:12, said:

I understand, but this partner is not open to taking on new conventions or modifying existing ones*. With a different partner we did agree to try a modified Jacoby that may have been linked from a post in this forum, but we ultimately went back to the standard version. The problem is Jacoby comes up so infrequently that we don't get anywhere near enough practice with it to commit the modified version to memory. How do you practice a convention that only comes up a few times a year at most

I found that the standard Jacoby 2NT was underused so I'm now experimenting with a modified version. I found that whether I had 3 or 4+ GI or GF I tended to end up in the same contract, but just told everyone whether I had 3 or 4+. I'm now working with a 3+ limit+ version which has had good results so far. It's simple to remember and so far I've not seen adverse results. Simply, I respond using a combined hcp/modified losing trick count so opposite 1-1NT I have:
3-8/7.5 10+
----3 asks which with 7.5 or fewer losers
3-7 12+
3-6.5 14+
3-6 16+
Higher- 5.5 or fewer 17+ Italian cue bids with 3NT showing 2+ losers (unlikely, but allows partner to stop in 3NT)

The key issue I find is that my declarer play can struggle opposite 3 rather than 4+ even when the contract is makeable. The modified looking trick count partly factors in distribution, but keeps the opponents in the dark.
0

#30 User is online   DavidKok 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 629
  • Joined: 2020-March-30
  • Gender:Male
  • Location:Netherlands

Posted 2021-July-08, 12:18

View Postmw64ahw, on 2021-July-08, 11:14, said:

I'm now working with a 3+ limit+ version which has had good results so far. It's simple to remember and so far I've not seen adverse results.
I personally play 'De Maas', where 2NT shows a limit raise or minimum GF with 3(+) support. I've had beautiful results with it. The auction 1M-2NT; 4M is very common and a big winner at IMPs.
0

#31 User is online   mw64ahw 

  • PipPipPipPip
  • Group: Full Members
  • Posts: 251
  • Joined: 2021-February-13
  • Gender:Not Telling
  • Interests:Bidding & play optimisation via simulation.

Posted 2021-July-09, 00:07

View PostDavidKok, on 2021-July-08, 12:18, said:

I personally play 'De Maas', where 2NT shows a limit raise or minimum GF with 3(+) support. I've had beautiful results with it. The auction 1M-2NT; 4M is very common and a big winner at IMPs.

Not come across 'De Mass' and can' t find any references, but I assume its straightforward. Any further context?



0

#32 User is online   DavidKok 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 629
  • Joined: 2020-March-30
  • Gender:Male
  • Location:Netherlands

Posted 2021-July-09, 03:02

De Maas, in Dutch. You can pull the site through Google Translate to get this, but it messes up the suit symbols. I personally translate all those point ranges into 'between 8.5 and 7 losers', though with 8 (or fewer) points and 8.5 losers I prefer a simple raise and with a good 14 (or more) points and 7 losers I prefer to bid 2, showing mild SI with a delayed raise.

The way I play it is slightly different compared to this linked version.
For me 1M-2NT; 3oM shows 5-4, GF, choice of games, no SI. Over 1M-2NT; 3 I play a different structure (3 3-card support, any other answer shows 4(+) and shortness, 3M 4(+) support and no shortness). The goal is to bid 1M-2NT; 4M most of the time, 1M-2NT; 3 asks for min/max (I play that responder may also suggest 3NT with a maximum with 3-card support over 3, but typically responder will choose 3M/4M), bidding 3 or jumping to 4 (the latter shows a real suit) is SI and the rest of the bids are choice of games.

Edit: a more legible (and slightly different) version in English.
0

#33 User is online   AL78 

  • PipPipPipPipPip
  • Group: Full Members
  • Posts: 896
  • Joined: 2019-October-13

Posted 2021-July-09, 07:41

View PostStephen Tu, on 2021-July-08, 10:59, said:

This is often field dependent. In a really poor field, the more you should play safely to make your contract, because of people who have bidding accidents and miss game, or who butcher the play completely and go down. The better the field the more risks you should be taking.

The things to ask yourself are:
1. Is a significant chunk of the field going to miss bidding the game, or is everyone in it? If people are missing game, tend to play safe.
2. Do I think I have received an unusually favorable lead not given at other tables, or has the defense already blundered a trick later in the play? If you are already one trick up on the field, going for a second trick usually gains nothing or not much, while giving it back is disastrous, so play safe.
3. On the flip side, have the opps appeared to make an unusually good lead for them not found at other tables, and you think you are a trick behind the field, and the hook (which field isn't taking, on a different lead) would bring you back to even?
4. Am I in a higher scoring game than the field is in? Am I in 3nt when the field will be in 4M, or vice versa? Is the field going to take this hook or not, and how many tricks are they going to end up with? Can I do something so that my score will surpass the likely field result if my contract is unusual, or is the die already cast?
5. What are my actual odds of success? If it's a straight 50/50 finesse, I am basically never going to take it if it risks the contract, because of stragglers who somehow aren't in game. But reasonably often there is a restricted choice situation (e.g 1/2 finesses with xx opposite AJT), or some opponent has been counted out to have 2x number of cards in the suit in question (if LHO has 4 cds in suit, RHO is 2, in a vacuum of other relevant info LHO is 2 to 1 to have the critical card). Or maybe an opponent has preempted and the hook against his partner in some suit is a lot more likely to succeed based on vacant spaces. In these situations, if everyone's in the same game, you should be risking your contract for these type of probable overtrick.

There's a good chapter on MP play in Bill Root's "How to Play a Bridge Hand", a great book overall, worth getting.


A good example is Bridge Master Beginner level A-29. Trivial at IMPS, win the lead, duck a club, cold if clubs are not 5-0. What do you do at MPS? Go for the 3-2 break which is about a 2 in 3 chance and go off on a 4-1 break, or take the safety play, assuming you have no information about the opponent's hands and it is likely almost everyone will be in 3NT? My guess would be go for maximum tricks which will be a winner in the long term given the odds of a 3-2 break.
0

#34 User is offline   Stephen Tu 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,912
  • Joined: 2003-May-14

Posted 2021-July-09, 08:36

View PostAL78, on 2021-July-09, 07:41, said:

A good example is Bridge Master Beginner level A-29. Trivial at IMPS, win the lead, duck a club, cold if clubs are not 5-0. What do you do at MPS? Go for the 3-2 break which is about a 2 in 3 chance and go off on a 4-1 break, or take the safety play, assuming you have no information about the opponent's hands and it is likely almost everyone will be in 3NT? My guess would be go for maximum tricks which will be a winner in the long term given the odds of a 3-2 break.

It's easy this one since you know everyone should be in 3nt. You go for max probable tricks, and win MP against the safety player 68% but lose 28%. Note if you had a suit like AKQxx opposite xx, 7 cd fit rather than 8, if you had to say get 4 tricks from this suit (and no other entry), you *should* duck a round at MP, even though going for max *possible* tricks is not ducking and hoping for 3-3. That's because 3-3 split is only 35.5%, while 4-2 is 48.5%. Ducking once would win more tricks (4 vs 3) more often, than it would lose (4 vs 5). You are going for max *probable*, not max *possible*.

The trickier ones are when you aren't sure everyone is in the best contract, but you really have no idea how big a portion of the field is likely to miss so it's hard to calculate the relative odds.

For me, generally at MP I am not going to be looking for "onerous safety play" lines that involve giving up a trick unnecessarily more often than just going for max probable tricks. I expect the field to get to the games I get to. If I ever do such a safety play it's usually in some slam where I don't think the field will reach, or the opponents have doubled me and I have a safety play to make, or generally situations where I think I'm in an unusual contract in some way, and that the overtrick is extremely unlikely to matter much, but making or not almost certainly will.

1

#35 User is online   mw64ahw 

  • PipPipPipPip
  • Group: Full Members
  • Posts: 251
  • Joined: 2021-February-13
  • Gender:Not Telling
  • Interests:Bidding & play optimisation via simulation.

Posted 2021-July-09, 10:16

View PostDavidKok, on 2021-July-09, 03:02, said:

De Maas, in Dutch. You can pull the site through Google Translate to get this, but it messes up the suit symbols. I personally translate all those point ranges into 'between 8.5 and 7 losers', though with 8 (or fewer) points and 8.5 losers I prefer a simple raise and with a good 14 (or more) points and 7 losers I prefer to bid 2, showing mild SI with a delayed raise.

The way I play it is slightly different compared to this linked version.
For me 1M-2NT; 3oM shows 5-4, GF, choice of games, no SI. Over 1M-2NT; 3 I play a different structure (3 3-card support, any other answer shows 4(+) and shortness, 3M 4(+) support and no shortness). The goal is to bid 1M-2NT; 4M most of the time, 1M-2NT; 3 asks for min/max (I play that responder may also suggest 3NT with a maximum with 3-card support over 3, but typically responder will choose 3M/4M), bidding 3 or jumping to 4 (the latter shows a real suit) is SI and the rest of the bids are choice of games.

Edit: a more legible (and slightly different) version in English.

Thx
0

Share this topic:


  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users