[pilun]

When you learn relay, you want to relay out everything, even if it's just for practice.
That's when we see eight bids to reach everyone's 3NT.

You discard that approach soon enough. Your occasional gains are more than offset by the help you give the defence, plus there's the time issue.

These days we seem to relay much less often.
For instance we might bid 1♥ - 4♥ on hands where responder can picture partner with a perfect minimum (that principle from Jeff Rubens) or at least a perfect non-maximum.

We relay opposite 11-15 with invitational hands.
1♥ is (11)12-15, 4+♥s, not 4♠. Responder relays with 11+, planning an invite next.

Very common is
1♥ - 1♠
2♣ = balanced, includes 5♥332.

Then 2NT and 2♥ are invitational. 2♦ is GFR. What is 3-of-a-suit?

We could play them as invitational 6-carders but we bid 1♥ - 3♣/♦/♠ as intermediate.

The Balanced Hand principle (which I associate with David Morgan) says that a shapely hand opposite a balanced one is better off describing, rather than asking.
We've all been there - holding a singleton opposite describer's 4-card suit, wondering.
With that in mind, we tend not to relay with good support for partner and a singleton:

♠Axxx ♥KQxx ♦x ♣KJxxx

We play 1♥ - 2NT as limit+, often with a side singleton. We might also do it with a balanced hand, again requiring opener to show shortage. So it's Jacoby-style. More hands where we used to relay.

Back to
1♥ - 1♠
2♣

3-any by responder is shortage, even 3♥. Shortage is usually a singleton, though a small doubleton is possible in some auctions.
1♥ - 1♠
2♣ - 3♠ is more stopper-ask, remembering opener has denied four spades.

♠AKxx ♥Qxx ♦x ♣KQxxx

would bid 3♦ over 2♣. Finding shape instead may not help much. Feed information that will help the balanced hand decide.

In any relay auction where partner's last bid showed some balanced hand, we play all non-relay bids by asker as shortage.
For instance

1♣ - 2♣ = 8+ bal, with a 4-card major (2♦ is the same with no 4cM)

2♦ = shape ask
2M = natural, 5 cards, minimum, unsuited to relay (where you judge that finding shape may not be enough)
3x = shortage

3♠ would typically be 1-3-5-4, or the same with a 6-card minor.
You are saying "You'll need a double stopper in spades to bid 3NT."

You might try it with 1-6-3-3, removing 3NT (or 4m) to 4♥ to show this. Neat!

You don't do it with 1-4-4-4 say, because 4♥ may be better and partner may do the wrong thing.
What you can do with that shape is relay once, then show your shortage next time, so

1♣ - 2♣
2♦ - 2♠ = hearts & a minor
3♥ = shortage

We also break opposite a positive to 1♣ on some minimums, short (0 or 1) in partner's suit.
Reverse relay is hard work for minimal gain, so we switch to natural

1♣ - 1♥ = 4+ spades, not flat

♠x ♥AQx ♦KJTxx ♣AQxx

1NT (or maybe 2♦) shows this.

[DavidKok]

I think this is a great post and a clear explanation. Some of the information leaves some room for debate though. Also let me suggest a different tongue-in-cheek guiding rule: "The less often you relay, the better it is for your score". This seems partially compatible with your observations, discarding multiple relay sequences to make room for dialogue.

More seriously, I think there are two primary situations where relay can be very strong:

• The balanced hand, ideally a strong balanced hand, asks while the unbalanced hand describes.
  
• We are in a slam auction, maybe a miracle slam auction, with a big fit and lots of extra values.

When I first learned relay it was advertised to me as being efficient, concealing information, helping with most game decisions, being obviously superior for slams, and arguably even helping on partscore deals as a single game forcing bid frees up the rest of the system for other hands. I no longer believe the marketing. Only the slam bidding claim seems to hold, in a weaker form - all the others don't match vugraph results, simulations and personal experience.

I think choosing fast arrival over fishing for a perfect hand is a good idea. More generally, I think there are a lot of strong hands where it would be preferable to have a dialogue auction over a relay auction. However, to me the choice of removing big fit hands seems like the wrong choice - those are the precious slams where relay hopes to compensate for its losses.
Relay breaks are really difficult if responder can have a reverser. Breaking into dialogue bidding is challenging if the only suit shown need not be the anchor suit of the hand. Personally I would not consider the 1=3=5=4 hand shapely enough to break the relay.
Lastly I don't like putting invitations in the relay. If there is any interference life can get very difficult. What are your competitive agreements after a fouth seat overcall?

[pilun]

DavidKok, on 2026-August-02, 23:50, said:

Relay breaks are really difficult if responder can have a reverser. Breaking into dialogue bidding is challenging if the only suit shown need not be the anchor suit of the hand. Personally I would not consider the 1=3=5=4 hand shapely enough to break the relay.
Lastly I don't like putting invitations in the relay. If there is any interference life can get very difficult. What are your competitive agreements after a fouth seat overcall?

We have relay auctions to slam. We also have relay auctions that go nowhere. It's tough because the decision has to be made early on. We missed

♠KQJx ♥K ♦KQT ♣JTxxx

♠ATxx ♥Axx ♦A9xxx ♣x

North opened 1♦ (spades). South found shape and strength, then gave up.
North had that sinking feeling relay players get with stuff in the short suits.
So it turns out that 2NT over 1♦ would have been better. I think South needs better trumps for that. Tricky.

Some (not many) of our relays are GI only. Over 1♥ (hearts, not spades) we do

1♠ = 11+ any
1NT = 6-10, could have spades
2♣/♦/♠ = natural "less than invitational"

So no spot for 11-12 balanced, or a 3-card limit raise.

1♥ - (nb) - 1♠ - (2♠)

figures to be a problem. Opener can double for take-out, canape to a minor (admittedly strength still unclear) or volunteer 2NT.
If 2♠ is passed round to responder, 3m is forcing, double is takeout, often a 3-card limit raise (then 2NT scrambling).

Not so great but we prefer to keep the non relay stuff natural. Not keen on transfer responses.