Join Bridge Winners
Transfer Honeymoon Relays
(Page of 5)


This article is something of a muse out loud - looking both for constructive praise (of good ideas), criticism (of bad ones) and further suggestions in context.  If you aren't into relay methods, please tune out now, but if not ...

... Honeymoon* Relays (typically on top of a MOSCITO platform) have been around since the 80's, but I haven't heard much about them since the early 90's.  If you want a primer on them, look at this PDF:

Note, I am not advocating MOSCITO itself - nor I am specifically against it.  My interest is in this case is purely in the creation of a generic relay platform which is relatively efficient, but not optimised, which can be applied quickly to any arbitrary Strong Pass or Strong Club opening structure - mainly as a rapid yet competitive test vehicle**.

Why Honeymoon Transfer Relays (HTR's)?  Well, most relay methods when actually played are optimised to varying degrees so that the hand which shows shape is less likely to be declarer, so why not try and partially optimise a relay structure out of the box.  As I deal with systems which on occasion show shortage as well as length, please note that I am actually advocating two types of "transfer"

  1. Transfer TO  length, e.g. bid 's to show length, but also
  2. Transfer FROM  (unknown) shortage, e.g. bid 's to show shortage


Why transfer from shortage?  So that the opposition cannot DBL the bid by relay responder (RR) showing the shortage, nor a relay continuation by relayer (R), to show length in the suit, for lead/overcall/save purposes.


* so called, as first timer partners can in principle play them with minimal/no discussion

** for my pet/preferred methods, I have a symmetric-like structure which caters for showing shortage first, but if I'm trying to convince partner to play a new method, I don't want them having to go through the pain of learning it - just to test the efficacy of the openings themselves (which is far more important)

So, what is a substantively sufficient rule set to define Transfer Honeymoon Relays?  Here's a first cut

(1) If there is known to be a shortage, but which shortage is not known, show it by transferring from it.

e.g. In a method where 1 shows 9-13 hcp short (0-1 length) in one or both minors, as in one I'm interested in, then after 1 R, 1 shows short 's, 1 shows short in both minors, 1NT shows a non-maximum with short 's, 2+ shows maximum with short 's, forcing to game, with other shape according to subsequent rules.

NOTE 1a:  The reason for 1 versus 1NT+ being the way around it is, is that it's one step more efficient for the more common hand type of the two.  Perhaps this logic should be a rule in itself?

NOTE 1b:  The reason for 1NT being a MIN will be covered below.

(2)  If not already in a game force auction, then the first step to the next R shows a non-maximum in context, the second step and above show a maximum, forcing to game, with other shapes shown according to the subsequent rules

e.g. After 1 1 1 as above (short 's), then 1 R, 1NT shows a non-maximum, 2+ shows a maximum

NOTE 2a:  The reason for the "non maximum" nomenclature, rather than the easier to roll off the tongue "minimum", is to handle wide ranges.  After, say, a strong 1 opening (e.g. 15+ hcp) it is common to play 1 as a wide-ranging negative, e.g. 0-8 hcp, after which 1 R, 1 might be 0-5 hcp with 1NT+ showing 6-8 hcp and forcing to game.  Hence, 1 shows a non-maximum, rather than a minimum, and might be followed by a 1NT (or 2) further relay, applying the same rule again, i.e. first step, 0-2 hcp, other steps, 3-5 hcp.


The next two rules cover the main shape showing principles

(3) Length not already denied (starting with 4+) is shown by transferring to it as cheaply as possible, with one caveat:  it denies length in the next suit unless you also have the suit after (suit order being defined as ->->-> revolving, i.e. 's also follow 's).

NOTE 3a:  The caveat above is best explained by two examples.  Imagine you held 4=4=3=2 (Bridge World style nomenclature) in response to a 2 R.  If you bid 2 to show 's, then 3's to show 's, you'd be a step higher than having bid 2's to show 's then 3 to show 's.  However, if you instead have 4=4=1=4, then it's better to show 's then 's then 's (2 2 3 3 3) as if you show 's first, you won't actually be able to conveniently show 's below 3NT.

NOTE 3b:  If you've previously denied 4+ length by bypassing the opportunity to transfer into a suit, transferring into it on the next time around shows exactly 3 length.

e.g. As above to the 2 R, it goes 2 2 3 3 3.  This shows a 3=4+=4+=2-, or exactly 3=4=4=2 if known to be balanced.  With 's and 's (and not 's), RR would have shown S's first, therefore they can only have 3.

(4) Transferring into NT shows the most balanced (least shapely) hand possible suggested by the previous auction, with one caveat:  if there is known shortage, the meaning of a transfer into NT and a transfer into the shortage (explained later) is switched.

e.g. In a method where 1NT shows a BAL range, 2 is R, 2 might show a non-maximum (if appropriate) but in any case, 2 would show a 4333 (but not 4=3=3=3, for reasons described below)

NOTE 4a:  "Most balanced" starts as a 4333 if that's possible, and goes out from there.

NOTE 4b:  When a shortage is known, showing it again shows a void.  For the same reason we transferred from it originally, we switch the meaning with the NT transfer as alluded to above.  So, using the a different example this time, a 1 opening showing 9-13 hcp with short 's, then after 1NT R, 2 would show a non-maximum, 2 shows the most balanced hand possible (in this case 4=1=4=4) and 2 shows a heart void. 

To test your (my?) understanding of some of the above concepts wrapped together, 2 specifically denies a heart void with both 4+ and 4+, as with this type, we'd show 's, then 's then the heart void (1 1NT 2 2 2NT 3 3)

NOTE 4c:  Related to the above, with a 4=3=3=3, show minimum shape first, then spades, to save a step (analogous to showing spades then hearts with the 4=4=3=2 type).  That is, 2 2NT 3 rather than 2 2 3.


The next two rules cover jump variants to shape showing.

(5) A jump transfer to NT describes a 5/5 (read as 5+/5+) in context.  If shortage is known, then the 5/5 is clarified as far as possible by exclusion (bidding the suit you don't have).  If the shortage isn't known, then the 5/5 is clarified according to normal length showing principles.

e.g. After 1 showing 9-13 short in one or both minors, as before, then 1 R, 2 is a 5/5 maximum with short clubs.  After 2NT R, 3 shows &, 3 shows &, therefore 3 shows & by elimination.

e.g. After 1 showing 15+ artificial, 2 shows an FG 5/5, 2NT is R, 3 shows & black (with &, show 's first), 3 shows & minor (with &, we would somehow show 's first) which leaves 3 and 3+ to show the other two combinations in some order.  Of these, 3 should show & (so as not to bid if we don't have them), so 3+ shows & (zooming, which we'll come to later).

(6) A jump transfer to a suit shows (at least) two more length than a transfer and repeat transfer would - so usually 6+.

e.g. 1 showing short hearts, 9-13 hcp, 1NT R, 3 shows 6+ (most likely shape 6=1=3=3).

NOTE 6a:  With a 6/4, it might well be correct to show the four card suit then jump in the 6 card suit, e.g. 1 15+ hcp as above, 2 showing , 2 R, 3 showing 6+.  This is true with MOST suit combinations.  However, with adjacent suits, and in particular, & , you should probably suppress the 's (initially, and perhaps entirely:) and jump transfer into hearts.

NOTE 6b:  Logically, when showing 6+ as in the example above with 3+, you by default zoom into controls (once again, more on that later).  However, with uncapped length, it is perhaps better to zoom showing extra length?  Something to consider ...


That's most of it.  The rest is tidying, and (in the case of control asks) subject to one's own preferences.

(6) Basic rules:

  • 3NT is always to play by R
  • Breaking relay is natural and tends to deny slam interest, but is FG still unless opposite a known non-maximum
  • The last possible shape asking R is 3 (note, this is different to the 3 suggested in Buchen's work).  In responding to a 3 R, focus on the majors if they remain relevant - 3 would tend to deny extra spade length if relevant, 3NT would tend to show it.

(7) Control asks can be made explicitly using bids from 3 through 4.  In order of preference, if enough steps are available, they are

  1. A=3 based (in which case, the base step is 60% of the MIN expected high card strength, rounded down, so for a 9-13 hcp opening, 5.4 rounded down to 5)
  2. A=2 based (in which case, the base step is 30% of the MIN expected high card strength, rounded down, so for 9-13 hcp opening, 2.7 rounded down, so 2).
  3. A=1 based (in which case, 0/3, 1/4, 2, with "kicking" for below slam sign-off when containing 3 or 4).

NOTE:  If the last RR shape response is 3 or 3NT, you lose the type [3] control ask above.  If the last RR shape response is 4+, you lose the [2] & [3] control asks above.

There is some assumed knowledge in "base", "kicking" and "zooming" above, which I haven't explained but am happy to if needed.  I won't go into Denial Cue Bids and similar here, as that risks going down a rabbit hole when the above and previous is far more important.

As I mentioned at the start, I would be grateful for your input.

Ian C

Getting Comments... loading...

Bottom Home Top