[jillybean]

Hi, could someone please explain relay bidding?
Thank you

[hrothgar]

This question is MUCH trickier than it initially appears. I’ve never seen a “good” working definition of the word relay or even relay system.

At the most basic, a relay is artificial bid that asks partner to describe his hand. As an example, many people use a 2C response to a 1NT opening bid to ask whether opener has a 4 card major. While most people consider this an example of the Stayman convention, it is actually one of the oldest examples of a relay. This example also demonstrates one very common feature of relays: TYPICALLY, the relay asking-bid is a first step response (2C is one bidding step above 1NT)

Where life gets complicated is that most people want to differentiate between “relays” and “asking bids”. For example, the ACBL places lots of restrictions on the ability of players to employ relays, but allows considerable flexibility in using asking bids. In my mind, the crucial difference between relays and asking bids is “recursion”: Suppose that I make an asking bid and you make a relay. If I have the option to make ANOTHER asking bid over your response and get more information about your hand then these bids are relays. This definition is far from universal, however, it is the one that I feel most comfortable using.

With this said and done, relay bidding methods are typically based on one player making a series of asking bids while the other player describes his hand. Normally, the relay asking bids are broken into 3 distinct stages.

Stage 1: Defining the relay responders shape
Stage 2: Defining the relay responder’s approximate strength
Stage 3: Placing specific controls.

There is some argument whether it is better to initially show shape or strength. [Ultimate Club initially shows strength followed by shape. MOSCITO, Viking Club, and Ice Relay show shape followed by strength].

In addition to these three stages, there is another critical component to relay type methods: How does one “turn off” the relays. In many case, the relay asker will want to stop relays and transition to normal bidding. Alternatively, the relay asker often will need to stop making asking bids and signoff in game or even a part score.

In any case, this probably is enough of an introduction. I’d be happy to go on in more depth about relays in general or symmetric relay in particular if there is any interest. BTW, here is a quick quiz for anyone who things that their good at relays. Please explain what the following image has to do with relay methods:


http://sprott.physic...ls/fracday6.gif

[the hog]

You have got me with that question, Richard. How is the fractal like a relay?

I would say an asking bid is not a relay as it asks a question about a specific suit eg Gamma 1C 1S 2S (what are your S?)
Relays follow a defined path - shape strength etc as you say. Interesting argument.

[hrothgar]

Fractals and relays share a couple important characteristics:

1. Well designed relay structures (including Symmtric Relay) are self similar.

2. The amount of information that can be packed into a relay structure is modelled by the Fibonacci Sequence. Fibonacci sequences also feature prominantly when studying fractals (Easiest to see when looking at a Mandelbrot sets)

Folks might find the following web page amusing
http://www.jimloy.com/algebra/fibo.htm

[jillybean]

Thanks hrothgar, I see there is a lot more to this topic but you have answered my question completely, thats all the information I need for now - thanks!

[luis]

I've found Richard's introduction excellent and thanks for the pointers to the Fractal and the fib page, nice brain teasers :-)

I've stopped to think about the difference between a relay and an asking bid and I think that the best definition -in my opinion- is that a relay doesn't disclose any information about the asking hand while an asking bid directly or indirectly discloses some information.

In Ron's example the gamma asking bid discloses that the hand asking has spade support and thus is asking for the quality of the trump suit (spades) so they are probably going to play a spade game, etc etc.

So a relay auction defined as an auction where one hand bids only "relays" is a half-duplex exchange of information where only one player sends information to the other, natural/asking bids/non-relay auctions are full-duplex since both players send information to the other.

In terms of information theory this is of-course inefficient in uncontested auctions, I wonder what are the right tools to measure the effectiveness of both approaches in contested auctions where natural methods make sense since both players may be able to take a sensible decision at the right moment.... mmmmh... interesting.

[hrothgar]

I've stopped to think about the difference between a relay and an asking bid and I think that the best definition -in my opinion- is that a relay doesn't disclose any information about the asking hand while an asking bid directly or indirectly discloses some information.

From my perspective, the problem with this approach is that there are always negative inferences available based onthe decision NOT to break the relay.

Lets consider a very simple example like the MOSCITO auction 1C - 1D

1C is strong artifical and forcing
1D shows a game forcing hand and denies 5440 shape or a solid 7+ card suit

At this point in time, the strong club opener has two choices:

1. He can bid 1H as a relay and ask about responder's hand
2. He can bid 1S+ and show his hand using the same response schedule

Normally, responder will make and asking bid with a strong hand (11+ Slam points or so).
He will show his hand with a minimum strength unbalanced opener.

In this case, the decision to make a 1H asking bid implictly provides information about the relay askers bid.

There are equivlent relay breaks throughout the relay structure where the relay asker has the option to transition to natural bidding, ask for stoppers, ...

[the hog]

So a relay auction defined as an auction where one hand bids only "relays" is a half-duplex exchange of information where only one player sends information to the other, natural/asking bids/non-relay auctions are full-duplex since both players send information to the other.

In terms of information theory this is of-course inefficient in uncontested auctions, I wonder what are the right tools to measure the effectiveness of both approaches in contested auctions where natural methods make sense since both players may be able to take a sensible decision at the right moment.... mmmmh... interesting.


Maybe we should all work on a system where you can relay or not, depending on:
a) Hand type
Local system regulations, (lol).

We used to have lots of relay breaks to show limited 1C openers eg
1C 1H (Spades 9+)
2C (H + D, 15-17, with 18+ you relay)

This meant you could often place a contract in 2-3 bids without giving away much info to opps. Now with 1D pos to 1C, you can choose to relay or not; Justin and I played that 15-18 showed, 18+ or freaks asked.

[hrothgar]

Here is an interesting relay break structure from "old fashioned" MOSCITO.

Assume that you are playing a Strong Club opening.
Bids from 1H+ show game forcing hands.

Paul Marston argues (and I concur) that relay breaks should show unbalanced hands with shortness in the prime suit shown by responder. For example, if a 1H response to a strong club opening shows a positive with primary spades, Opener's relay break shows a singleton or void in spade. The "logic" behind this structure is motivated by two observations:

1. In general, relay structures work best when the balanced hand is asking and the unbalanced hand is showing
2. Its best to escape a misfit as low as possible.

The relay break is constructed as follows:

Opener will show suits from longest to shortest
Ties are broken in order H > S > C > D

Step = Relay
Step +1 = 4+ Cards in suit A
Step +2 = 4+ Cards in Suit B
Step +3 = 3 suited hands (5440 or 4441)
Step + 4 = Single suited in suit C

The structure is designed to revolve arround step +4 (a raise of responder's suit) which always promises a single suited hand pattern.

Otherwise, a standard 2 suited symmetric resolution scheme is used.

For example, Suppose that I opened 1C on

J
AK843
AK42
K42

Responder bid 1H to show 5+ Spades

The auction would proceed

1C - 1H [1C = strong, 1H = 5+ Spades]
1N - 2C [1N = 4+ Hearts, 0-1 Spades, 2C = relay]
2D - 2H [2D = Two suited with Hearts and Diamonds, 2H = relay]
2S - 3D = 1543 shape

After the 2S bid,

2N = reverser
3C = 5+/5+ shape
3D = 1543 shape
3H = 1642 shape
...

[mishovnbg]

Jeff Rubens:

"Most relay methods I consider best in a given context follow one of three styles: (A) fixed style, in which the relay is (in normal follow-up sequences) limited in strength (e.g., nonforcing Stayman), or perhaps upwardly unlimited (e.g., "forcing to game"); ( skip style, in which the relayer can later show a weak or strong hand but not one in-between (e.g., the cheapest response to one of a major shows, in pointthink, 7-10 or 13-plus); © absolute style, in which the relayer describes nothing, requiring partner to act robotically (as in many situations in full relay-based systems; Americans are most likely to have encountered this style in the Ultimate Club of Mike Becker and Ronnie Rubin)."

Misho