[hrothgar]

Hi All

As many of you know, I've spent a fair amount of time over the years thinking about methods for providing full disclosure in electronic playing environments. (I still recall debating the issue with Matt Clegg a decade ago). There have been a number of threads on this forum that have touched on related issues. Accordingly, I think that it might be worth while to create a thread where I can describe what I think "ought" to be done. Equally importantly, this will provide a vehicle by which others can critique these suggestions... Before going any further, I should preface matters by indicating that my primary interest is describing methods that I feel are appropriate for "serious" tournaments. I don't necessarily believe that the same structures are necessary for casual play at open tables (who are you going to ask for adjustments). Even so, I believe that it is important that the “serious” tables follow the rules as to provide appropriate examples and guidance the environment as a whole. My beliefs regarding disclosure are shaped by two very simple and (hopefully) uncontroversial assertions. However, most people disagree with the conclusions that I draw:

Assertion 1: Partnerships have an obligation to provide full disclosure regarding their agreements

Assertion 2: Bidding history is the best way to examine partnership agreement

Lets examine each of these in turn.

Implications of Assertion 1: On several occasions in the past, I have suggested that BBO should adopt an electronic convention file that mapped a bidding sequence onto an hypertext string. For example, the sequence 1D – (P) – 1S – (P) – 2H would map onto a string that defined the meaning of the 2H bid. Playing 2/1 the sequence might read “5+ Diamonds and 4+ Hearts, 16+ HCP”. Playing MOSCITO the sequence could read “6+ Hearts, single suited hand” or some such. I have also suggested a number of advantages that would accrue from implementing such a tool. I believe that this “Convention File” system would be an extremely valuable tool for teaching new bidding systems. I also believe that it would make disclosure much quicker and easier. For the purposes of this discussion, I will simple note that Convention File system would also guarantee that alert strings that are included within the Convention File would unambiguously reflect partnership agreement, thereby conforming to the Laws as they are currently written. Many appeals hearings in Face-to-Face bridge hinge on whether an incident reflected a misbid or misinformation. Appeals committees typically rely on written documents such as System notes or Convention Cards to reach their verdict. The methods that I suggest (essentially) short-circuit this process. The Convention Card is now the one providing explanations. I’ll note in passing that the system that I recommend is not a panacea. There will always be bids that fall outside the Convention File. More importantly, their will often be disconnects between what players think that the system means and what the system actually provides for. Case in point: Lets return to player who opened 1D and then reversed into Hearts holding a 2443 13 count… The alert string provided by the Convention File would bear no relation to the actual hand held by the bidder. In turn, this could damage the non-offending side. All that I can offer in this case is two basic observations: First – If we presume that the convention file is accurate, this system permits us to make an unambiguous ruling based on a misbid rather than misinformation. Second, as I noted earlier, I believe that the Convention File system would be extremely useful as a teaching tool. Ultimately, the way to avoid these types of issues is through better education.

Implications of Assertion 2: Assertion 2 states that bidding history is the best mechanism to judge partnership agreement. One of the trickiest issues for players is accurately defining the meaning of different bids. Lets consider a simple example such as the strength shown by a 1NT opening. What happens when different players use different “metrics” to decide whether or not a hand is strong enough for a 1NT opening bid? Walter the Walrus lives and dies by HCP, the Hog uses a much more fluid system to evaluate hand strength, and the Secretary Bird insists that these two need to provide some mechanism to communicate. I argue that the “best” way for these two to communicate is for the Hog to provide the Walrus with all of the hands where he has opened 1NT in this partnership. The Walrus then has the option to collapse this set of hands into any one of a variety of summary statistics describing hand strength. For example, if the Walrus is ONLY comfortable with HCPs, he would be given a Probability Density Function describing the distribution of HCP. In the case that the Hog can’t provide a suitable bidding history, he should use a Hand Generator to create 100 odd hands consistent with the bidding so-far.

Please note: I recognize that what I am suggesting here is wildly impractical. Given the “will” convention files could be implemented today. I’d even argue that they should be implemented today. I don’t necessarily believe that every player wants to be configuring his own expert system to analyze the opponent’s bidding history. With this said and done, it does make for an interesting thought experiment…

[Blofeld]

Lovely idea.

[pigpenz]

I had an odd thing happen against a pair in an acbl game....they were playing new minor forcing over nt rebid but they werent reallly playing it. After the hand I asked few questions as they never alerted the rebids afterwards. The hand that opened
and rebid has three card support but only bid two of the major and the one using new minor forcing had only nine hcp. When asked they said they were new players and didnt have sophisticated meanings to the rebids. Now They both had "J" by their names which means they have won at least 100 bbo points.

So looking at their hand records from the past that I found that they used it quite regulary and it seemed to work in their favor quite a bid for stealing hands cause noone wanted t balance agaisnt this. Now to me this isnt full disclosure.

[1eyedjack]

I like the idea of a system file with hyperlinks, accessible to opponents. It would certainly be an order of magnitude better than any method currently in use (and the existing BBO CC design could use a LOT of improvement). Perhaps the coding is technically simple, but I foresee a few possible problems, perhaps not insurmountable but requiring care at the planning stage:

Despite the opportunities afforded by technology, examples may arise where there has to be a trade-off between ( a ) the completeness of disclosure within the file and ( b ) ease of access of the information sought. Where the file is being used as a reference tool after a potential infraction, typically after a hand is concluded, it is a no brainer: the complete disclosure is the only way to go. Not only do you have additional time in which to seek the information but you can also by that time use the assistance of the "offending" pair, who are familiar with the structure of the file, to guide you through the notes without a risk of UI.Where the file is being used live by opponents during the progress of a hand rather more care is required if they are to succeed in distinguishing the wood from the trees.

Convention cards as they currently stand have the advantage of familiarity. If you want to know the meaning of a jump overcall you know precisely where to look within the opponents' CC, no matter how complex their system is otherwise.

The hyperlinking opportunities undoubtedly cut down the problem. Intelligent rules imposed upon the author of the CC regarding the structure of the file will also cut down the problem. Neither will eliminate it, although it would not take much to arrive at an improvement on current methods.

Take a couple of fairly trivial examples:

Example 1

You choose to agree that a 1NT overcall is 15-18 in direct seat after a 1-suit opener by the opponents. Do you separately define (1C)-1N, (1D)-1N, (1H)-1N and (1S)-1N? Or do you permit a structure that allows for (1-suit)-1N to be described in a single expression? If you choose the latter, what if another pair vary the definition according to vulnerability? With a manual scribble on a convention card there is scope for the author of the CC to disclose in the most efficient manner that his intellect devises. The forced structure of the CC, combined with forced limitations on the total amount that you can disclose therein, affect what is disclosed within the CC and where it is disclosed, but in such a manner that an opponent has a fair chance of finding it if it is relevant.

Example 2

With my regular partner I have a fairly complex (as they go) flow diagram that determines whether a double is take-out or penalty orientated, as well as including examples where a double is a strictly artificial convention. The flow diagram is quite efficiently designed and compact, but only because in the "white space" on a CC I have autonomy over that design. Granting an author of an unlimited-sized CC carte blanche over creating such structures within the CC would be quite neat, and doubtless helpful in an appeals hearing, but of dubious benefit to an opponent in flight.

I guess it would also be quite nice for the CC to be linked to the bidding history so that where a relevant sequence develops it automatically ties the explanation to the bid within the bidding history, but I can predict that that would be a decade or more away yet. A rudimentary auto-alert facility (perhaps with manual override) may be a bit closer to immediate implementation.

If it were possible to create a statistical analysis of hands played, it would be nice if that were automatically cross-checked against the system file so that you could see, for each defined sequence
( a ) the number of times that the sequence has been used where the hand is consistent with the description in the system
( b ) the number of times that the sequence has been used where the hand is marginally outside the description of the system
( c ) the number of times that the sequence is completely at odds with the description

Item ( c ) should ideally be subdivided into psyches and misbids, but that would require manual input at the discretion of the user so would not be possible in practice (even assuming that any of the rest is possible). Item ( b ) would be useful to identify implicit agreements, ie those who habitually open 1NT with 11 points having described it as 12-14 (even if responder never fields).

Personally I think that the statistical element is a flight of fancy, and probably should be kept to a separate thread to avoid cluttering that which is immediately feasible.

[keylime]

Richard,

Ditto, ditto, ditto, and ditto.

Pigpenz,

Send a group of hands like that into Abuse. It's clear they have a clear implicit agreement to psychicly bid NMF, and that breaks about 3-4 different guidelines.

[ArcLight]

In addition to conventions, there is also the shared unannounced information from negative inferences.
Say a partnership has an agreement that a double of an enemy splinter bid calls for leading the lowest unbid suit. The opponents make a splinter bid, and pard does not make that double. You now have a negative inference to steer you away from that particular lead.
The declarer should be entitled to this information as well, shouldn't he?

Do you alert the opponents "By the way, in this situation we play that a double of your splinter calls for leading Diamonds"?


How about this idea:

Someone writes a "convention checker" program, that analyzes a set of games to measure the consistency of a conventions use. For New Minor Forcing it could measue average strength, length, frequency, etc. Then you could run it against certain pairs. I'm not saying it would be all that easy to do this, as a lot of borderline cases would be involved. But it might be useful.

--------------------------------------------------------------------------------

I was playing against a pair with a convention card. One of them false carded and false lead every hand, sometimes multiple times per hand. Ex: They had Lead K from KQ on the card, yet this one person would always lead the Q. I asked their partner "Does you partner false card a lot, (pointing out the lead)." I also asked about this persons frequent psych bids, and playing the higher card from a touching sequence instead of the lower, when following suit. I was told no, he doesnt make psych bids, and he can card in what ever way he likes.

A player may make what ever non standard carding plays , leads, bids they like, but if their partner is aware of this, they have an obligation to announce it. And more importantly, they shouldn't have secret signals, which in effect the did have.

(I left the table after a while, and mafrked them as enemy)

[hrothgar]

I agree with much of what 1eyedjack has to say... In particular, he very properly identifies that enormous care needs to be spent designing such a system. There are a lot of trade off's between simplicy/ease of use and flexibity.

I've actually discussed this topic in depth with a couple friends who are professional programmers. Both of them independently reached the same basic conclussion... Ideally, they believe that this system should be designed the same way many programming systems. At the top level you have a programming environment that is designed for ease of use. The output from this programming environment gets compiled into nice compact machine language

I also agree that any such system should be viewed as a supplement to traditional convention cards rather than a replace for convention cards... I've always felt that an "otpimal" disclosure system would include 4 key components:

1. A short/easy to understand summary of the system

2a. An announcement describing the meaning of specific bids
2b. Suggested defenses to particular complcated bids

3. A convention card optimized to describe negative inferences (bids available but not chosen)

[hotShot]

A CC is one thing, a compleat system description is something else.

I can look at a CC to see if "may contain 5-card Major" is checked, but i can't read a full system description while playing.

Creating something like that makes only sense if you plan a long term partnership.
With a pick-up partner you might have only agreed on 5-card Majors and aome NT range.

But i agree, an electronic playing environment should supply means to define a full system description or a good CC. Opponets should be provided to look at some sort of partnership history of opponents, to see if a sequence like the one that just happend, had been played before.

[hrothgar]

hotShot, on Jul 20 2005, 09:49 PM, said:

>With a pick-up partner you might have only agreed on
>5-card Majors and some NT range.

To some extent this reflects a decision regarding which convention card to load rather than the mechanism by which disclosure should be achieved

>Creating something like that makes only sense if you plan a long term partnership.

In an ideal world, players might agree to play a detailed standard system like BBO Basic... Here, once agin, I consider the options that a convention file system would provide for learning new systems to be particularly useful.

[Echognome]

One thing that may be nice is for a system to be "defined" through play and alerting. Suppose you sit down opposite a partner and choose to play SAYC, 2/1, MOSCITO, Acol, Precision, WJ2005, etc, or other. The program may have a set of default bids and bidding sequences already defined. You can also have some options within these systems to check, for example, Raptor NT, Defenses to weak and strong NT, different blackwoods, etc. Now, when you play the first few times that a bid comes up it offers you a meaning and you can either click to confirm that meaning or type in your own alternative meaning. These agreements are then saved for your partnership in a file on both you and your p's computer. Everytime you play these files are updated. It might also give you a clickbox to select "psych" so that the regular meanings are shown, but it knows that you haven't now defined partnership agreement.

Now, the problem with this system is that it will remind you of your own agreements while playing in a serious event. So, perhaps during a serious tournament there is a different system in place that utilises the information in your file. It will automatically alert the opponents what your bid means, but you (or your partner) can also alert. In order to avoid too much information, perhaps the computer can be trained to only alert bids that need alerting.

Ideally this becomes somewhat like programming artificial intelligence. The more you play, the more accurate your alerting becomes. Perhaps you can choose whether to play in learning/teaching mode or "just for fun".

[luis]

When I saw the title of the thread I was scared thinking "oh no, not again" but I'm glad I clicked since I found your ideas refreshing and can be the start of several good things.

For example about disclosure it would be nice to have a "history agent" available so when you play against a pair you can get some information about at least their openings, for example:
They open 2h and you can at the given position/vulnerability the average # of HCP they had, the average number of hearts they have, if they had a 4 card major before, side suit etc.
For example:

2h opening (Mr Foo and Dr Evil)
NV vs VUL (2nd position)
Average HCP: 6.66 (lowest: 2HCP, highest: 8HCP)
Average # of cards: 5.50 (5 cards 23 times, 4 cards 3 times, 6 cards 31 times, 7 1 time)
Side 4 card major? (Yes 2 times)
Can have a void? (No occurrences registered)

1NT opening (Mr Foo and Dr Evil)
NV vs NV (1st position)
Average HCP: 15.54 (lowest: 14 highest: 16)
HCP per frequency (14: 3 times, 15: 17 times, 16: 2 times)
Can have a Singleton (2 occurences, x of spades and K of clubs)
Unusual distributions: (3622 (1 time), 3262 (1 time), 1345 (1 time), 3451 (1 time)

For a 15-17 1NT you can see how many times it was with 14 (13 less etc) how many times with a singleton, how many times with 6322 or 5422 etc.
We can then work what stats will be displayed on request about openings and we'll have good disclosure about opening style.

Maybe I'm just nuts :-)

Luis

[Echognome]

I think Luis' idea is nice, however I can imagine that the data issues involved with that are quite intensive. That would effectively mean that one would have to keep data on every hand played by every player. Although this is currently kept in myhands, I can imagine it would be quite slow to access the data and perform some routine statistics on them.... unless... one defined a few useful statistics to be kept on a person's computer everytime they played. For example, if we thought opening 1NT and 2-bids were the most common ones to want data on, then that could be done. However, even with 2-bids, how are you going to account for my current 2-bid structure with my most regular partner:

2♣ = 44 or better majors
2♦ = multi
2♥ = hearts and a minor
2♠ = spades and a minor

The statistics would certainly make little sense for these in terms of the suggested statistics (with the exception of perhaps hcp).

[coyot]

You would need to keep the statistics individually for each player - and when playing against a pair, view their individual statistics and go nuts trying to figure out what will happen :-).

[pigpenz]

luis, on Jul 20 2005, 03:02 PM, said:

2h opening (Mr Foo and Dr Evil)
NV vs VUL (2nd position)
Average HCP: 6.66 (lowest: 2HCP, highest: 8HCP)

Hey Luis I resent the DRevil implications:
Full disclosure is full disclosure......look at the question posted about what type of hands would you opene weak twos on......weakt two are pretty common prolly i guess about 98% of players play them.

But what is not disclosed is how your partners tendencies are. if you know you partner opends most of the time on jxxxxx your opps really are entitled to that info, but i have never seen anyone say we open undisciplined two bids

[luis]

pigpenz, on Jul 20 2005, 09:05 PM, said:

luis, on Jul 20 2005, 03:02 PM, said:

2h opening (Mr Foo and Dr Evil)
NV vs VUL (2nd position)
Average HCP: 6.66 (lowest: 2HCP, highest: 8HCP)

Hey Luis I resent the DRevil implications:
Full disclosure is full disclosure......look at the question posted about what type of hands would you opene weak twos on......weakt two are pretty common prolly i guess about 98% of players play them.

But what is not disclosed is how your partners tendencies are. if you know you partner opends most of the time on jxxxxx your opps really are entitled to that info, but i have never seen anyone say we open undisciplined two bids

When I'm asked I usually say "can be dramatic". I think that's a good description of some of my weak 2s

[bendare]

Pigpenz--

You struck a chord with me, Pig! It seems some of the players in ACBL tourneys post CCs such as 2/1, nmf, 4sf, bergen, rev drury, etc. when they either have no clue what ANY of it means or if they do they have UI to know when their bids are for real and when not.

An example of one pair, who has such a CC posted to include Reverse Drury-----

p p p 1s p p double 2s p p p.

Opener, in 4th seat, had 10 HCP with 5 bad spades and rebid those bad spades after her pd had passed originally and THEN passed her 1 spade. Her pd, with 4 spades and 11 HCP passed not once, but twice. (no misclick could be claimed.) As per your NMF example, this particular pair frequently does not alert nor do they have their NMF bids (or most other systemic bids as their cc states.) Yet, when their NMF bid is real, I have seen them alert it when they DO HAVE their bid. Go figure!!

It is maddening, and yes, it does affect opponents decisions, not only in balancing, but on defense or declarer play should they win the auction. The results from these improprieties affect not only their opponents but the entire field. (The particular pair I reference have a J and a Q on their profile, meaning they have played and placed enough in online ACBL games that they should know and be held responsible for abiding the rules and laws of ACBL.)



Keylime--

So---------- you send a group of such hands (and much worse than Pigpenz or I have given as examples) to Abuse, and you send another group and another group, and THEN WHAT??? The pair I referred to is still playing regularly in the ACBL tournaments. On one occasion I just asked Abuse to look at an entire 12 board tourney where 11 of this pairs 12 boards were clearly more than bizarre in bids or plays.


I would certainly like to see the "laws" of bridge, full disclosure, rule of coincidence, etc. enforced in online tourneys and particularly in ACBL tourneys, where Zero Tolerance is announced at the beginning of each tourney. Also the directors announce that a CC MUST be posted or players will be expected to PLAY the default SAYC CC posted for them. That should mean, as in FTF ACBL games, that players MUST play per their own posted CC. It seems the ZERO TOLERANCE only refers to "rudeness".


Online bridge is a horse of a different color and has its own unique problems. I understand it is virtually impossible to enforce the "laws of bridge" in online bridge, but I do think the most blatant abuses can and should be dealt with appropriately by BBO OR in the case of ACBL games by ACBL also. If participants in ACBL online tourneys do not know or understand the systems they claim to play on their posted cc, they must be made aware that is not permissable and that it will not be permitted. Personally, I think every "suspected abuse" reported to BBO abuse (that occurred in an ACBL game) should be forwarded and reviewed by ACBL for appropriate action BY ACBL according to ACBL's rules and regulations. If there is any perception by ACBL that it is not in their best financial interest to "offend" or "alienate" frequent online "paying customers," I beg to differ. I feel strongly ACBL online games would grow leaps and bounds if ACBL did their best to keep their online games up to the standards most of us have known, abided and respected in ACBL FTF games.

When such improprieties are repeated again and again by the same players and are allowed to continue , should we just throw up our hands. give up on hope of a fair and square game online, forget it and stop wasting our time discussing cheating, ethics, rules, laws, regulations, etc in these forums??



In general-----

Bridge, and online bridge, is and should be for everyone who enjoys the game regardless their skill level. Not all aspire to "improve" their bridge but rather just want to enjoy the "kitchen table" game as they know it. There is NOTHING wrong with that. In fact, it is a wonderful outlet, especially for elderly social bridge type players. BUT----- players should be aware that when they play in certain competitions such as ACBL games they will be expected and must play by the "rules" laid out by that body.

We ALL want to encourage beginners and certainly some leniency is given beginner players in FTF ACBL games --as it should be here---just not indefinitely. I think that was Pigpenz's point in noting these players had Js on their profile. Obviously, from the example he gave, these are not "good" players, but they also obviously have played enough and "placed" frequently enough to have won these Js and Qs on their profile and it is high time they know and observe the rules and laws of the games in which they participate.


BBO is an awesome site and I appreciate the privilege of playing here. However, it IS frustrating when playing in what I presume should be a reasonably "serious" game but instead seems more "anything goes." Indeed these are NOT World Championship games. I have frequently heard the ACBL tourneys here compared to "club games." I would agree with that comparison regarding variety of skill levels, etc. but the clubs I have played in ARE serious about enforcing the rules, laws, and ethics as laid out by ACBL. Club games are generally very friendly, but not friendly to an anything goes attitude.

Now that I have sounded off here, perhaps I should have instead originated a new subject called "SOUND OFF!"

Bendare

[inquiry]

Echognome, on Jul 20 2005, 04:35 PM, said:

That would effectively mean that one would have to keep data on every hand played by every player.

There is a program/database that already does this. For example, your first hand on BBO as echognome occurred on March 11, 2004. The hand you played was this, you were South playing with panzade against JRShip and Miss E. You played this hand at 9:09 PM (probably Las Vegas Time, but I am not sure). If it matters, you won 1.80 imps.

Scoring: IMP

West North East South

 Pass  1♦    7NT   Dbl
 RDbl  Pass  Pass  Pass
 

---

H6 H3 HJ H4
DA D2 D9 DJ
DK D4 C3 D3
DQ SA C5 D5
DT HQ S2 D7
D8 CK H5


During March and April of 2004 you (not your partner) opened 1NT 16 times on BBO main room, your hcp range was from 11 to 17. You opened

11 - twice
12 - twice
14 - four times
15 - four times
16 - twice
17 - twice

Once you had a five card heart suit. You frequently were 4-4 in majors.

[Echognome]

Right. I knew the data was available... I didn't know how quickly it was accessed though. I knew from myhands I could bring up the data, but it usually takes around a minute to process which would be too slow for practical bidding.

As per my NT ranges you can see how important it is to associate that with partners as with some I play a weak NT, with others a strong NT and with my regular partner a variable NT (including 9-14 in 3rd NV).

[kenberg]

Hats off to you for your effort in this. I will try to say something useful.

There are always two related but separate issues:

Working out the optimal rules

Providing an environment where these rules are understood and followed.

Far more often, I see posts and hear comments about the second of these issues. Most of us (I think) accept that partnerships will always know more about each other's tendencies than can ever be summarized neatly, even with the best of intentions. Example: The other day my partner opened 1C, the opponents boought the auction, and early on it became clear that my partner had exactly five cards in the majors. I would have bet my house that this partner had at least five clubs and I defended the hand on that assumption. I have had partners who might open 4-4 in the minors with 1C, but this partner never. I don't think this aspect of bridge has a technological (or any) answer.

On the other hand, pairs vary greatly about their willingness to disclose unusual agreements. Quite often this is due to not knowing what constitutes an unusual agreement. Example: 1C-1D-1NT. At least one of the opponents might have a four card major and of course maybe they both do (it happens). If opener might skip over a weak major to rebid NT when holding a flat hand, that's fine. If he regularly rebids 1NT when holding 4-4-2-3 knowing he is safe in doing so, that's a non-standard agreement (by my understandings) and we, his opponents, need to hear about it. These issues arise in f2f bridge and cause consternatiuon, the same happens here.

I wish you luck. From my viewpoint the most important thing by far is a clear presentation of what is expected.

Ken

[epeeist]

Interesting discussion. I have three questions (well, more, but I'm only asking three) and one comment.

First, in respect of the discussions regarding data, what of people (like me) who tend to play with a wide variety of partners, whether in pairs or individuals, using different systems and conventions?

Second, what about "evolution" of partnership bidding agreements? E.g. maybe a pair starts out using SAYC, then 2/1, then tries precision, then...(etc.). Or even adding/changing conventions would have an effect.

Third, what about data on carding?

As for the comment, a previous poster noted the problem of people not knowing the conventions they claim to use. I think the problem was not with not knowing the conventions, but with not alerting/explaining properly. If one makes e.g. a Michaels cuebid, the correct explanation is NOT "Michaels" but what one's partner will understand it as meaning in that situation (e.g. if nonvulnerable over a minor suit, IF this is what partners have agreed, "5-11 HCP, at least 5-4 in majors" or whatever). If partners don't know how to use a convention, BUT explain what they think it is properly, then opponents get alerted to the partnership understanding of what that bid means -- even if it has no name known to bridgedom.