[aguahombre]

Actually, my contention hasn't changed. For the person not signalling, because his partner doesn't need the information, it isn't a signal at all.

[pran]

 Trinidad, on 2013-February-10, 22:28, said:

No, I don't. But then again, I said that I think encrypted signals should be allowed.

My question was not whether this should be allowed (for me that is a clear yes). The question is whether it is an encrypted signal (to me that is also a clear yes).

You seem to define an encrypted signal as a signal that isn't allowed. And since the above situation should be allowed, this cannot be an encrypted signal. The usual order is to define things first and then say whether they are allowed.

I asked this question because David claims that this isn't really an encrypted signal because the defenders don't know more than declarer about the key (the ace). So I ask him for his opinion if the defenders do know the key and declarer doesn't, by giving two situations:
- one where the defenders know explicitly that declarer can't have the ace.
- one where the defenders know by inference that declarer can't have the ace. (If he has the ace, it doesn't matter what they signal: the contract can't be broken.)

To me, those are encrypted signals. I don't have a problem with those, since I would allow encrypted signals anyway. But if I see signals with an encryption (right side up or upside down) and a key (possession of the ace) that determines the encryption then I will say that these signals are encrypted.

Rik

A clear example of an encrypted signal is where the key is based on something unrelated to the actual hands, like the date or day of the week, the number of tables in the room, the weather and so on. Both defenders (but not declarer) then know the key because of some Concealed Partnership Understanding which is illegal in itself. On the other side Helge Vinje in his advanced defence system has signals depending on for instance the count of cards held by dummy in the suit led. His signals are definitely not encrypted.

If we follow these principles then varying signals according to (presumably) which defender holds a particular card, has any interest in the signal as such, or on similar key factors should be (and is IMHO) fully legal even if the key factor in the particular situation might be unknown to declarer.

[Trinidad]

 pran, on 2013-February-11, 03:40, said:

A clear example of an encrypted signal is where the key is based on something unrelated to the actual hands, like the date or day of the week, the number of tables in the room, the weather and so on. Both defenders (but not declarer) then know the key because of some Concealed Partnership Understanding which is illegal in itself.

That is not what is meant by an encrypted signal.

 pran, on 2013-February-11, 03:40, said:

On the other side Helge Vinje in his advanced defence system has signals depending on for instance the count of cards held by dummy in the suit led. His signals are definitely not encrypted.

And neither is that.

 pran, on 2013-February-11, 03:40, said:

If we follow these principles then varying signals according to (presumably) which defender holds a particular card, has any interest in the signal as such, or on similar key factors should be (and is IMHO) fully legal even if the key factor in the particular situation might be unknown to declarer.

That is closer to an encrypted signal.

The typical example of an encrypted signal is when declarer's length in a suit is known, e.g. after a Stayman or Puppet Stayman sequence. As a result, the defenders know the distribution of this suit while declarer doesn't. When the distribution in hearts is known, one could agree to use standard signals when one has an even number of hearts and upside down with an odd number of hearts.

In the above description of the method, the key is explicitly separated from the definition of the signal and it is easy to see that it is encrypted. However, it is easy to define the count signals as showing count in (the suit declarer plays and hearts). This description of the signal doesn't look encrypted, because the key is embedded in the description rather than explicitly separated. But is the same signal, the difference is that the encryption is described implicitly.

If you have the agreement to signal honest count with the ace and the exact opposite without the ace then that is an explicit description of an encrypted signal.

If you have the agreement to signal the amount of cards other than the ace then that is an implicit description of the same encrypted signal.

Encrypted signals are not disallowed by the laws, but many NBOs disallow them in their regulations.

Rik

[PeterAlan]

Pran, your post seems to me to be rather muddled.

 pran, on 2013-February-11, 03:40, said:

A clear example of an encrypted signal is where the key is based on something unrelated to the actual hands, like the date or day of the week, the number of tables in the room, the weather and so on.

The fact that such signals are not permitted has (or should have) nothing whatsoever to do with encryption - it's because once you allow unlimited external factors to be taken into account in signals then the potential information content of a particular play (or sequence of plays) becomes much wider, and you can't limit things to binary keys.

 pran, on 2013-February-11, 03:40, said:

Both defenders (but not declarer) then know the key because of some Concealed Partnership Understanding which is illegal in itself.

This is not necessarily so.

Both these points are illustrated by the proverbial LOLs - "Maisie knew my lead was a singleton - I led with my left hand. It's on our system card you know." Greater information content, no CPU, clearly out of order. Similarly, even if not concealed, it would be wrong to use the number of fingers holding the hand to convey the length of the ♥ suit ...

The basic principles are those enshrined in Law 16A. Regulators have seen fit to go futher in restricting the use of that AI in encrypted signals. I've not thought it through enough to have formed a firm view on whether this is a good thing or not, but it's clear to me that the issue of encrypted signals is about how AI is used and not about how UI might otherwise be brought into the picture.

 pran, on 2013-February-11, 03:40, said:

If we follow these principles ...

You need to delineate some principles clearly enough for others to use them to determine the boundaries of what they would or would not permit. IMO, you haven't done this.

 pran, on 2013-February-11, 03:40, said:

... has any interest in the signal as such ...

This quasi-test seems to me to be asking for MI trouble. There may be some clear-cut cases - there are plenty more where (a) the defender has an interest in a signal (and declarer knows it), but partner doesn't realise it; (b) the defender has no interest in a signal (and declarer knows it), but partner thinks (s)he has; etc. Declarer has to know the potential signaller's state of mind.

[paulg]

 Trinidad, on 2013-February-11, 04:31, said:

If you have the agreement to signal honest count with the ace and the exact opposite without the ace then that is an explicit description of an encrypted signal.

If you have the agreement to signal the amount of cards other than the ace then that is an implicit description of the same encrypted signal.

I think we'd all agree with the first.

The second is more contentious if your agreements are always to signal the number of small cards, whatever dummy holds. I accept that the consequence can be an encrypted signal, but it is not by design.

And would you describe Fantunes lead system, where the lead gives count of the number of small counts, as encrypted?

[Trinidad]

 paulg, on 2013-February-11, 05:52, said:

I think we'd all agree with the first.

The second is more contentious if your agreements are always to signal the number of small cards, whatever dummy holds. I accept that the consequence can be an encrypted signal, but it is not by design.

And would you describe Fantunes lead system, where the lead gives count of the number of small counts, as encrypted?

I fully agree with you on this one. A signal that was never intended to be encrypted may become encrypted due to accidental circumstances.

A common agreement is to lead 4th best from a picture card and second highest from a worthless holding. If a pair uses that agreement on every hand, then one day a player may lead the 2 of a suit (i.e. small doubleton or 4th best with a picture card) when dummy may hit with AQJ in the suit led. Now if third hand holds the king, the lead ends up being nicely encrypted. He can signal encouragement when declarer inserts the queen and when partner gets in, he can continue the suit. Third hand's king wins and gives his partner a ruff.

So, if the agreement is to always signal the amount of small cards in your hand, then in the case of the ace missing it is a coincidence that it has the effect of encryption. But if you agree with partner that for the specific case of a hold up play you signal the number of small cards in the suit then it has the intent to be encrypted.

The fact that perfectly normal signalling methods can end up being encrypted shows that it would be a good idea to allow encrypted signals. What is a TD to do when someone leads to 2 from a worthless doubleton/four to a picture card. Adjust the score since a perfectly normal standard lead becamen encrypted and, therefore, forbidden?

Please, please don't explain to me that the TD will not do that because it is a standard leading convention and only happened to turn encrypted without intent. I can guess that myself. But think what happens when players start to design their leads and signals in such a way that the probability that they will end up encrypted is maximized. Are we then talking about encrypted signals that sometimes have an easily accessable key or about normal signals that often end up encrypted?

Rik

[blackshoe]

I think you people are discussing a very flawed definition of "encrypted". It is ridiculous to say a signal becomes encrypted because of some coincidence which has nothing to do with the agreement as to the meaning of the signal.