[rbforster]

I've always been interested in applications of encryption to bridge and was recently reminded of this when I ran into Peter Winkler at the DC Nationals (who first introduced these approaches to bridge - see some articles here). He's got a new book coming out, Crypto-Bridge, and was even kind enough to give me a copy .

Anyway, this got me thinking about some of the opportunities for encryption in bridge, and I thought I'd offer a few and hopefully others could add their suggestions. Here are a couple examples in the bidding:

1. In an ace-or-king cue bidding situation (not cueing shortness), once your side cues a suit twice you encrypt you later actions based on who holds the A vs the K.

2. encrypted king ask, like specific kings after 4N blackwood/keycard followed by 5N asking for kings (and confirming all keycards), you vary how you show your kings based on who's got the A♠.

3. pairs of special raises, like Jacoby or Drury, where one raise promises 1 of the top 2 trumps. Then you follow this up with rotated game tries or cue bids if partner confirms holding the other top honor.

4. special doubles, like a Rosencrantz redouble (1m-1M-X-XX), showing Ax or Kx in partner's major, which is later used to encrypt good/bad actions in competition

In general, you first need to establish a key, such as which of the top trump honors you hold, and once the key is established, you can then use it to encrypt your subsequent bidding so the opponents will have less of an idea what to lead. In some examples, like Blackwood-then-5N, you know for sure that a key is established; in other cases, such as when one person shows a top honor in support, their partner will often but not always have the other (since they bid the suit and showed some strength), which leads to a likely but not sure key.

There are even more tricky examples, a few of which are mentioned in the new book. Recalling that generally an encrypted situation arises where your partner can tell based on his hand what the meaning of your bid is, but the opponents can't tell which of several possible meanings are shown since you aren't forced to disclose the contents of your hand. An example of this is a double based on length or shortness:

1M-P-1N-P
3M-P-P-X

where X is either penalty (lots of length in their suit) or takeout (shortness). Partner is expected to look at his hand and figure out which, but the opponents (especially the 3M bidder) can't tell at the time of his bid. You could even extend this to a much more common situation like

1m-X where 1m showed 2+, and double shows 5+ or 2- in that suit.

While applications of encryption to bidding are pretty rare and specialized, there are lots of applications of encryption to defensive carding. Unfortunately, unlike encrypted bidding (which is generally allowed), encrypted carding is largely not allowed. Examples of encryped carding include:

1. against 3N, lead 3rd/5th from a good hand but 4th from a weak one. Figure the opps have 25 points, and your hand is "good" if you've got 8+ (more than half of the at most 15 remaining for your side). Declarer will often not be able to read the lead until much later in the hand, and your partner will be able to read it immediately (at in cases where the contract is in danger). Since many hold-up plays and blocking plays require guessing the correct layout to play for at T1 or T2, declarer may be at a disadvantage until too late.

2. leading standard or upside-down attitude against a suit contract based on trump parity, when declarer has shown a known trump length (like a stayman auction or a weak two opening). Upon seeing dummy, partner knows how many trump you have, but declarer can't figure it out until he draws trump.

3. when declarer ruffs in hand, you and your partner know who holds the smallest card remaining in that suit. all further signals, like attitude and count, are standard or UDCA based on who holds this smallest card (and you don't discard it unless you have to).

If you can think of other ways to apply encryption to bridge, feel free to share them. I know some of these methods are getting forgotten since the regulators have banned them for years now, but hopefully some of the senior member can share examples from back when these were still being played in high level competitions.

[Echognome]

I always thought encryption was disallowed as signals, but not in the bidding, although it may vary by regulating authority or I may be outdated.

[rbforster]

Echognome, on Aug 18 2009, 12:15 AM, said:

I always thought encryption was disallowed as signals, but not in the bidding, although it may vary by regulating authority or I may be outdated.

That's correct. Bidding conventions aimed at setting up encryption aren't restricted any more than any other conventional raises or artificial calls, at least in the ACBL.

[awm]

It seems like there might be some interesting encryption opportunities in relay bidding, and they might actually allow you to compress information more efficiently too.

For example, suppose instead of showing specific suits in response to a strong club, I show "a five card suit headed by the ace only, and a lower four-card suit headed by the king and queen." Odds are fairly good that partner can figure out which two suits I have based on his honor holding (and if he can't, our chances of slam are dim anyway). You'd need to use some relay breaks to say "I can't figure out your suits, please clarify" but on those auctions where you don't need the breaks you've probably saved space as well as made things tough for the opponents on opening lead.

[Free]

awm, on Aug 18 2009, 07:10 AM, said:

It seems like there might be some interesting encryption opportunities in relay bidding, and they might actually allow you to compress information more efficiently too.

For example, suppose instead of showing specific suits in response to a strong club, I show "a five card suit headed by the ace only, and a lower four-card suit headed by the king and queen." Odds are fairly good that partner can figure out which two suits I have based on his honor holding (and if he can't, our chances of slam are dim anyway). You'd need to use some relay breaks to say "I can't figure out your suits, please clarify" but on those auctions where you don't need the breaks you've probably saved space as well as made things tough for the opponents on opening lead.

Your setup doesn't work. You have 4 suits to distinguish. However, using tophonours gives you a lot more combinations (8 to be exactly). You could combine them (A or KQ, nothing or AKQ) but that's an encrypted encryption. Even using only A and K gives you 4 possibilities. And note that the key "C/D/H/S" is unique, while the key "-/A/K/AK" is not (I can have 2 suits headed by AK, but I can't have 2 ♣ suits). So compressing information is impossible.

[rbforster]

In relay bidding, denial cue bids sometime group honor holdings like "AK or neither" or "AKQ or none", on the assumption that the strong hand can almost always tell which. You can then encrypt later responses to honor location based on if it was all or none. Combining this with keycard or similar alternatives to denial cuebidding for when opener has a wide open suit, and you could be pretty sure that your side would know what was being shown.

Adam's example of first responses to a strong club could still have potential. Even if there are more possibilities, if you have responder show a likely one first (and opener can read it) you're off to a great start and can encrypt responder's side suit, etc. If opener can't read it, you've wasted a little space and the remainder of that auction will be less efficient. For example,

1♣-1♥ shows a 5+ major with 1 major ace
1♣-1♠ shows a 5+ major with 0/2 major aces

After this, the cheapest step would confirm the encryption and relay for shape, while 2♣ could ask responder to transfer you to his major (and continue with an unencrypted auction 1-2 steps higher).

Normally you show the major and then maybe which honors much later (main suit, side suit, exact shape, honors being a typical order). But if you show the honors earlier in the sequence you can sometimes encrypt the rest. In addition, right-siding is less of a concern during relays if you're not going to give away responder's hand during the bidding on the times where he'll end up being declarer.

Another couple examples of encrypted bidding:

1. disciplined weak 2 bid (2/3 top), then 2N asks for an encrypted feature based on the missing honor (which responder will usually have).

2. 3N showing a solid minor suit, and then control or shortness asking bids are encrypted based on which minor (on the assumption the asker will have an A/K/Q in the other minor to know). Of course some of the defenders may also be able to figure this out if they also hold an honor in the other minor, but they might not be the one on lead.

[hrothgar]

One minor suggestion: For better or worse, labelling a method as "encrypted bidding" invites the regulatory authorities to take a nice big ***** on your chest. A lot of the topics that are being discussed have very little to do with encryption per see and an awful lot to do with compression.

There isn't much established nomenclature in this area. If I were trying to kick off a discussion of this type of topic I know what vocabulary I'd choose.

[RichMor]

I think remember an example from 'Super' precision where a 3NT response to an opening 1♣ showed some solid 7-card suit. Opener could usually figure it out or could relay with 4♣ to ask.

RichM

[CarlRitner]

So where does Keycard Blackwood responses fall in here?

5C = 1 or 4 key cards (when playing 1430) where responder knows and the asker can supposedly figure it out. We assume the opponents might or might not know the actual answer.

That's compression and not encryption, right?

[rbforster]

Carl, on Aug 18 2009, 10:54 AM, said:

So where does Keycard Blackwood responses fall in here?

5C = 1 or 4 key cards (when playing 1430) where responder knows and the asker can supposedly figure it out. We assume the opponents might or might not know the actual answer.

That's compression and not encryption, right?

Asking a question where you can likely guess the answer among two possibilities is compression. However, it sets the ground for encryption after that since you can use as a key which possibility is shown. This was in my example about encrypted king asks where you start with RKC, confirm all the keycards, and then use who holds a particular ace to encrypt the answer the kings question.

[aguahombre]

A question based on my ignorance: When, after the auction is over, you are explaining it to the opponents, do you say "6D showed the club king"? Do you say, "6D showed the club King if I have the Spade Ace, and the Diamond King if I don't."?

[akhare]

aguahombre, on Aug 18 2009, 12:00 PM, said:

A question based on my ignorance:  When, after the auction is over, you are explaining it to the opponents, do you say "6D showed the club king"?  Do you say, "6D showed the club King if I have the Spade Ace, and the Diamond King if I don't."?

It seems that the explanation should be based on apriori facts, i.e., "6D showed a minor suit king". You may be privy to which specific king responder holds based on your own hand.

IMO this is no different from infering in an either / or DCB auction ("either AKQ of X or nothing"), where you can resolve the condition based on your own holding, but one or more opps may not be able to tell.

[aguahombre]

akhare, on Aug 18 2009, 12:13 PM, said:

aguahombre, on Aug 18 2009, 12:00 PM, said:

A question based on my ignorance:  When, after the auction is over, you are explaining it to the opponents, do you say "6D showed the club king"?  Do you say, "6D showed the club King if I have the Spade Ace, and the Diamond King if I don't."?

It seems that the explanation should be based on apriori facts, i.e., "6D showed a minor suit king". You may be privy to which specific king responder holds based on your own hand.

IMO this is no different from infering in an either / or DCB auction ("either AKQ of X or nothing"), where you can resolve the condition based on your own holding, but one or more opps may not be able to tell.

Ok, that was the reason for the question. Disclosing information known to the bidding side, but not to the other side doesn't need to occur. But we are free to guess which one of the two has a particular Ace if we want, and deduce therefore what King was shown.

[CarlRitner]

Bridgeguys, a rather comprehensive bridge site, says encrypted calls and signals are banned by the ACBL, but a search of the ACBL site finds several references to signals only.

Are we pretty confident that encrypted bids/calls are approved, or at least not banned?

[Cascade]

akhare, on Aug 19 2009, 05:13 AM, said:

aguahombre, on Aug 18 2009, 12:00 PM, said:

A question based on my ignorance:  When, after the auction is over, you are explaining it to the opponents, do you say "6D showed the club king"?  Do you say, "6D showed the club King if I have the Spade Ace, and the Diamond King if I don't."?

It seems that the explanation should be based on apriori facts, i.e., "6D showed a minor suit king". You may be privy to which specific king responder holds based on your own hand.

IMO this is no different from infering in an either / or DCB auction ("either AKQ of X or nothing"), where you can resolve the condition based on your own holding, but one or more opps may not be able to tell.

No it shows something more specific than that.

It shows both ♦K and ♠A or ♣K and no ♠A you need to tell the opponents that.

I mean it is not a guess for them. They or one of the opponents may be able to figure out this key for example someone might have the ♠K and therefore know that a cue-bid was the ace. Or the shower might become dummy and then the defense can figure out what is happening by whether or not the ace shows in dummy.

[rbforster]

Carl, on Aug 18 2009, 12:47 PM, said:

Bridgeguys, a rather comprehensive bridge site, says encrypted calls and signals are banned by the ACBL, but a search of the ACBL site finds several references to signals only.

Are we pretty confident that encrypted bids/calls are approved, or at least not banned?

Yes. One of our regulars here asked the ACBL specifically about this and was surprised when they said that encrypted bids where allowed under GCC (subject to tempo and other conventional restrictions). There is nothing about encrypted bidding specifically being disallowed, and so long as you can make the conventional bid in question legally - such as a jump raise showing support (ala Bergen), a game forcing artificial response, or any constructive call starting with opener's first rebid - you're welcome to give it a try.

[rbforster]

aguahombre, on Aug 18 2009, 12:00 PM, said:

A question based on my ignorance: When, after the auction is over, you are explaining it to the opponents, do you say "6D showed the club king"? Do you say, "6D showed the club King if I have the Spade Ace, and the Diamond King if I don't."?

I believe the latter is the correct full explanation. They aren't entitled to know whether or not you have the Spade Ace, just like they don't get to know whether a RKC sequence showed 1 or 4 (even if asker can tell) or if an A-or-K cuebid was the ace or the king.

[nigel_k]

Rob F, on Aug 18 2009, 01:22 PM, said:

One of our regulars here asked the ACBL specifically about this and was surprised when they said that encrypted bids where allowed under GCC (subject to tempo and other conventional restrictions).

If my opponents wanted to do this, I would really hope they were allowed to. It sounds like a recipe for frequent disaster with a miniscule upside.

[FrancesHinden]

While I like the theory, I am a tad pessimistic about the frequency of application.

The problem with the encrypted bidding ideas is that the potential for using the agreements is so rare. To take one of your examples "once we have cue bid twice in a suit where we wouldn't cue shortage".... to be honest, once I've got to that point in the auction it's usually going to be over pretty soon afterwards.

Your example 2 (specific king ask): this works (in the very rare case that it comes up) - but what is the benefit of encrypting the replies?

I would have thought this type of thing would come up more often in relay-type methods where you have a lot of artificiality at lower levels. But are you then getting a significant benefit from the encryption?

Quote

1. against 3N, lead 3rd/5th from a good hand but 4th from a weak one.  Figure the opps have 25 points, and your hand is "good" if you've got 8+ (more than half of the at most 15 remaining for your side).  Declarer will often not be able to read the lead until much later in the hand, and your partner will be able to read it immediately (at in cases where the contract is in danger).  Since many hold-up plays and blocking plays require guessing the correct layout to play for at T1 or T2, declarer may be at a disadvantage until too late.

2. leading standard or upside-down attitude against a suit contract based on trump parity, when declarer has shown a known trump length (like a stayman auction or a weak two opening).  Upon seeing dummy, partner knows how many trump you have, but declarer can't figure it out until he draws trump. 

Frances's special rule of encryption: never play a signalling method that depends on believing the opponents' auction.

Quote

3. when declarer ruffs in hand, you and your partner know who holds the smallest card remaining in that suit.  all further signals, like attitude and count, are standard or UDCA based on who holds this smallest card (and you don't discard it unless you have to). 

This is the one I like. It works any time you know declarer is out of a suit (he doesn't have to 'ruff in hand').

Thinking of a different thread, playing this is also good practice for keeping track of all the pips in a suit...

Pity it's illegal.

[aguahombre]

agree with Nigel, and hope people start using encrypted bids. great way of getting around disclosure in the few cases where opening lead is critical, and finding out which king is missing after seeing dummy is too late. And when it goes into use, I will have the following auction:

1NT* alerted as 15-17, unless our last board at the previous table was a N/S hand, in which case it shows 12-14. The opponents have not yet played that board, so they do not get to know. they can guess, based on their holdings, so everything is OK.