[Free]

chicken, on Apr 6 2005, 07:36 PM, said:

Free, on Apr 5 2005, 11:09 AM, said:

chicken, on Apr 5 2005, 04:18 PM, said:

in my opinion this is definately a brownsticker convention, good for this site and
A-Level tourneys, but not allowed in 90% of the tourneys all over the world. as far as i know (and i play X-fer preempts with str 2-suiter option as well) wbf calls this a hum convention as long as no anchor suit (dont know the precise technical term in english) is announced.

Total rubbish! The anchor suit is only needed for a weak version, so the preempt suit is known (transferred). Apparently you need to read the rules, because you don't even know what kind of system YOU play.

although i dont like the tone of ur reply u may be right. however when we first played a rather similar sys we were penalized by the german bridge federation. the TD told us that there must be an anchor suit for both variants. so since then we played 3club = weak diamonds or STR 2-suiter with diamonds and so on. this might be a specific german ruling, or the TD was silly. obviously the rulings seem to differ a lot.

I didn't mean to offend you in any way, and I still don't want to...

I doubt that the description of "brown sticker convention" and "hum" are different in some countries (what's the WBF for?). However, some countries just add specific rules, like "rule of 18", or specify what kind of 'psychs' are allowed, or forbid 1-level openings with less than 8HCP, or many other stuff.

When one needs transfer preempts with an anchor suit in both possibilities (whenever you're weak or strong) are easily modified, just always show the transferred suit...

[inquiry]

Ok, let's forget the for a minute all the different opener's rebids. Instead lets deal with one specific auction and how the bidding will continue to see how the system works in a simple format.

As I said earlier, i have coined the term Misiry for Misho - Inquiry. True Misiry is the preemptive hands. Then, we have Misiry loves company hands which re the big hands. So, let's examine this specific auction, opponents silent

3C-3D-3H. Here, the 3H rebids shows Exactly four losers and at least 10 cards in the majorss. Responder counts his covers and decides if he should signoff in partscore, game, slam or grand slam. Or, if he has potential cover cards, he can cooperate with opener to see if the potential covers are working.

The first "maybe " cover cards are CONTROLs in the non-anchor suits. After you determine their usefulness, you can ask about wayward sisiters (two anchor suit queens).

If responder is uncertain if one or more of his potential cover cards is "good" or not, he can explore by misiry cue-bids and then search for wayward sisters. The misiry cue-bid is where responders first option is to bid in a non-anchor suit where he lacks any potential cover card (SEE THE VERY RARE EXCEPTION LATER0. If two quick tricks are off in the misiry bid suit, opener bids the cheapest anchor suit as pass/correct. Let;s see how this works on our example auction

3C-3D-3H-?

• Pass = I have hearts and not even one cover card
  
• 3♠ = I prefer spades, and I lack even 1 cover card
  
• 3NT = to play, I know you have a two suiter, but, I think this is best
  
• 4[cll] = I lack club ACE, King, singleton, or void
  
• 4♦ = I have A, K, singleton or void in clubs, but no such holding in diamonds
  
• 4♥ = I have one or two cover cards and hearts. If two, I don't hold possible cover for a wayward queen
  
• [4sp] = same as 4♥, but I prefer spade to hearts
  
• 4NT = I have diamond king, and a control in clubs, can be Ace or King.
  
• 5♣ I have ace of diamonds and king of clubs
  
• 5♦ = I have AK in one non-anchor and A in the other. It doesn't matter which.

In general, after the misiry cue-bid opener may want to ask about the control in the other off suit, especially if a suit bypassed. We use the cheapest non-anchor suit bid (which can include pass if the opponents intefere) to ask about the controls in the other suit. On an auction where a suit is bypassed, like a 4D bid here, a 4S bid would ask about the nature of the club control. Opener ONLY ASK in theory if the club KING is wasted. Responder typically signs off with the club king, and continues with the ACE. This is important, if OPENER doesn't ask about the other anchor suit, he must either know which is held, or either control is a cover card (xx in opener hand). This later case is important as it means with AK in a bypassed suit, both are working!!

If partner DOESN'T skip a suit, the meaning of simpliest non-signoff is the same, only the ACE in the other suit is useful. Any other bid (besides cheapest signoff and cheapest non-signoff) shows any cover in the other suit is useful. So after the initial misiry cue-bid and opener's response to it, responder is well positioned to determine how useful his noin-anchor suit cover cards are. Let's examine some partial hands to see how this works (I will show only the non-achor suits - clubs and diamonds)..

Now, what if opener doesn't issue the one step warning that "other suit" king is wasted? The next bids are conventional.

D-Kxx   
C-xxx
3C-3D-3H-4C-4H
1.   Here, 4H is a warning, we are off two clubs. You stop, pass 4♥ or move to 4♠
3C-3D-3H-4C-4S
2.   Here 4S says, no two club losers, but since it is the cheapest non-signoff, it issues a warning that only diamond ACE would be a cover, so you don't count the diamond King as a cover card. Easy?
3C-3D-3H-4C-4N
3.   4N says that all is ok and gives responder a chance to show potential covers for wayward queens. And if room exist below five of the cheapest anchor suit, opener can show grand slam potential (no quick loser in either non-achor suit) and where coverage for a wayward queen would be an extra cover. Here you would count diamond King as full cover, and diamond AK as two full covers.

So what is a wayward queen? xx in one anchor suit with adequate support for the other, or xxxxx in one of the anchor suits. Over 4NT responder can bid 5C/5D to show coverage for wayward queen in lower (5C) and higher suit (5D). If opener was going to try for grand slam, he could have bid 5C and 5D as well, with the same meaning. Opener can jump to 5H to show either substitute queen is working.

This appoach allows you to explore if specific covers exist.. consider this hand by opener....

AKJxx
AKT9x
x
Kx

Let's see some auctions..all start 3C-3D-3H (major 2 suiter, 4 losers)

3C-3D-3H-4C-4D-4H <<--- 4C denied club control, 4D asked about diamonds (cheapest non-signout other non-anchor ask), 4H was signoff, no diamond ACE.

3C-3D-3H-4C-4D-4N <--- working diamond ACE, no substitute queen, now opener can ask for four card support if missing one.

3C-3D-3H-4D-5H-6S<<--- 4D promised ACE or king of clubs, denied ace of diamond or king of diamonds, Five diamonds showed missing both queens. 6S shows a substitute queen of spades (length) or queen of diqamonds (shortness).

This gives an idea how this works. Looks like Misiry is good name for it.. :-)

Ben

[fred]

inquiry, on Apr 6 2005, 01:19 PM, said:

Well, I see more and more transfer preempts. For instance, I know that Russ Ekeblad and Ron Rubin use 3C as weak tranfer to 3D, 3S as weak transfer to 3H, and 2NT as weak transfer to "a minor" I think, since they seem to have either. I have not seen a 3H opening bid by them, presumably it is not spades, as they alert their 2S bid as weak two or three spade bid. Their 3S bid seems to be a very good suit (AKQTxxx is the example I saw).

Russ and Ronnie use 3H as weak with both majors. They are my regular teammates so I am in a good position to tell you that they have had a couple of impressive results with this convention (not that this proves anything).

Russ loves talking system and I am sure he would be happy to answer any questiosn you might have about their (very interesting) methods.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[whereagles]

3H as both majors is actually a though nut to crack on 2nd seat because you can never be sure it will be passed out or not. Tremendous pressure because you're given a chance to be passive or active and both can turn out terribly well or terribly wrong.

Still, technically it shouldn't be too hard to defend because you definitely don't want to play in one of the majors (you might want to do that if it were 2H showing a 44 major or better), so you have a couple of bids available:

3H ..?

X = good balanced hand, penalty oriented
3NT = natural, usually based on a long minor
3S = 54 minors either way or 55, good hand
4m = natural, very good hand
4NT = 55 minors, very good hand
4M = 55 minors, very good hand, void
Pass + double = penalties

The problem would be to judge at table which action is best.

[pclayton]

Checking in late on this discussion as usual......

This whole argument about 'frequency' relating to Ben's multi-preempts is hogwash.

Just because a hand with a preempt hand comes up (say) 5x times more often than the strong 2-suited variety is irrelevant. Usually the responder to the transfer preempt will just accept the transfer and play it there. Positionally, its better for the long suit to be on the table, so transfer preempts will frequently gain in this scenario.

Even when responder does hold a hand that wants to bounce, frequently, the preempter's LHO will have taken a call anyway. Responder will usually be able to guess what hand pattern partner has judging by what action the direct seat takes. I'd say that responder is only at a slight disadvantage.

Ergo, the only real positive gain for standard natural preempts is when the call gets past preempter's LHO and preempter's RHO has a monster. Having to guess at the 4 or 5 level is a lot tougher than taking a shot at the 3 level while the multi-transferer works out his hand pattern.

I will also admit that its problematic for the responder to psyche over one of these multi preempts.

I don't think responder is at any less of a disdvantage then when the partnership uses a traditional multi 2♦ (either weak 2 in major or a 2N opening). Responder is very limited on the amount of 'bouncing' he is able to do, since opener can have THREE hand types. As I understand Ben's system, opener is limited to basically two types: a normal preempt or a strong 2 suiter with a known anchor suit.

__________________________________

I think I'm going to explore these further. I'm flabbergasted that the ACBL allows these as mid-chart, I would think they wouldn't be allowed at all.

For a pair that plays a strong club, I think these fill a very important role. Strong 2 suiters are very vulnerable to preemption after Opener has to use a strong ♣. Even if the opponents don't get busy in the auction, frequently, opener can't sort out the suits until the 3 or 4 level anyway, so the 'constructive' sequences gains very little.

Multi-preempts would also gel very nicely with my 2♣ opening. Two suiters frequently get lost after 2♣ - 2♠ / 2N / 3♣ (3,4 and 5 controls) response, so these would aid greatly.

The need for a natural jump shift by opener seems to be all but eliminated. These could be converted over to mini-splinters, or the weaker distributional 5-5 or 6-5 hands.

I would be careful with 5-5-0-3 patterns. I can see how the fragment can get easily lost after a multi-preempt.

I will follow this thread with interest..........

[hrothgar]

pclayton, on Apr 7 2005, 06:41 PM, said:

Checking in late on this discussion as usual......

This whole argument about 'frequency' relating to Ben's multi-preempts is hogwash.

Just because a hand with a preempt hand comes up (say) 5x times more often than the strong 2-suited variety is irrelevant.

If you methods cause you to win 16 IMPs 5% of the time, but lose 2 IMPs, 95% of the time, then by definition, the expected value for the method is

(16*.05) + (-2*.95) = .8 - 1.9 = -1.1 IMPS each time the opening crops up...

Care to explain why the relative frequency of the hand types is "Hogwash"?

[hrothgar]

pclayton, on Apr 7 2005, 06:41 PM, said:

Usually the responder to the transfer preempt will just accept the transfer and play it there. Positionally, its better for the long suit to be on the table, so transfer preempts will frequently gain in this scenario.

Even when responder does hold a hand that wants to bounce, frequently, the preempter's LHO will have taken a call anyway. Responder will usually be able to guess what hand pattern partner has judging by what action the direct seat takes. I'd say that responder is only at a slight disadvantage.

The issue regarding the transfer preempt is NOT related to the constructive response structure, but rather to defensive methods...

Transfer opening bids provide the opponents with enormous amounts of additional bidding space. Given that you've just opened at the three level, my preference would be to use this space to provide a healthy range of penalty oriented measures...

In short, we're gonna rip you apart when its right and still be able to show a wide variety of constructive hands...

[pclayton]

hrothgar, on Apr 7 2005, 07:55 AM, said:

pclayton, on Apr 7 2005, 06:41 PM, said:

Checking in late on this discussion as usual......

This whole argument about 'frequency' relating to Ben's multi-preempts is hogwash.

Just because a hand with a preempt hand comes up (say) 5x times more often than the strong 2-suited variety is irrelevant.

If you methods cause you to win 16 IMPs 5% of the time, but lose 2 IMPs, 95% of the time, then by definition, the expected value for the method is

(16*.05) + (-2*.95) = .8 - 1.9 = -1.1 IMPS each time the opening crops up...

Care to explain why the relative frequency of the hand types is "Hogwash"?

The prior arguments for

"are multi-preempts good"

vs

"are multi-preempts bad"


were based on a function of frequency. I would argue that frequency of the event has little to do with the outcome, since there are mitigating factors that are very difficult to quantify.

Richard, your idea about sampling 100 (or 1000) hand makes sense however, since you will come up with an aggregate total of gains vs losses, since each hand can be be evaluated on its own merits.

But to say method A wins "x" IMPs "y"% of the time (vs method B ) is the wrong approach.

[pclayton]

hrothgar, on Apr 7 2005, 08:06 AM, said:

pclayton, on Apr 7 2005, 06:41 PM, said:

Usually the responder to the transfer preempt will just accept the transfer and play it there. Positionally, its better for the long suit to be on the table, so transfer preempts will frequently gain in this scenario.

Even when responder does hold a hand that wants to bounce, frequently, the preempter's LHO will have taken a call anyway. Responder will usually be able to guess what hand pattern partner has judging by what action the direct seat takes. I'd say that responder is only at a slight disadvantage.

The issue regarding the transfer preempt is NOT related to the constructive response structure, but rather to defensive methods...

Transfer opening bids provide the opponents with enormous amounts of additional bidding space. Given that you've just opened at the three level, my preference would be to use this space to provide a healthy range of penalty oriented measures...

In short, we're gonna rip you apart when its right and still be able to show a wide variety of constructive hands...

Well, certainly....

In Preempts A-Z, a defense to Namyats is discussed (I think its this - I don't have a copy at hand):

After an opening 4♣ call (showing ♥'s):

Double: Cooperative
Direct 4♥: Takeout of hearts
Direct 4♠: Stronger 4♠ call

After the 4♥ conversion:

Double: Penalty
4♠: Weaker (sac?) than direct 4♠


Its curious that the argument now is focused on transfer preempts in general, rather than Ben's multi structure. If given a choice, I prefer transfer preempts, but the added two-suiter is a net bonus, especially in my system architecture.

[hrothgar]

pclayton, on Apr 7 2005, 07:18 PM, said:

The prior arguments for

"are multi-preempts good"

vs

"are multi-preempts bad"

were based on a function of frequency. I would argue that frequency of the event has little to do with the outcome, since there are mitigating factors that are very difficult to quantify.

Comment 1: My argument was related to the relative frequency of the weak single suited hand relative to the strong 2 suited hand pattern.

Comment 2: The frequency of events actually has a very large impact on their viability:

Case in point: Assume for the moment that you are considering playing the "Hibernian" 3♦. This opening is wonderous thing. Your expected gain from the Hibernian 3♦ is +2 IMPs any time the hand comes up. However, there is a catch... The Hibernian 3♦ opening is very much a top or bottom sort of bid. 50% of the time that this bid comes up, you expect to win 16 IMPs. However, the other 50% of the time, you expect to lose 12 IMPs.

My willingness to play this method would very much depend on its frequency. Assume for the moment that the bid came up once per session. That -12 IMPs means that every other session, we expect to be knocked out of contention because the Hibernian 3♦ opening backfired. In contrast, lets assume that the frequency of the convention was such that it came up 4 times per session. In this case, adopting the convention becomes MUCH more attractive since the frequency of the event dampens the variance...

[inquiry]

Well thanks to richard, we all know that any transfer preempt loses on average 2 imps everytime you use it. Fred should IMMEDIATELY alert his teammates (Ron and Russ) and have them stop this sillyness right away.

But as I showed earlier, the real world frequency of weak to strong hands is not 5 out of 100 as richards latest post uses in his calculations (to over dramaticaly emphasize his point), but more like 4 out of 10 (that high even surprised me). So if his frequency and imp estimates was right 16*0.4 + (-2)*0.6 = net plus 5 imps for the opening 3 bid. Of course, +16 imps is also wrong.

(As an aside, an opening bid of 3 earns these average scores on BBO over 3 million hands....

3C  +0.40 imps,    54.31% MP
3D  +0.41 imps,    58.24% MP
3H  +0.14 imps,    48.60% MP
3S  +0.23 imps,    48.49% MP

So, I guess these "small" average plus is what is at stake here. Do you lose this advantage and go the other way? I think Richard underestimates the problem of bidding over the transfer preempt that may not have the promised suit, and that may be weak or strong, but time will tell. I will post another study, which no doubt he will object too as well. But while he claims "we're gonna rip you apart when its right and still be able to show a wide variety of constructive hands... ", I think this has yet to be shown. So how he gets his +2 imp and his 95% of the hands is still highly questionable. I guess real world experience like that by people who use such bids is a better indicator than what richard thinks will happen. From Fred's post it seems Ron and Russ are happy with it, I wonder if they would be if they are being "ripped up" when they use the transfer preempt?

Ben

[hrothgar]

inquiry, on Apr 7 2005, 09:21 PM, said:

Well thanks to richard, we all know that any transfer preempt loses on average 2 imps everytime you use it. Fred should IMMEDIATELY alert his teammates (Ron and Russ) and have them stop this sillyness right away.

At no point in time did I suggest that the numbers that I provided reflected the expected gains/losses of the particular methods being discussed. Rather, I provided an illustration regarding why the frequency of different methods might matter.

The fact that I chose an example in which one hand type was 20 times more frequent than the other might be considered a clue that this wasn't reflective of the methods being discussed...

[hrothgar]

inquiry, on Apr 7 2005, 09:21 PM, said:

I think Richard underestimates the problem of bidding over the transfer preempt that may not have the promised suit, and that may be weak or strong, but time will tell.

When devising defensive methods, I'd pretty much ignore the possibility that you held a strong 5-5 hand. If I hold a strong hand, the conditional probability that you also hold a strong hand drops fairly significantly...

I don't claim that I'll never get burned, however, I'm willing to tolerate the rare chance of a large loss for a decent chance to pick up some IMPs here or there...

>I guess real world experience like that by people who use such
>bids is a better indicator than what richard thinks will happen.
>From Fred's post it seems Ron and Russ are happy with it, I wonder if
>they would be if they are being "ripped up" when they use the transfer preempt?

I've see a LOT of weird methods in my time. This is the second time that I can recall seeing a transfer preempt in an otherwise natural bid.

Bergen and Cohen used a 2-under preempt style, however, this was a deliberate choice to compensate for the extremely undisciplined preemptive style that Bergen used. The pair needed extra bidding room to sort out where to play.

I have no knowledge regarding why Ekblatt's partnership has adopted this stuff. Certainly my own track record versus Russ leave me little room to critique his methods... With this said and done, I suspect that his 3♥ bid must encompass multiple meanings (maybe a good bad split range?) to justify the transfer...

[pclayton]

Ben:

I'm going to steer this discussion off on a tangent for a second. Did you and Misho discuss the possibility of these multi preempts in sort of a "suction" format?.

Here's what I'm thinking:

2N = preempt in ♣'s OR ♦'s + ♥?

3♣ = preempt in ♦'s OR ♥'s + ♠?

3♦ = preempt in ♥'s OR ♥'s + ♣?

3♥ = preempt in ♠'s OR ♣'s + ♦?

3♠ = solid suit OR ♣ / ♥ or ♦ / ♠

The reason why I'm considering this is because the 3 loser hand is solved easily enough by my 2♦ opener. Because its such a moose, we have room to figure out what we need to know.

So what I'm thinking is that the big 2 suiters are 4 - 5 losers. But also - since Opener can have a smaller set of hands, that responder can bounce; to preempt, OR to show a certain # of cover cards.

Thinking out loud - 3♣ - 4♥ might show three cover cards, but a club / heart fit. If the hand fits hearts - great - we might have a slam, otherwise, it might be a hand where its a cheap sac against 4♠. Something like: xx, Axxx, Axxx, Kxx feels about right.

[inquiry]

Here is an independent study, so you can validate what is happening (I assume most of you don't have bridgebrowser). One of richards claims is that I "cherry picked" good examples for my method. Another is that the frequency question of weak to strong, and a third is how well does it work.

So I approached this in a way anyone can check to be sure I was not “hand picking” the strong variety of hands, I started by gong though the BBO vugraph archives looking for hands that fit the requirements for MisIry love Company hands. These are, in general 4+ controls (four requires exceptional suit quality or extra length). No more than 4 losers, at least 10 cards in two long suits, no fewer than five either. And a minimum of around 15 hcp, which can be shaded downward with a void and 6-5 or 7-5, it is fitting cards, not HCP that win tricks here.

I stated with the first two Italian champtionship sessions from today and go down the list of archived vugraph to Icelandic team trials, first session. For Frequency testing, this accounted for 512 hands of bridge, and 13 of them fit the requirements of strong, 4 losers or less, and they get to open (There were a few others, but you didn’;t get to open). For the record, this is a frequency rate of 2.53%, higher than I estimated. It is also worth noting that there were only six hands where the table I was checked for hand patterns opened with a preempt at the three level in a suit (lots of weak twos and opening 4 bids, but so few at three level). So in these hands, the rate of strong to weak would ACTUALLY have favored the strong hands by 2 to 1.

Below I show the 13 strong hand varieties, suggest the MisIry bidding sequence, and describe what happened at the table. Judge for yourself its usefulness, the frequency, and how hard (or easy) these auctions are.

This is the first I found Victorian Pennant SF 4 of 4, board 20

X
AKxxx
Qx
AKQxx

AQx
J9
ATxxx
J9x

One pair ended up in 4H, lost two tricks. The other ended up in slam, but played 6D om 5-2 fit. That actually made. 6C is the best slam, of course.

3C – 3D – 3N - ? This is an ackward one for South. Two aces usually are worth two cover cards, even if partner is void in one side suit, as a pitch is possible from the other. The AQ of spades might be worth two on a hook. The doubleton heart is possibly worth a cover card if the heart queen is missing. There is too little room to investigate, however, as the bid to show this hand would be 5D (ace of spades, and K or A of dicamnods). But with support for clubs, that takes you beyond the “Safe” level for five clubs. So here, Responder has to guess to try for 6C or perhaps stop and ply 3NT (wrong sided no doubt), or bid 4H/5C. I think 6C is a reasonable gamble, as exploratory room does not exist, and the J9's in both partner's suit might prove useful after all. This hand actually makes 7C on the spade hook because the heart queen is missing (and hearts are 3-3).

Hand 95 from Brazilian Team Trial finals, was
AKQxx
AKxxx
T
Kx

Txx
Xx
A9xx
Qxxx

3C-3D-4H-? < - - opener shows 5-5 or better in majors, 3 losers. South has diamond ACE, likely cover, and doubleton heart, maybe cover if wayward queen. We haven’t covered how to proceed here. We will save this hand for later. Note, Not the best slam as requires great luck in the heart suit. but if you push to slam, it happens to make. I just point out you know for slam, you will have to be ruffing a heart, and partner will have to be missing the Queen.

Hand 96, same event.
QJ9xx
Void
AKQJx
AQx

T87
Qxx
986
KJTx

3D-3H-3S-4S-pass. Easy. One pair got to 6S, which wasn’t a success. Remember the 3S bid can be three or four losers, and with fit, responder bids game. You will see this type of hand fairly often. If partner bids 3H showing major two suiter, you NEED a cover to bid game, as that is always 4 losers.

Board 73 of C_N_Echipe Diviizia A standza 5 is interesting. Both pairs stopped in 6H. With MisIry, grand slam is "automatic". Let's see why.

AQ9xx
AKxxx
A
Qx

Kx
QJxxx
Qxx
ATx

3C-3D-3H-4D-4S-5D-7H

Here 3H = major 2 suiter, exactly 4 losers. 4D deines diamond control while showing Ace or King of clubs, and shows slam interest, 5C is third non-signoff step (first is 4S, 2nd is 4N), it shows opener is missing heart queen (we knew that), but promises the spade queen. With both missing queens and no grand slam chance (club ace missing), opener would jump to 6H (pass/correct) as responder must have enough cover cards since all his “maybe’s” are working (and if he had no maybe's he would have signed off instead instead of 4D). If partner has the club king (so that he knews we have the club ACE), he would bid 4NT, so when our “club ACE” is working, we show where we hold replacement for a wandering queen, and rather we have sure 3 covers or 2 covers with this wandering queen cover. So here, 5D shows not only missing heart queen, but also shows missing club King, you will find this bid useful on LOTS of hands.

So responder knows his partner has two clubs, without the King, five spades, no K, five hearts, no Q. That is his four losers, so he must have diamond void, or more likely diamond ACE stiff. We get one club pitch on the Spade Queen, and one on the fith spade, so we take 5H, 4S, 1C, 1D, and 1S ruff, and 1 club ruff for 13 triccks, so 7H is bid. No doubt spiral scan finds this as well, but otherwise this is hard to duplicate by other methods.

Board 7, Icelandic championship, round 11
Jx
A9xxx
Void
AKQTxx

5
KJx
AQxxxxx
8x

3C-3D-3N-6H

Not as nice as other auctions, but no easy way to bid. The KJx of hearts might be two covers (it turns out it was), Diamond AQ might be 2, likely 1 but maybe none (it was none), spade stiff might be one (it was). Six hearts makes on a heart hook. Neither team sniffed at slam, at 50% (or less), maybe best not too, but would work here.

If you bid 4NT as south instead of 6H, north knows the diamond honor is “wasted” and will bid 5C and you end up in five hearts.

As the last hand showed, not all hands are magic. Here is board 29 from Brazilian Team Finals, semi-finals (2 of 4).

KQT87
A9
6
AQT54

93
QT2
AQ98xx
J7

3C-3D-3S-4S-all pass. As you can see, 4S is not horrible contact. You may escape losing a club, a heart and a spade. 3NT is down on a heart lead (it made at both tables). 4S has some play (clubs 5-1, spades 3-3, diamonds 1-5) but is not a sterling contract. I know 3NT can be beat, and it looks like 4S will be hard to make too.

QT
AKQ986
KTxxx
Void

Axxx
Void
QJ932
Qxxx

3H-3S-3N-4C-4N-6D

3N showed red 2 suiter, 4 loser. 4C showed no control in clubs, 4N was third non-signoff step, it showed, whatever is spades is working (hence partner has xx or Ax, but we are looking at ACE so he has two without the king), and that he (opener) is missing diamond Queen (we know that anyway). This also shows that partner has the heart queen. So we know, three of his sure four losers, which are two spades and the diamond queen. With so many diamonds, we can envision the chance to throw all our spades away if partner is missing a club or diamond honor (then he must have AKQ of hearts). If he is missing a heart honor, then we might avoid heart loser all together due to our five card support in diamonds. So, we bid the slam. Both teams easily bid this as well, perhaps with less certainty after 1H-1S-3D-5D-6D. They could be off two quick spade losers.

Here is a more fun one from Icelandic Team trials, round 9 (board 9).
AKTxx
X
A
AJT982

8
AKQJ53
xxxxx
x

3C-4H-6H is one quick way, 4H shows solid type suit, desire to play in hearts, even if partner has club-spade two suiter. With lesser type holdings bid 3H then 4H. One pair reached 5S down one, the other got nicely to 6H. If you like as opener you can invite to six hearts, partner will bid it. This hand is huge win for these methods

Icelandic Team trials, round 8, Board 5
xxx
QTxx
xxx
9xx

AQxxx
AK
AKxxx
2

3D-3H-3S-4S,

Ok, 4S with NO potential covers is hard to swallow. At least on earlier hand you had the king of clubs or king of diamonds as possible cover. But the rule I live by is over 3S with a fit, carry on to game. Since partner may have 3 loser hnad for the 3S rebid. Pass of 3S is also a reasonable option (The rule is to bid 4S with fit even without cover, but is 3433 a fit?) 4S makes as there is a doubleton spade King.

The very next hand (board 6) was
AQJT7
Void
AQx
KJxxx

Xx
Jxxxx
KJ9
xxx

3C-3D-3S-4C-pass

Here you have option of playing 3S (don’t raise, no fit), or 4C. 4C makes, 3S is problematic on heart lead. Also here, the diamond KING is not a sure cover (less than 50% cover). Remember if partner has Great 3 losere hand with nice spades, he would bid 4S (exceptional 3 loser hand), which is what you would need to make 4S with this doubleton support. Eiethe contract is better than 3NT which was played with these cards at one of the tables. And since 4C makes, it beats 2S making 2 (Which is what happened at the other). These methods also avoid the hopeless 3NT contract.

Here is a hand from the same match… This type of hand lacks the right number of controls (only 3) and HCP, but if you choose to open these with great shape and when the major is strong, you will still do well. I show this as an example. Board 10

AKQxxx
Void
Kx
JT8xx

xx
xxxxx
ATxx
xx

3C-3D-3S-4S-pass

Here diamond ace is a more likely cover than diamond King on earlier hand, and 4S is as good a guess as 4C since you have two cards in each. As you can see 4S is surely likely to make, and can make five. The contracts were 2S and 5Hx when this hand was played. I would open this hand 3C because I am never letting them play 4H, I may as well get both my suits into the auction.

Board 7 Icelandic Team champtionship, round 3
Void
KJ
AQJT95
KQxxx

AKQJx
Txxx
Xxx
X

3D-3S-4C-5D-pass (after 3S, 4H would be responders own two suiter), or
3D-3H-4C-4H-5C-5D-pass

A word about this second auction. 5C is normally “we are off two quick tricks in hearts”. But it can also safely be, “any cover cards in the other major is wasted”. The reason why, if partner had three sure cover cards (say DK, C-AK), he will bid the slam despite your 5C bid. You simply can not be off two quick hearts in that case. But this warns that AKQJ of spade is worth exactly nothing.

Ben

[inquiry]

pclayton, on Apr 7 2005, 03:17 PM, said:

I'm going to steer this discussion off on a tangent for a second. Did you and Misho discuss the possibility of these multi preempts in sort of a "suction" format?.

Here's what I'm thinking:

2N = preempt in ♣'s OR ♦'s + ♥?

3♣ = preempt in ♦'s OR ♥'s + ♠?

3♦ = preempt in ♥'s OR ♥'s + ♣?

3♥ = preempt in ♠'s OR ♣'s + ♦?

3♠ = solid suit OR ♣ / ♥ or ♦ / ♠

The reason why I'm considering this is because the 3 loser hand is solved easily enough by my 2♦ opener. Because its such a moose, we have room to figure out what we need to know.

So what I'm thinking is that the big 2 suiters are 4 - 5 losers. But also - since Opener can have a smaller set of hands, that responder can bounce; to preempt, OR to show a certain # of cover cards.

Thinking out loud - 3♣ - 4♥ might show three cover cards, but a club / heart fit. If the hand fits hearts - great - we might have a slam, otherwise, it might be a hand where its a cheap sac against 4♠. Something like: xx, Axxx, Axxx, Kxx feels about right.

Misho actually prefers, I believe to use 2NT as transfer to clubs, or a strong two suiter with hearts and not clubs.

Then 3C is transfer to diamonds, or strong two suiter in the black suits,
then 3D is transfer to hearts, or strong two suiter with spades and diamonds
then 3H is transfe to spades, or strong minor two suiter.

So all hands with hearts start 2NT, all other two suiters have the first suit bid and not the "transfer" suit. I simply refused to give up 2NT. I still liek to play with the field, as I humbly think I play as well or better than the average player. Giving up 2NT was just too radical for me despite the fact that I have multi 2D to show a big hand.

Ben

[pclayton]

hrothgar, on Apr 7 2005, 09:13 AM, said:

Case in point: Assume for the moment that you are considering playing the "Hibernian" 3♦. This opening is wonderous thing. Your expected gain from the Hibernian 3♦ is +2 IMPs any time the hand comes up. However, there is a catch... The Hibernian 3♦ opening is very much a top or bottom sort of bid. 50% of the time that this bid comes up, you expect to win 16 IMPs. However, the other 50% of the time, you expect to lose 12 IMPs.

My willingness to play this method would very much depend on its frequency. Assume for the moment that the bid came up once per session. That -12 IMPs means that every other session, we expect to be knocked out of contention because the Hibernian 3♦ opening backfired. In contrast, lets assume that the frequency of the convention was such that it came up 4 times per session. In this case, adopting the convention becomes MUCH more attractive since the frequency of the event dampens the variance...

This is an interesting thought - although I don't know that I follow, (or agree) with all the logic.

Certainly there are other "high-variance" calls that are realtively standard. Some that I can think of are the stopperless 3N call in the balance chair, overcalling 'under the gun' vulnerable over a high-level preempt, etc.. Just because my downside might be significant doesn't mean the call is wrong, because of the potential upside.

Going back to your example of the Hibernian 3♦. The EV of the call seems to be +4 IMPs, although the downside may be huge.

I think you are trying to equate these high-variance situations to the following:

You want to give me 2:1 odds on a coin-flip for a million dollar bet. While it makes economic sense to take the bet, if I lose, I'm wearing a barrel and making a sign of out cardboard boxes advertising my ability to work for food.

However, if I'm playing a match against Meckwell, I definitely want to INCREASE my variance, especially when the variance call has a positive EV. If I'm playing a weaker team, I agree with Richard - keep the variance lower. Even if the EV is a slight negative, I still want to get the throw some variance into the game.

Sometimes you have to cut the dogleg, sometimes you have to keep it in the middle.

[hrothgar]

inquiry, on Apr 7 2005, 10:35 PM, said:

Here is an independent study, so you can validate what is happening (I assume most of you don't have bridgebrowser). One of richards claims is that I "cherry picked" good examples for my method. Another is that the frequency question of weak to strong, and a third is how well does it work.

So I approached this in a way anyone can check to be sure I was not “hand picking” the strong variety of hands, I started by gong though the BBO vugraph archives looking for hands that fit the requirements for MisIry love Company hands. These are, in general 4+ controls (four requires exceptional suit quality or extra length). No more than 4 losers, at least 10 cards in two long suits, no fewer than five either. And a minimum of around 15 hcp, which can be shaded downward with a void and 6-5 or 7-5, it is fitting cards, not HCP that win tricks here. 

I stated with the first two Italian champtionship sessions from today and go down the list of archived vugraph to Icelandic team trials, first session. For Frequency testing, this accounted for 512 hands of bridge, and 13 of them fit the requirements of strong, 4 losers or less, and they get to open (There were a few others, but you didn’;t get to open). For the record, this is a frequency rate of 2.53%, higher than I estimated. It is also worth noting that there were only six hands where the table I was checked for hand patterns opened with a preempt at the three level in a suit (lots of weak twos and opening 4 bids, but so few at three level). So in these hands, the rate of strong to weak would ACTUALLY have favored the strong hands by 2 to 1.

Now I'm REALLY confused

As I noted earlier, two suited hands with 5+/5+ shape are pretty rare

5-5-2-1 patterns make up 3.17% of hands
5-5-3-0 patterns make up .9 % of hands.
6-5-1-1's are .71%
6-5-2-0s are .65%

All told, this makes up 5.63% of hands. Even if we add in 7-5s and the like, we aren't going to get up past 5.8% or so...

The interval 17-40 HCP encompases ~5.45% of all hands
16 HCP = 3.311%
15 HCP = 4.424 %

Hands with 15+ HCP = 12.78% of the hands

The combination of BOTH 5+/5+ shape and 15+ HCP is extremely rare.
Roughly .7% of all hands would qualify for the strong hand type.

This back of the envelop calculation is pretty close to AWM's script which estimated the frequency at .826% (I suspect that the difference is based on the fact that he was strictly looking at losers while I'm stuck with HCP)

The data that you are presenting shows that the strong hand variant is four times are frequent across a reasonably large sample. Worse yet, you're only looking at Opening bids... The odds of this happening are absurdly low. We are INCREDIBLY far out of the tail of a distribution... (Far enough tht I'd be very suspicious of the Hand Generator being used)

I haven't played much with Bridgebrowser and don't really know is capabilities; however, could you please check the following:

First of all, you noted that you looked at 512 hands of Bridge. However, how many BIDS did you consider? For example, if the auction start P - P - 3♣, this would account for three bids and might substantially modify your frequency estimates...

Next: Only look at North Hands (I'm not interested in who opened or gets a chance to open, but rather, simple questions of frequency)

1. What percentage of the hands have a 5-5 or 6-5 hand pattern?

2. What percentage of the hands have 15+ HCP?

3. What percentage of the hands have both 5/5 and 15+ HCP?

4. What percentage of the hands have both 5/5 shape AND 4 or fewer losers AND less than 15 HCP?

[hrothgar]

pclayton, on Apr 7 2005, 11:28 PM, said:

Going back to your example of the Hibernian 3♦. The EV of the call seems to be +4 IMPs, although the downside may be huge.

EV = 50%(+16) + 50%(-12) = 8 - 6 = 2

You point relating variance to the qaulity of the oppsoition is certainly valid. COnsider the following: If your main goal in playing bridge is to accumulate master points, the difference between placing 12 and placing 13th is immaterial, however, the difference between 1st and seconde is HUGE...

(Completely off topic, I'm busily reading through some books on the derivatives markets right now. There is actually a formal term - "negative convexity" - that refers to this type of phenomena)

[mikestar]

pclayton, on Apr 7 2005, 08:28 PM, said:

hrothgar, on Apr 7 2005, 09:13 AM, said:

Case in point:  Assume for the moment that you are considering playing the "Hibernian" 3♦.  This opening is wonderous thing.  Your expected gain from the Hibernian 3♦ is +2 IMPs any time the hand comes up.  However, there is a catch...  The Hibernian 3♦ opening is very much a top or bottom sort of bid.  50% of the time that this bid comes up, you expect to win 16 IMPs.  However, the other 50% of the time, you expect to lose 12 IMPs.

My willingness to play this method would very much depend on its frequency.  Assume for the moment that the bid came up once per session.  That -12 IMPs means that every other session, we expect to be knocked out of contention because the Hibernian 3♦ opening backfired.  In contrast, lets assume that the frequency of the convention was such that it came up 4 times per session.  In this case, adopting the convention becomes MUCH more attractive since the frequency of the event dampens the variance...

This is an interesting thought - although I don't know that I follow, (or agree) with all the logic.

Certainly there are other "high-variance" calls that are realtively standard. Some that I can think of are the stopperless 3N call in the balance chair, overcalling 'under the gun' vulnerable over a high-level preempt, etc.. Just because my downside might be significant doesn't mean the call is wrong, because of the potential upside.

Going back to your example of the Hibernian 3♦. The EV of the call seems to be +4 IMPs, although the downside may be huge.

I think you are trying to equate these high-variance situations to the following:

You want to give me 2:1 odds on a coin-flip for a million dollar bet. While it makes economic sense to take the bet, if I lose, I'm wearing a barrel and making a sign of out cardboard boxes advertising my ability to work for food.

However, if I'm playing a match against Meckwell, I definitely want to INCREASE my variance, especially when the variance call has a positive EV. If I'm playing a weaker team, I agree with Richard - keep the variance lower. Even if the EV is a slight negative, I still want to get the throw some variance into the game.

Sometimes you have to cut the dogleg, sometimes you have to keep it in the middle.

Richard is quite correct that a high probability high variance method is a better choice than a low probabiltity high variance method.

To make it clear, let's look at your 2:1 on a coin toss example. For a single toss, your expectation is to win $500,000 but the bet is scary because 50% of the time you lose everything.

Now let's say the coin will be tossed twice, with the same bet on each. Now you expect to win $1 million, but your chance of losing money has gone down to 25%: You are +$4M if you win both tosses (25%), +$1M if you win one and lose one(50%), and -$2M if you lose both(25%).

Add more tosses and your expectation goes up and your chance of losing goes down.

If you are interested in more in-depth explanation of the concept, google "gambler's ruin": http://www.google.com/search?hl=en&q=gambl...G=Google+Search