by Jeff Rubens

Did you ever wonder what joggers think about during their long hours pounding along? For me, it is insufficiently diverting merely to be running and listening to music. Usually, I also analyze relay methods. Although, predictably, most of my mental researches have led to negative conclusions (as Thomas Edison once said, I now know 50,000 ways it doesn't work), occasionally I mine a nugget of practical value. Why, then, have I not been sharing these results? Three reasons: (1) It takes some doing to persuade players to adopt a relay technique, which is always substantially more complicated than more natural methods. (2) It is very difficult to present a relay method. To give incomplete explanations reduces persuasiveness; to give full details risks glazing the eyes of all but true believers. (3) In all but the shortest auctions, it is impossible to describe a relay method accurately except in the context of other partnership agreements. As there are many such contexts . . .
   In the case of relay methods in what theorists call "third-suit situations" (natural suit one-bid, simple suit response, simple same-suit rebid; example: one diamond -- one spade -- two diamonds -- ?), I'm running out of excuses. In "The Bourke Relay," (July 1996 Bridge World) David Bird and Tim Bourke have done a fine job with (1) in a small space. Examples from the Master Solvers' Club dramatically support their view that standard third-suit bidding is so inept as to make it worthwhile to consider artificial methods. Regarding (2), successful experiments by producers of match records, notably Eric Kokish and Tony Sowter, have made abbreviations an acceptable form of providing telegraphic descriptions of meanings. It is now only a small step to add my two cents to this evolving technique. There is still no known way to get around (3); to cope with that problem, I intend to cheat.

Relay Styles and TSAR

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"); (B) 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); (C) 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).
   This article describes a significant exception to the three-style standard, a rarity that I have discovered over all those many miles, a principle for relay bidding when one player has described a limited hand with a long minor suit (plus some other distributional restrictions) through a sequence ending at two of the long minor. In standard methods, this occurs in third-suit situations, defined earlier. TSAR, an acronym for Third Suit Alternative Relay, named for its most important standard-bidding application, designates a principle of bidding in situations meeting the requirements in italics earlier in this paragraph. Let's call the player opposite the long minor the Tsar; he can have absolute control over the auction if he wishes, but also has alternatives that are neither fixed style nor skip style. For purposes of examples, let's assume this sequence:

1 1
2    ?

To read the next paragraphs, you will need to know this useful but uncommon definition: quasi-game means "three notrump or four of the agreed minor." Typically, when an action is forcing to quasi-game the partners bid stoppers up the line to see if three notrump is playable. If it isn't, the bidding may -- but might not, of course -- stop at four of the minor. In such a situation, a takeout of three notrump is strong.

Under the TSAR principle, the Tsar's bids have these meanings:

(a) In his own suit: two is natural, at most mildly invitational; three is invitational, suggesting a strong suit; four is natural, a mild slam-try with a powerful suit (all as in Bridge World Standard).

(b) In notrump: two notrump is natural, invitational, suggesting a misfit for partner's suit; three notrump bars partner. This is mildly different from standard, where two notrump is ambiguous as to hand-type, fit or misfit. Under TSAR, two notrump says the Tsar intends the partnership to take approximately two tricks in each suit should partner pass, an extra one somewhere should partner raise; perhaps,

A Q x x x   K J x   x   Q 10 x x.

Using three notrump as natural may seem -- well, natural -- but in fact it is a hairline decision.

(c) In partner's minor: to three is natural, forcing to quasi-game, asks partner to show stoppers at the three level; to four (and, for that matter, any non-original-suit action above three notrump) is something special. This is quite different from standard, where the direct raise to three is invitational. Under TSAR, the Tsar shows support and his general intentions. In contrast, in standard, to show his strength he would have to make up a bid that is artificial, does not show the fit, does not focus on the main issue, and in some cases is dangerous to bidding health because partner might raise (or, in some cases, otherwise preempt the auction undesirably).

For a raise to three diamonds in the example sequence, the Tsar might hold any of these hand-types:

A Q x x   A J x   Q x x   x x x.

A Q x x   x x x   Q x x   A J x.

A Q x x   x x x   A Q J   x x x.

Because the Tsar has announced interest in stoppers, rather than in shape, after the direct raise the meanings of bids on the three level are stopper-oriented. In the example sequence, over three diamonds opener could bid three hearts with a heart stopper, three spades with a club stopper, three notrump with both, above three notrump with neither. Not to worry at this time about secondary support for spades. If that were paramount, the Tsar would have relayed, as we will see below. Also, with a hand like,

K 10 9 x   A x x   Q x x   K Q x,

the Tsar will relay, intending to head in one direction opposite short hearts but in another opposite short clubs.

(d) The cheapest new-suit bid is a relay, artificially asking for information about shape and strength, but restricted to hands of these types: (d1) invitational strength with a (possibly mild) fit for partner's minor; (d2) quasi-game-forcing or stronger, with a fit for partner's minor; (d3) three notrump or greater strength but lacking a shape that can be shown directly.

Bourke-relay users may find it helpful to think this way: Under TSAR, responder relays on the same hands as in The Bourke Relay except that TSAR switches the notrump-try strong hands to three of the long minor and replaces them with the invitation-with-fit hands. This is the essence of TSAR. It is another peculiarity of TSAR that even though it is usually more useful to have relayer's partner offer an additional piece of description when relayer is strong, in this case it works well the other way. I must admit that it took me almost 2,000 miles to understand exactly why this is true here although not elsewhere, but most of those miles were run before I had formalized the Useful Space Principle.

Under TSAR, it is easy to figure out when to relay: You have an invitational-strength hand based on fit, or a stronger hand with which you can't do anything else. The key to later bidding is that with invitational strength, the relayer will not himself cross the three-of-opener's-minor barrier, and neither will opener in the absence of extra values. (Thus, after a relay, when neither partner has extra strength and no new fit is discovered, the final contract will always be three of opener's minor.) With a stronger hand, the Tsar can show his own shape as soon as convenient, or keep relaying to find out more about partner's hand.

The tradeoffs among standard, Bourke and TSAR are these: Standard bidding is clearly worst, and by a lot. It is somewhat simpler, but not really that much -- consider that responder will have to bid some new suit with all three of the notrump-try examples shown under (c), above. Bourke is a hefty improvement over standard, in exchange for only a moderate amount of added complexity. Most regular partnerships will be better off with Bourke than with Standard. TSAR is perhaps a significant technical improvement over Bourke, but clearly enormously more difficult to implement. The amount of increased difficulty depends partly on how well you know the rest of your system.

(e) Other new-suit bids are used to fill in whatever hand-types cannot be bid conveniently otherwise. Their possible meanings are highly dependent on the specific sequence.


When I was at an impressionable age, I was strongly affected by New York City bus advertisements for stenography lessons in a method named, as I recall, Speedwriting. These ads read something like, "if u cn rd ths u cn gt a gd jb w hi pay." In what follows, I will use a form of Speedwriting to condense the relay-method descriptions. To emphasize my contention that the abbreviation system is simple, I will not tell you what the abbreviations stand for, just as in the original ads. [However, for the HTML version posted on the Internet, I must remark that it is inconvenient to use individual less-than or greater-than signs in text, so I have used two of each.] It is my hope that purveyors of abbreviations will be able to get together (via the Internet, perhaps?) in, say, a few years' time, to establish a universal set that everyone will be willing to use.

I have already explained why it is impossible to give details of a relay method to suit all partnerships. In the Speedwriting demo that follows, my main purposes are (1) to give you a sample to work from, and (2) to demonstrate a variety of useful relay-method construction techniques. To broaden the illustration of TSAR, I am going to change the example sequence, but I cannot give a more detailed explanation unless I make some assumptions. For instance, in my new example sequence, which is going to be,

1 1
2    ?

to provide full details for your partnership, I would have to know, at the least, when, if ever, you would rebid two diamonds holding three hearts and six diamonds, or holding four spades and seven diamonds, or holding four clubs and six diamonds. I am going to assume that your answers to these questions would be similar to my own: I would usually rebid two hearts with three-six in the reds, choose a rebid with four-seven in the pointed suits based on suit quality, and rebid two diamonds with six-four in the minors only with the minimum half of the valuation range for a two-diamond rebid. Chances are, you won't prefer my style all the way down the line, so you will have to make adjustments to suit your own preferences.

Obviously, each partnership must determine the level of complexity it wants to use. Well, there's no free lunch. If you want to win more with TSAR, you have to be willing to work for it. Using a relay method without full discussion is akin to walking in a cleared minefield from the latest war. Probably, you won't get blown up, but . . . In order to limit the complexity of the demonstration model, I use the term "fancy," followed in parentheses by a logical interpretation for the agreed suit, to incorporate such devices as you may prefer among void-showing, ace- or key-card asking, exclusion asking, trump-asking, or whatever. I have also suppressed some details in long or unlikely sequences.

If u cn rd ths, u cn lrn hw to bld yr own sys:

1 1
2    ?

2 : < I.
2 : R. F3.
2 NT: I. misfit.
3 : T [Opener bids as he would over a two-spade relay, or three diamonds if he would have bid no higher than that. Then, over 3 : 3 = s, short ; 3 =s, short ; 3 NT(F) or higher: big red two-suiter.]
3 : FQG. Stops?
3 : I.
3 : F1. s.
3 NT: Bar.
4 : N. Slam??
else: fancy ().

1 1
2 2

2 NT: 3=2=6=2 | 2=2=6=3 | 1=2=6=4* | max, ?=2=7=?
3 : 3=1=6=3 | 2=1=6=4* | 3=0=6=4*
3 : min, ?=< 3="7"=? or > 7 s
[Then, 3 : R.; 3 : s]
3 : ?=3=6=? (special hand, else would have raised hearts immediately)
[Then, 3 : R.]
3 : ?=3=7=?
3 NT: max, ?=< 2="7"=?
> 3 NT: max, > 7 s

*recall that possession of this shape implies a minimum two-diamond rebid

1 1
2 2
2 NT   ?

3 : R. [Then, 3 : min; 3 : max, ?=2=7=?; 3 : max, 3=2=6=2; 3 NT: max, 2=2=6=3. Over 3 , 3 : R (then, 3 : 1=2=6=4; 3 NT: else); also 3 : R (then 3 NT: 3=2=6=2; higher: else); note how the Tsar asks the question he wants answered, in this case inquiring either about four clubs or about three-plus clubs, a valuable "tight-space" technique in constructing relay methods.]
3 : I.
3 : I. (a "raise")
3 : FG. s.
3 NT: Bar.
> 3 NT: fancy ().

1 1
2 2
3    ?

3 : I.
3 : R.
[Then, 3 : 3=1=6=3 or 3 NT: else]
3 : FG. s.
3 NT: Bar
4 : FG. s.
else: fancy ().

These charts conceal some very close decisions. I usually decide such questions by following this principle: Make the least likely hand-type the one that is hard to bid. Sometimes, even that guideline is inadequate, so this tiebreaker is needed: Make the hand-type with which it is easiest to cheat hard to bid.
   For example, in the current context, I would like to make one diamond -- one heart -- two diamonds -- three clubs invitational, showing a concentrated two-suiter, something like,

J x   A Q J x x   x   K Q 10 x x.

However, I can't afford it. The concentrated heart-club type is uncommon. Hands with shaped support for diamonds, such as,

x   K Q x x x   K x x   A x x x,

are much more frequent, and it is essential to describe these below three notrump (as a few example 3=2=6=2 opening hands, some with two spade honors and some with no spade honors, will easily convince you). But Murphy's Law demands that anyone playing the system as shown will be dealt a series of invitational-strength heart-club hands! If you get one of those, you can cheat as follows: With two diamonds, or a singleton diamond honor, treat it as a diamond raise. With honor-low or three low in spades, bid two notrump. With six hearts, or five very strong ones, bid two hearts (which, if you use weak jump responses, preserves game chances) or three hearts. With a minimum, pass two diamonds. With a maximum, bid three spades, showing a strong heart-club two-suiter (forcing only for one round; for example, you can pass a simple preference to four clubs).

Another hand-type, less common, slighted by this particular implementation of TSAR: game-going strength with five-six (or more) in the majors. With that, the best you can do is to keep relaying while praying for rain. In evaluating the degree of difficulty you are likely to get into with this type, note that if the Tsar finds opener with two hearts, he will usually want to play in hearts even when opener has three spades.
   Let's look at some other kinds of tricky hairline decisions you may have to make. After, one diamond -- one heart -- two diamonds -- two spades (relay) -- two notrump (basically, two hearts) -- ?, it is not necessary to use three diamonds as invitational; you can relay with those hands. Then, you could use three diamonds as a transfer to hearts, invitational or stronger, with opener bidding three hearts with a minimum, or higher with a maximum, thus freeing up bids at the three level (for major five-sixes, for example). Or, you could make three hearts nonforcing, three diamonds stronger in hearts, and still have three spades available for five spades. However, relaying with the invitational hand will give the defenders more information about the shape of declarer's hand at three notrump. As against that, some of that information has already been given away. Another hair to split: If three diamonds is invitational, the three-club relay need not serve that function and the replies to it could make it easier for opener to show shape, and harder to show range, rather than the other way around (as I have it).

TSAR in Other Situations

The idea behind the TSAR principle was (to my knowledge) first suggested by the late Monroe Ingberman for bidding after a natural, limited two-club opening, the kind popular in strong club systems. Using a home-brewed, TSAR-based method, he and I bid about 250 practice deals (think of that as two full years of Challenge the Champs) where opener's limited, long-suit two-club opening denied a four-card major. Post-mortem analysis strongly suggested that the TSAR principle was greatly superior to all competing products. My restricted experience in other two-club-opening contexts limits the value of my opinion there. Still, for the record, I believe that the TSAR approach, modified to match the system of course, will work better than anything I have seen for bidding over any specific form of natural two-club opening.

In contrast, there are other, somewhat similar, sequences in which TSAR is not helpful. For example, after a weak two-diamond opening it would be foolish to give up the preemptive raise to three diamonds even for a substantial constructive gain; the value of relaying for shape is lower when responder has not (as he has in third-suit sequences) announced his own longest suit, and (also unlike third-suit sequences) the probability that the notrump declarership belongs opposite the long minor is very high (thus you want to be careful about which partner first says the word "notrump").

Other auctions in which TSAR is a candidate for best approach are less common. Anyway, I see it is time to go to the track. I'll try to check in with more relay results after the next ten thousand or so miles.


This section is devoted to weird, wild and wacky material. For bridge friends, lovers of arcana, pursuers of special interests, and anyone intrigued with a particular facet of the game of bridge.