[1eyedjack]

This is a hand that I kibbitzed today in Unia Lezno v Computerland viewgraph

Scoring: IMP

---

East-West = Unia Lezno
North-South = Computerland
West North East South
Bizon Balicki Kiwalski Zmudzinski
 -     -     -     Pass
 1♣!   1♦!   1♠!   Pass
 1NT   Pass  2♣!   Pass
 2♦!   Pass  2♠!   Pass
 2NT!  Pass  3NT   Pass
 4♣!   Pass  4NT   Pass
 7♥    Pass  Pass  Pass
 

We are not told much about the complex relay methods employed except that 1♣ by West was strong, and 1♦ overcall showed ♠.

We are not told what constraints there are on the 1♦ overcall but I speculated that a 2♠ raise by Zmudzinski might have disturbed their relay structure. Although we cannot be sure, it was suggested to me in the forum that the raise would have been ineffective in stopping their relays despite that (even after taking into account the availability of pass and double) it would remove two steps from their continuations.

Perhaps he had no way of knowing, but a 3♠ raise by Zmudzinski could have caused some problems. Goes for 1400 at best defence but may get out for 1100 if he gets to ruff a ♥ (unlikely I know). Still, that is a 2 IMP gain against 6♥+1 in the other room

So my questions are this:

There must be a point at which intervention consumes so much space that relay continuations are impractical. What is that level and, if it is such a level that it removes available steps (as in the suggested 2♠ raise above) how do you cope with the displaced hand types? Do you simply shift all of the continuations up by the relevant number of steps?

Also, if the intervention is so cheap that it actively gives the opponents additional bidding space (as in the 1♦ overcall above) how do you take advantage of that additional space? Do you just shuffle all responses down a step?

If you are playing some sort of Symmetric relay such as Moscito where they are symmetric in an absolute sense rather then in a relative sense then then it looks as though your problems are multiplied by intervention (if you want to preserve the relay structure), but I may be mistaken.

[hrothgar]

The extent to which an overcall adds or subtracks from the bidding space available is a function of the level of the overcall.

Assume for the moment, that relays are established

The Relay responder (RR) has just bid 1H (hypothetically) showing an unbalanced hand with 4+ Spades: In this case, double adds two steps of bidding space
Pass by the Relay Asker is the relay ask. at which point in time,

(a) XX by responder is first step
(B) 1S is second step
...

In a similar fashion, a 1S overcall adds 1 step of bidding space:
Pass is the relay ask and

(a) X by the relay reponder is first step
(B) 1N by the relay responder is second step
...

A 1N overcall is neutral.
A 2C overcall consumes 1 step of bidding space
...

The decision when to break relay is question of system design.
Here is the set of rules that I prefer:

1. Double or Redouble by R is always to play
2. Double or Redouble by RR is to play if RR could have 4+ cards in the suit.
3. Otherwise, relays continue so long as the level is < +3
4. If relay is established, but then broken due to high level interference, new suits are game forcing and natural.

[slothy]

You didnt mention the fact Monocular-Jack that he made an overtrick

[1eyedjack]

Hi

I appreciate the feedback. I am still a little confused.

I wish to concentrate on occasions when active bidding space is consumed.

The symmetric relay system in Moscito is based on "absolute" rather than "relative" symmetry. Any 7330 hand is resolved at the level of 3♠ in an uncontested auction.

The way you describe it (if I get it correctly), if a competitive bid removes bidding space but you decide that the space consumed is sufficient to justify continuing with relays, then the solution is to shift the bids up by the amount of space consumed.

Thus, if the competitive bid removes precisely one step, then 7330 hands would be resolved at the level of 3NT.

I have a slight problem with this. One of the advantages hailed on behalf of symmetric relays is the ease of memory. If we assume (with I think some justification) that most 1C auctions will be contested, this argument dissipates somewhat.

Many years ago I developed a fairly simple relay precision system. In response to 1C, 1D = neg, 1H = 5+Spades unbalanced, 1S= balanced, 1N = 5+Clubs unbalanced, 2C = 5+Diamonds unbalanced, and 2H = 5+Hearts unbalanced (2S+ showed 4441 hands and certain solid suit hand types).

After (say) 1C-1H, 1S = relay, and first step shows extra Spades, 2nd+ steps pattern out with minimum Spade length. If you show extra length with 1st step then after relay, 1st step shows extra length again and so on. This method has some symmetry but not in the sense of showing a particular shape at a particular absolute bid, but rather in the sense that it shows the same hand type relative to the starting point (1C-1H-1S-1N = 6+S, 1C-2D-2H-2S = 6+H, etc).

I don't claim that this is more efficient in an uncontested auction. It almost certainly is not. I do think that it is no more difficult to remember, and, more importantly, if you are going to continue with relays after space-consuming intervention, it is (in my opinion) probably easier to remember and apply, because you are already in a step-counting mindset rather than a specific bid mind-set.

There is also a problem of shifting priorities: If the opponent's bid is to be believed (and perhaps it should not be), then showing (or denying) a guard in the opponent's suit assumes an importance that did not exist prior to the intervention, and the pre-designed relay continuations are probably not geared to giving that information at a stage in the auction as early as the priority indicates.

Just some thoughts. Not convinced about any of it just yet.

[1eyedjack]

slothy, on May 2 2004, 01:25 PM, said:

You didnt mention the fact Monocular-Jack that he made an overtrick

Umm, yes I did. At least I quoted that in the other room it was 6♥+1. A bit difficult to make an overtrick in 7♥ in the room I was kibbing

[hrothgar]

>The way you describe it (if I get it correctly), if a competitive bid removes >bidding space but you decide that the space consumed is sufficient to justify >continuing with relays, then the solution is to shift the bids up by the amount of >space consumed. Thus, if the competitive bid removes precisely one step, then >7330 hands would be resolved at the level of 3NT.

Correct.

>I have a slight problem with this. One of the advantages hailed on behalf of >symmetric relays is the ease of memory. If we assume (with I think some >justification) that most 1C auctions will be contested, this argument dissipates >somewhat.

The ease of use argument continues to apply. In order to learn apply a relay sequence, players need to learn a relay sequence and Symmetric is still, by far the easiest relay structure to learn. Moreover, once you actually memorize a relay strucutre, its fairly easy to "hop" up for down a fixed number of steps.

>if you are going to continue with relays after space-consuming intervention, it is
>(in my opinion) probably easier to remember and apply, because you are >already in a step-counting mindset rather than a specific bid mind-set.

You are still in a specific bid mind set, however, its the specific bid that has changed.

>There is also a problem of shifting priorities: If the opponent's bid is to be >believed (and perhaps it should not be), then showing (or denying) a guard in the >opponent's suit assumes an importance that did not exist prior to the >intervention, and the pre-designed relay continuations are probably not geared >to giving that information at a stage in the auction as early as the priority >indicates.

Relay auctions are intended to determine shape at a low level.
Relay auctions are not intended for low level discovery of stoppers.
If you care about stoppers, then break relay and transition to natural bidding.

[luis]

1eye:

I play the following simple rules with my pds when playing a relay system:

The relay chain is broken if it consumes 2 steps. If it only wastes 1 step we follow with relays 1 step up.

Example:

1c - 1h
1s

Now a 1NT or 2c overcall waste 0 spaces, 1NT actually adds 1 space since you can pass or double 1NT. Over 2c pass=1NT and double = 2c.
2d wastes 1 space so we follow 1 step up (pass=1nt, dbl=2c, 2h=2d etc)
2h breaks the chain

When the chain is broken relayee bids naturally and relayer still relays.