[inquiry]

So far, my studies have been how to optimize bidding with the strong hands. I did do a study of what I will be missing by not being able to open 3C. An opening 3C is a small winner over thousands of hands at both imps and matchpoints. So I know I go in with a slight deficiet for that reason. What anticipate that I will make up for this on the big hans. The ratio of bid hands to preempt hands is roughly 1 to 3 to 1 to 4.

On the other hand, it is hard to caluclute the net negative affect of a 3D bid with hearts when everyone else opens 3H. A couple of issues,

3H-X-P-P-P where they got you, cancels out with
3D-(P)-3H-P-P-X.. .I the first case, WEst makes take out, East passes, in second case WEST can't make takeout but can make penalty.

So there is some finnagliing needed to even estimate this. IF you ahve any ideas how to do it, let me know.

Next, sometimes the bidding will go...

3H-P-5H/6H meant as preemptive. I can't see it every gong
3D-P-5H - unless responder has big diamond fit as well, so this really bump the preempt weapon is removed.

Next, the psyche-type 3NT is still here.. on this type auction
3H-(P)-3NT, now becomes
3D-(P)-3NT, if partner has a big hand, you will hear about it now.

[hrothgar]

inquiry, on Apr 6 2005, 12:21 AM, said:

On the other hand, it is hard to caluclute the net negative affect of a 3D bid with hearts when everyone else opens 3H. A couple of issues,

3H-X-P-P-P where they got you, cancels out with
3D-(P)-3H-P-P-X.. .I the first case, WEst makes take out, East passes, in second case WeST can't make takeout but can make penalty.

So there is some finnagliing needed to even estimate this. IF you ahve any ideas how to do it, let me know.

As you note, issues that are dependant on whether a specific hand is held in direct seat or balancing seat will cancel out and are largely irrelvant to the discussion at hand.

What would worry me is the following:

Your transfer preempt provides the opponent's with significant extra bidding room.

If I open a 3♦ preempt showing hearts, I'm giving RHO two shots at the apple. RHO can doube 3♦, cuebid, 3♥, and pass then double 3♥, bid a direct 3♠, and pass then bid 3♠ to show different types of hands...

I haven't spent much time thinking about an optimal defense to this method, but I'd probably use something like the following

Direct seat double of 3♦ = Cooperative oriented penalty double of 3♥
Direct seat 3♥ Cue bid = "Classic" takeout double
Direct seat 3♠ = Standard 3 Spade overcall
3NT = To play, typically based on a running minor
4♣ = Clubs and Spades
4♦ = Diamonds and Spades
4♥ = Hearts
4♠ Good hand with Spades

Assume for the moment that I pass...

When 3♥ comes back to me...

X = Pure penalty double
3♠ = Weak takeout of hearts
3NT = Both minors
4♣ = Single suited with Clubs
4♦ = Single suited with Diamonds

I'm sure that with a little work it would be possible to make dramatic improvements on this structure...

In short, the transfer oriented preempt gives the opponents enormous opportunity to desribe their hands... I wuldn't go so far as to say that it completely destroys the vlaue of the preempt. However, methods that permit the opponents to easily show

1. A cooperative oriented penalty double of our 3 level preempt
2. A takeout oriented double of our 3 level preempt
3. A pure penalty double of our three level preempt

While simultaneously assisting in showing constructive hands would give me significant pause for thought

[inquiry]

Yes, you get several bites of the apple, but not all is eqaully delicious. Let's take a look at a few of the bites....

Bite one...
"Direct seat double of 3♦ = Cooperative oriented penalty double of 3♥"

Bidding might go 3D-(X)-P-? What does your partner do, especially if not particularily strong? If he passes, the auction maybe over when opener has a strong hand with diamonds. Remember, parnter of the opening bidder is not forced to bid if you double. Ouch. Better be sure about this, the apple will bite back 30% of the time.

Bite two...
"Direct seat 3♥ Cue bid = "Classic" takeout double",
This one can bite back as well. First way this can bite back is when at the other tables it went 3♥-X-all pass. Your 3H cuebid takeout has taken this weapon away. Second way this could be wrong is when you are short in hearts, and opener is strong.. because he lacks hearts too. You are surely in trouble now....

The other bids seem ok, although the 4H bid surely helps the other side as far as I can see as often as it helps you. Of course when you bid the two suit things, this will often be more painful when opener doens't have hearts, because he surely has one of your suits, and he doesnt have hearts and neither do you, which increase misfit changes.
4♣ = Clubs and Spades
4♦ = Diamonds and Spades
4♥ = Hearts
4♠ Good hand with Spades

Ok, now we look at the hand where you pass, and the bid comes back to you in three hearts...
Ok, when you can double 3H for penalty, that is great, and a big win for you. But what about hands where you can not make a penalty double? You may have a lot of problems competing now. If you bid 3S as weak takeout, you just let us off again when your partner is stuffed in hearts. This means your direct actions might have to weaker when you don't have hearts, which will hurt both your cooperative doubles and your direct cue-bids, putting more pressure on them.

The 3Nt minor two suiter, the 4C/4D natural is ok. But of course, the preempt has done what it was meant to do, put the last guess back on you.

For sure, more options is good for the defense. I don't mean to suggest that it isn't. But, there is also more risk too, when you take an action based upon opener having the transfer suit. If transfer WAS ALWAYS the transfer suit and always weak, you can eat it alive.... I agree... I am not so sure about these preempts, but my experience is fairly limited in competition so far.

Ben

[hrothgar]

inquiry, on Apr 6 2005, 04:50 AM, said:

Yes, you get several bites of the apple, but not all is eqaully delicious. Let's take a look at a few of the bites....

Ben, I don't think that its fruitful to argue about the specifics of this defense.
As I noted, its just something that i threw together off the top of my head.

The real point that you need to consider is whether you want to trot this method out against opponents who do have a good defense.

I suspect that the losses when you hold a weak hand are going to be frequent enough and large enough to swamp any benefits that acrue from slam invitational 5-5s...

[hrothgar]

inquiry, on Apr 6 2005, 12:21 AM, said:

The ratio of big hands to preempt hands is roughly 1 to 3 to 1 to 4.

This doesn't sound quite right... I don't have the right tools to do a script right now (running into trouble with my memeory manager). So, lets do a rough approximation


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

There are 6 2 suited hand patterns. Diamonds + A Black suit make up 1/3rd of all these hands. So, we're looking at roughly 1.87% of the hands.

In contrast, if we look at single suited hand patterns

6-3-2-2 patterns = 5.64%
6-3-3-1 patterns = 3.45%
7-3-2-1 patterns = 1.88%
7-3-3-0 patterns = .27%

Here we're looking at 11.24% of the hands...
Lets be generous and assume that half of those hands are too flawed to consider a 3 level opening. Lots of those 6322 are going to lack texture or some such. This takes us down to 5.62%. So, the frequency is pretty much equivalent.

Now, lets look at HCP...

Lets assume that a weak 2 bid shows roughly 6-9 HCP...
This interval encompasses roughly 32.83% of the hands..

At the other extreme, lets define strong two suited hand pattern as 17+ HCP...
The interval 17-40 HCP encompases ~5.45% of all hands

From what I can tell, your weak hands are roughly six times more likely than the strong variants.

Personally, my style of bridge has always emphasized trying to get things right with the common stuff and let the freak hands take care of themselves...

These methods seem suicidal at MP. Even playing IMPs, I can't see your gains being significant enough to compensate for your loses...

[inquiry]

Riichard,

Well, part of what I want to do is explore things like best defense. I happen to like pass then double as penatly. I like pass then rebid 3NT as minor two suiter. I like nonleapoing michales. I am not sure I like 3H as takeout, but 3Dx as cooperative penalty double imaybe ok. I thave to think about it. I tihink, however 3Dx should show diamonds, to prevent the 3Dxppp when opener has diamonds....

Ben

[awm]

My statistics indicate that the strong hands are surprisingly frequent -- up to 1 out of 3 transfer preempt openings if your three level bids are always seven card suits, or a bit worse if you often preempt six carders.

The loss of 3♣ as a preempt is significant, but you could certainly revise the system to include 2NT as either weak with clubs or a two suiter (basically what Ben originally proposed). If you're playing a strong club system, or certain versions of multi, you don't really need a 2NT opening and this bid is basically free.

So I think the real questions to be answered are:

(1) How much do you win/lose when you open a transfer preempt with a strong hand?
(2) How much do you win/lose when you open a transfer preempt with a weak hand?

I think we can agree that (1) is more wins than losses and (2) is more losses than wins. Of course, (2) especially depends a LOT on how your opponents defend the method. Then we need to weigh these things by the relative frequencies (case 2 is 2-3 times more frequent). Still haven't seen any examples of what happens at the table when opener has the weak hand and opponents have some reasonable defense in place.

I'd suggest the following defense:

Direct seat Pass = either a weak hand, or a strong single suiter
Direct seat X = cards, usually balanced around 14/15+ points, promises some cards in the actual bids suit (but Hxx is enough); this encourages partner to make a light takeout double after the transfer is accepted.
Direct seat bid (not accept of the xfer) = competitive, not super-strong
Direct seat "accept" of transfer = stopper ask, same as 2♥-3♥ for example
Direct seat 3NT = to play, but serious values (not the hands where you "guess" a 3NT on 16 bal)

If direct seat passes, then partner basically bids as if the weak bid was opened. After direct seat X and accept of transfer, partner is encouraged to act with light values. If direct seat passes and transfer accepted, no need to be super-aggressive with takeout shape.

In balancing seat after 3♦-P-3♥-P-P or the like:

Double = balancing takeout, weakish and good shape
Suit bid = a GOOD hand, forcing to game (otherwise would've bid directly to compete over strong hand possibility)

Double then double again is also takeout, but shows sound values.

Playing this defense, I don't see any real losses when compared to non-transfer weak hands. The person in direct seat gets to distinguish strength, gets to make stopper asks that usually aren't possible, and you can penalize a bit more often because of the value-showing double (which encourages partner's light takeout double in hrothgar's example). You avoid some of the blind guess 3NT bids after a 3-level preempt.

[Gerben42]

There is an important selection effect. If you have the weak variant and you are not in 1st seat, opponents are more likely to open in front of you than when you have the strong variation.

For example in 3rd seat:
The frequency of 6 - 9 HCP = 21.4% and no longer 32.8%
The frequency of 17+ HCP = 14.3% and no longer 5.5%

[inquiry]

Gerben42, on Apr 6 2005, 04:14 AM, said:

There is an important selection effect. If you have the weak variant and you are not in 1st seat, opponents are more likely to open in front of you than when you have the strong variation.

For example in 3rd seat:
The frequency of 6 - 9 HCP = 21.4% and no longer 32.8%
The frequency of 17+ HCP = 14.3% and no longer 5.5%

This selection affect is true. In fact, third seat is most frequently strongly hand, and third seat does have more frequent strong than first seat.

Ben

[Free]

yes, in 3rd seat you have a lot higher frequency of strong versions, but I've noticed that when you have a normal preempt, you'll experience more trouble as well. In those situations you really want to bid natural.

[inquiry]

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).

[inquiry]

awm, on Apr 5 2005, 11:22 PM, said:

So I think the real questions to be answered are:

(1) How much do you win/lose when you open a transfer preempt with a strong hand?
(2) How much do you win/lose when you open a transfer preempt with a weak hand?

I am willing to accept that, strong two suiters win in your case one, and by a fairly large margin. I am willing to accept that in case number 2, you have tilted the odds in your opponents favor a little bit when you hold the weak hand, but not as much as you may suspect. But because of the higher frequency of these bids, even a small amount will get amplified.

However, what you haven't factored in is a case #3. Where you DONT open 3C/3D/3H but show a strong two suiter naturally. Consider, this run of the mill auction....

1H-1S
3C...

and

1S-1NT (forcing)
3C...

How does not having the strong two suiter (however you minimum requirements apply) affect these auctions? You will be surprised.

Also since none of us like opening 2C with a strong two suiter, it also has some significant affects on 2C auctions where opener bids two suits (he can never have five plus in both suits).

So what I am suggesting is that it also provides wins on other auctions where you didn't open 3C/3D/3H as a strong two suiter, placing reasonable and definable limits on what these other bids show. And that has proven to be a big win for us, as we tick off in our heads the kind of hands partenr can have for his jump rebids based upon our own hand and bidding to date. It becomes much easier to visualize (since you will not invision a true monster).

Ben

[inquiry]

Traveller for Board 8 BBO tourneyment

Just ran across this little hand. Bidding the slam was worth 6imps in huge field.

Scoring: IMP

West North East South

Pass Pass 3♣ Pass
3♦ Pass 4♠ Pass
4NT Pass 5♣    Pass
 5♥ Pass 5♠     Pass
6♠    All Pass
4S shows club-spade two suiter, with 2 losers. West with spade queen ALREADY knows at least 6S will be bid. Maybe heart singleton is useful, maybe king of diamonds is useful, so he explores this possibility. 4NT as always shows second round control in hearts. That is not useful, so opener bids 5C. At this point, responder can now try to see if his diamond king so he bids 5♥ (promising a diamond value). Again opener knows heart ace is missing and no diamond card will be useful,, so he signs off in 5S. Of course, responder carrie to slam disappointed that neither red suit value was useful. He could just bid 6S over 4S and not try for grand slam, but that seems timid.

---

 

[hrothgar]

inquiry, on Apr 6 2005, 07:25 PM, said:

Traveller for Board 8 BBO tourneyment

Just ran across this little hand. Bidding the slam was worth 6imps in huge field.

Three comments:

1. As noted earlier, displaying hands that suit your methods does very little to validate the transfer preempts. For Christs sake, if have a mechanism to show a 5-5 and 17+ HCP, you had better get decent results when these hands occur.

2. 40 pairs played this boards. 19 of the 20 were bid the slam. Me, I normally assume that standard methods would allow me to score in the top 50% of the field on BBO...

3. If you are actually interested in advancing this discussion, the most useful information that you could provide would be information regarding the relative frequencies of the different hand types. As I noted earlier, my back of the envelope estimation suggests that the strong 5-5 hand occurs approximately once in 400 hands. I would hope that you have a more accurate analysis available...

[inquiry]

Gee richard, if you don't want to participate please don't bother. I point out this hand because it was referred to me by an interested party. I neither played in this event nor kibitzed it...and it is not in bridgebrowser database yet. There was a hand in Jr Flght yesterday, and BridgeBrowser tourney last night that also would have worked out well. As I said, I have thousands of these hands, downloaded in order (aka random), I would gladly in groups of 20 to interested people who want to try them out. There are lucky slams where three finessess later, the slam makes that this method will miss. There are hands that belong at the two level that this method witll get too high on (usually based on total misfit and horrible splits, btw). But the vast majority work well.

As for this hand, sure many people bid this slam easily enough (or it would be more than 6 imps), but change openers hand to Heart Ax instead of Kx, how many would have found the grand slam? This method makes it easy. Look at the auctions that got them to six, most EAST just blasted once partner showed "support", hopeing not to lose two hearts or one heart and one spade.

So far, your contribution has been limited to some useful ideas on defense to the opening bid, and a general critisim that in your opinion the transfer preempt concept is so horrible it will never work (not in so many words). I got it. I understand your "feeling". But unlike you I am still exploring the usefulness of this bid, and not only as it deals to when you use it, but also to when you don;t.


The fact is, I am still working on five major issues.
1) Best way to explore usefulness of "maybe covers" (side AK, and things like doubletons and five card support).
2) Best way to show type of hand over interference (it turns out to be easy to get across the two suits, but you lose some flexibility in describing if you ahve two or three losers.
3) Best type of interfernce.
4) How bad is the disadvantage with weak openers.
5) Waht happens to other auctions when you don't open 3C/3D/3H

Ben

[hrothgar]

inquiry, on Apr 6 2005, 08:00 PM, said:

Gee richard, if you don't want to participate please don't bother. I point out this hand because it was referred to me by an interested party... So far, your contribution has been limited to some useful ideas on defense to the opening bid, and a general critisim that in your opinion the transfer preempt concept is so horrible it will never work (not in so many words). I got it. I understand your "feeling". But unlike you I am still exploring the usefulness of this bid, and not only as it deals to when you use it, but also to when you don;t.

The fact is, I am still working on five major issues.
1) Best way to explore usefulness of "maybe covers" (side AK, and things like doubletons and five card support).
2) Best way to show type of hand over interference (it turns out to be easy to get across the two suits, but you lose some flexibility in describing if you ahve two or three losers.
3) Best type of interfernce.
4) How bad is the disadvantage with weak openers.
5) Waht happens to other auctions when you don't open 3C/3D/3H

Ben

The issue that I am trying to hammer home is related to the analytic frameworks that can be used to evaluate these types of methods...

From my perspective, the key issues to understand are

1. Relative frequencies of hand types. I'll ask, yet again, whether you can provide any kind of estimate... I find it astounding that you would devote so much time and effort worrying about complex auction continuations, but you can't provide such a simple piece of information.

2. Whether the methods have a positive expected value ...

3. The relationship between the variance and the frequency of the methods. I'm perfectly happy to adopt high variance methods like "Frelling Two Bids", so long as these methods also have a reasonably high frequency. I need the opening to come up multiple times each session to dampen out the swings...

[awm]

I've run this through my hand generator.

Hands with at least 5-5 in two suits and 4 or fewer losers:

826 out of 100,000 hands.

Of course, these are the first seat odds. Someone has one of these hands at the table almost four times more often than this, but sometimes they face an opening in front of them, etc. This is substantially more frequent than I thought it would be when I first read Ben's posts.

In contrast, how frequent are the weak three openings? It's hard to quantify because this depends a lot on partnership style (how often do you open weak 3 on six cards? are there any loser requirements? suit quality requirements? how does vulnerability effect things?) The simulation I ran was the following:

Weak three bids have between 4 and 10 hcp, exactly 7 cards in the longest suit, and 7 or fewer losers (note xx x xxx KQxxxxx has 7 losers so the loser requirement is not ridiculous). No specific suit quality requirements. No requirements not to have 4 cards in a side major, etc.

Under these constraints, weak three bids came up 1652 times out of 100,000 hands.

Of course, your mileage may vary. If you open a lot of six baggers at the three level then these openings will obviously be more frequent for you. If you drop the 7-card suit requirement completely (keeping the points and losers requirements and allow any 6+ cards) then the weak three bids come up 6401 times out of 100,000 hands.

So it seems the ratio of strong hands to weak (in first seat anyway) is between 1:2 and 1:8 depending on just how aggressively you open at the three level. Ben commented in his post that it was 1:3 or 1:4 which seems reasonable based on a "normal" preempt style.

This doesn't answer the questions though, because we really have to quantify the wins/losses. Ben's posted a lot of examples which support wins on the strong hands, although it would be interesting to know what proportion of the strong hands are relatively flat (i.e. he's given examples where the method wins, but I bet there are a fair number out there where the field is in a flat game or slam). The big question though, is how big are the losses when you open the normal preempt hands. Again this is hard to figure because you can't normally "just look at the hands" without considering the opponents defensive methods, etc. Perhaps Ben has some examples of hands he played (actually using the transfer preempts) in competition with weak hands? Even so, I expect that opening a transfer preempt with a weak hand is actually a win against weak opposition who don't have a good defense! So sample hands from BBO tourneys etc. may not help much here either.

[chicken]

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.

[inquiry]

I think I quantified the hand frequency where the opening bid is made from online bridge using bridgebrowser already in this thread.

Bridgebrowser has a dozen or so databases, some form okbridge, four from open room on BBO, three from tournments on BBO. When I run this program, I haven't figured out how to define look for number of losers, so I set min number of hcp, controls, and two suiter 5-5 or better. This turns up between 1.6 and 2.0% (depending on if you use 15hcp or 16 hcp as minimum and 4 or 5 controls as minumum). But, not all these hand have 4 or less losers. Some have five and occaassionally six. So lets be generrous and estimate that half have two many losers or too poor a major suit to open 3C/3D/3H, you are looking at a frequency of 0.8% to 1.0% of all hands.

Ok, I admit this is an approximate, I ahve found thousands of unique hands, but I have not plowed through all of them to see if thye have four losers or five. This approach includes as a non-match any hand where someone opens before you get a chance. This means at some table, someone might open light in front of this hand so you don't get a chance (shows up as a non-hand), while at others, the person passed in front of the strong hands so it counts. So, let;s consider this a real world frequency. Compare this with the real world frequency of an opening bid of 3C/3D/3H/3S. I did the same test, but instead simply asked for an opening bid of 3 of a suit, limiting opener only to 11 hcp or less (no legnth requirement). The frequency on a large test of over 1 million deals was 2.3%. This number needs no correction per se, as it is what is was. This includes frequency of such bids at all vulnerabilities, and includes reckless preempts and ultra conservative ones. I would think my rate slightly higher, but we will go with the measured rate here.

So the it is 0.9% of all hands strong (maybe slight unbid as 1/2 is a lot to throw out), and 2.3% of hands (may be less than what you would preempt with, but vul versus not, preempt are rare at imps, andmost of those were imp hands). So out of 3.2 such opening bids at this 28% would be strong, 72% weak, or nearly 4 out of 10. I am surprised at this number myself, I think because people don't preempt enough, or I underestimated the frequency of light opening bids or line prior to preempts, or I underestimated rate of vul preempts (indeed frequency of vul preempts turns out be only 1.5% over half a million hands).

So this is the third time I have provide a frequecy of these hands (justin and I discussed them in the other thread), amw has done so a couple of times as well. Experience shows they happen often enough, and EVEN MIORE importantly, they limit the jumpshifts by opener (which happen even more frequently).

Does the method have an expected positive value. This is harder to estimate, I guess. What would you accept as a reasonable trial? Would the first 250 hands that fit an opening bid of 3C/3D/3H/3S meet your requirement? How do we handle competition on these hands? What if opening lead makes a difference? Would PAr result using JACK to figure it out be acceptible to YOU? I have already satisfied myself that the net plus when strong is quite healthy.

As for the downside. It is clear that opening bids of 3C/3D/3H/3S is a winner (bridgebase shows after thousands of such bids, opening side with >50% for each and positive imps score for each (small plus each case). If enough people would adopt Misiry (my new nickname for this based on first three letters of Misho and last three letters of inquiry)... we could run such statistics.

I suspect playing moscito or other strong club system the need to open 3C/3D/3H as transfer preempt strong two suiter is much reduced. You can force quite nicely without it, and your jumpshift rebids after opening, say 1H are already limited by not opening 1C.

[hrothgar]

Here's what I would recommend for evaluating expected gain/loss for the "weak" variant of the transfer preempts.

You need some mechanism to automatically generate large numbers of hands. I've typically done this through scripting.

I use a deal generator to generate hands.
I use a bidding script to automate the auction
I use a double dummy solver to evaluate the final contract

In this case, I'd contrast a standard 3 level preemptive opening with a standard defense as opposed to the transfer preemptive opening and an "optimal" defense...

Generate a 1000 hands and compare results...

In an ideal world, the double dummy solver would be able to take inferences from the auction, but this is way beyond any of the scripting that my friends will do for me...