Dubious Logic

High Fidelity

You have to have a system; everybody knows that.

The hard part is getting a good one. In the main of course, you get out what you put in. I'm not sure how the question arose as we stood in the shop, first through the door early Saturday morning, but the answer was unequivocal.

'Over the years, the cost of a good hi-fi has kept pretty stable – about the same as a small family car.'

And if you were in any doubt about the definition of 'good',

'..That's one that will make you cry'

The guys made coffee, we settled into the sofa in front of the set-up which we'd asked to listen to and I reflected. My big budget now looked small; it had been bolstered by the insurance settlement after the theft of my motorcycle but however much the benchmark small family car cost, it figured to be a lot more than a ten year old bike. Then the man selected a track and June Tabor sang 'You Don't Know What Love Is'. I guess I'm easily moved.

Those days were not so long ago but we were talking vinyl. Records. Analogue. Natural sound, if the gods wanted us to listen to silvery disks they wouldn't have given us black plastic, grooves, styli and record covers. Paraphernalia, performance and romance.

I'd played a natural system for most of my life. Of course I fooled with conventions from time to time. But in all but a tiny minority of situations, I bid the suits I had and named the controls I held. The auction was a conversation, not an interrogation. But some time ago I started reading foreign magazines and then experimenting at home. As you might expect I was seduced by the bright shiny stuff. Digital. Symmetric relays.

True, I had incorporated some technology in the Zebra Club (scary bits left out of the Table Talk articles) but they were very rare, though I'm glad to say, effective, when they came up. This hand from the 2000 SCBA pairs final decided first and second place:

East just asked questions and West showed his hand. As an opening 1NT was 10-12, the next range of balanced hands was 13-16; a minimum any hand that didn't have 15 HCP in aces, kings and queens. 4♣ set diamonds and asked for key cards.

Knowledge of the shape is very important here – just put West's doubleton in clubs or hearts and slam would be poor except against exact cards. But I was fortunate to buy a dummy that meant the slam was almost cold. The spade king alone would have made the contract reasonable but it is easy to picture less suitable hands. With short spades, any outside king and queen would make 6♦ excellent - even against the field at pairs. Anne Wilmer / Ian Corbett were unlucky that the room started with a weak no-trump and didn't look any further than Stayman. It was an arrow switch too.

The auction above only became game forcing with East's 2♥; earlier actions might have been scrambling on weaker hands and despite the apparent use of space, West's shape might have emerged at an uncomfortable level, above 3NT. The natural domain of relay systems is the strong club. There you can create early game forces – positive over the club for example. The trouble was no-one wanted to play a strong club system. Not with me anyway. So I dreamt up a box of symmetric tricks to bolt onto our five card major system. It was 1993, they didn't want to play that either and I didn't spot that they weren't very good.

One reason for the lack of take-up – not the most important but as it is blame free I'll focus on it – was that the relays were a very small part of the system. Why learn all this tricky stuff if it came up so infrequently? Quite. The obvious solution was to invent a complete system and exploit the foolhardiness of youth. When Jonathan Green and I started playing together it wasn't entirely serious and we were game for anything.

Even the Suffolk Swede.

Lots of canapé and a strong club that was protected by having a weak no-trump ingredient. That's the Swedish motif and a good idea - the opponents can't be too energetic over your club opening as the hand might well belong to them. Ours was 17+ any shape or balanced 11-13 with no major. The responses showed five card suits and, many revisions later, my symmetric relay responses saw the light of day. The Swede got its first decent trial at the Tollemache 1998. It was a good event for it. Hardly any relay sequences came up. It wouldn't last for long.

My top five Swede slam disasters

The natural sphere of advantage for relays is in slam bidding. In practice it's not so easy. However, only a fraction of auctions will start with 'the method' – all your other openings have to take up the slack.

1) SCBA Club Teams of Four Final 2001

The strange thing is that this was a natural auction, including the opening bid. OK, 2♦ began a forcing sequence and 3♠ was fourth suit. Spade lead, won then ♣10 intending to unblock diamonds, take a second trump in dummy, thereafter more diamonds. South discarded on the first club. North ruffed the second diamond. In 6NT, our opponents took the same line except that their tricks were now 3 clubs, 4 diamonds and a spade for eight. Hearts were 3-3 queen on-side. That was 17 IMPs away.

So what's all this symmetry stuff anyway?

The very first hand is an example of step relays. Over the ask or relay bid – almost always the cheapest call – there is an 'agenda'. The informer is asked a question and differentiates holdings by steps. 2NT asked for suits and 3♦ showed some 5♦332 – 3♣ was 5 clubs, higher bids, 4-4 minors. Over the relay 3♥, West was showing his doubleton from the top, so spades happened to be the first step. The hand elaboration is reasonably safe. Yes, it would have been bad had West forgot and showed clubs rather than diamonds but that wasn't likely. More possible would be to miscount the doubleton reply, go low to high say. But that's not going to be so terrible; within the step relay structure the agenda has in-built fault tolerance.

The great attraction of symmetric relays is that all the shapes come out at the same level. There will be a bid that means, say, 5431. All 5431s will make that bid. That's good for memory and it's comforting to know you won't be carried 'over the top'.

The great danger of symmetric relays is that a single mistake gives rise to a completely different hand. Different long suit, different shortage, different pattern.

Neither method gains anything in terms of number of hands it can show. Let's say you have a relay of 2NT and you don't want to proceed beyond 3NT, how many hands can you show? Using step relays you might have 5 steps, 3♣, 3♦..3NT. Fine but not best. 3♠ and 3NT must show specific things as must 3♥ (when the relayer bids 3♠, you can only go 3NT) but why should 3♣? You can have another relay and extract 3 pieces of information from it. The same goes for 3♦. So:

Total 8. Counting these for a longer range – say starting at 1NT ending at 3NT, might well be tiresome from first principles.

If we're talking symmetric relays the same constraints about 3♠ and 3NT as information-points holds, but now we must count the paths to those.

Total 8

There are 5 routes to 3NT and 3 to 3♠. Continuing to work backwards, there are 2 to 3♥ and 1 each for 3♣ and 3♦. Fortunately Leonardo of Pisa, also know as filius (son of) Bonnacio discovered a neat way of counting these in the thirteenth century

The Fibonnaci sequence: the first two numbers are given and thereafter each is the sum of the two immediately previous. Many of nature's packing algorithms exhibit this outward pattern – stamens, spirals in shells and leaf buds – so maybe it's not so artificial after all.

All we have to do is select our information points to define the distributions we want and Fibonnaci will tell us whether its possible and almost how to do it. Let's say we start with 1♣ and partner responds 1 major showing the suit (naive but we'll let it pass), we bid 1NT to start the description.

We would like to show the second suit – out of three – then the third suit – out of two. That's six pieces of information that we need to cater for. The sixth number in the sequence is eight, the fifth only five. We therefore have to go as far as the sixth bid 3♣ for sufficiency. 5431 is a common type, so let's say that's what 3♣ stands for.

So over 1NT, we'll bid 2♣ with clubs, 2♦ with diamonds (distressingly natural this), ignore 2♥ but go to 2♠ and above with the other major. For the third suit if it's lower ranking bid, 2♠, otherwise 3♣. All the minor hands that bid 2♣ or 2♦ originally can get to the same structure of 2♠ and above (which is why we left out 2♥). The next information point, 3♦ can be 5521/5530 (see hand 3), 3♥ 6421 etc.

Example. 5=1=3=4; bids 2♣ [clubs], 2♠ [third suit, diamonds lower ranking], then 3♣ to show the 5431 shape. A 6=1=2=4 would bid the same except that the final call would be 3♥.

There are a lot of gaps in the scheme above – there will be two unused paths to 3♣ and five to 2NT. Some distributions only require 3 bits of information to describe them once the long suit is known we and can fit these into the holes. And that, give or take a bit (quite a bit) is what we play today. Oh and this is only for hand shapes, there's still the business of finding out if partner has the right high cards. Time for a hand.

2) London Business League 2001 v Law Society

There's a lot hidden here. One fact is that the annotations above are perfectly correct, we both were under the impression East had shown 2=5=4=2 (1N then 2♥ was correct). I couldn't bare to show a hand with 2½key cards as a 'bad hand for slam' (4♣) so admitted to my controls immediately. So much for judgement. With the ♣Q extra we would have had great play for a grand slam in clubs but as it was we were struggling in hearts. As the opening leader might have led a singleton club (he started a spade) I tried ♣A first and it transpired he held QJ alone. An apologetic 11 IMP gain.

When you have sloughed off the constraints of naturalness, you are free to address other issues. One aspect we have given a lot of thought to (and it's not finished yet) is that of 'right-siding'. Thus in our example scheme earlier over 1♣, it would be much better to respond 1♥ with spades and 1♠ with hearts, getting spades played by the opener and delaying a decision on hearts for as long as possible. There are two big reasons for this. The one club relayer will almost certainly be the stronger hand, and it will be undisclosed. It is very difficult to lead up to a hand that has bid no suits naturally either by length or by cue – one reason why missing 6♣ above was painful, we had done nothing to attract a diamond lead.

3) London Business League 2001 v IBM

One no-trump set a game force and relayed. Everything was fine until 4♦. East intended this as the start of a sign-off, West thought it set hearts. West imagined he had a good hand for hearts (known shape remember) and showed key cards. East took stock. East took stock again. As West I reviewed my auction.

No mistakes thought I until a cold, clammy realisation hit me. Over one club you can decide which of your suits with 5-5 to bid first. We choose the higher ranking if 5521, the lower if 5530. Thus 5-5 hands are always treated the same – if you hear a second suit that's higher than the first, the third suit is three cards. But you can't bend your whole opening one bids into that – even if you wanted to, we don't have a five card minor opening (2-level promises 6). So at this evolutionary stage there was an extra step upon which I had just alighted. Oops. Once again I had shown a suitable hand and when partner grabbed a lot of cards I half-expected 7♥. 6♠, phew!

Even better, he thought I was 5=2=1=5 (he had 6421 a bid earlier). We admitted that I might be 5-5 or 6-4 before the lead which was a small heart. I won the knave in hand and deliberated. The danger of a ruff seemed great so I lay down a top trump and got the queen from South. I would have been cold with trumps 2-2 but now I played a diamond to table and a club down. Good news/bad news as South won. He played a diamond. Eventually I took my ruff in dummy, crossing with diamonds twice. Hearts were 3-3, the whole hand took about thirty minutes (I haven't mentioned this aspect have I?) and was a push – the opponents had on their bidding socks too.

I should say that the original system with its canapé didn't last long. As we played it so infrequently, we tended to concentrate our memory efforts on the club sequences (you can see why). The four card major parts went un-revised. When Ed Colley came into the fold we made the shift to 5CM. You've no doubt noticed above that we play the same relay system over the major openings as we do to the responses to one club. We experimented with 1♦ promising a four card major but eventually adopted Ed's idea to show 3 or 4 spades. That allowed us to pre-empt from a modest base and it's been quite successful. The doubleton spade hands are the weak component in 1♣, so what do you with a singleton or void in the suit?

4) SCBA Winter League 2001 v Stanley

It came apparent before the lead that we were not on the same wavelength. There had been no explanations during the auction and East confessed he was bidding spades. The lead was the knave of clubs. I won the ace and South signalled encouragement due to his paucity of high cards for the upside-down method. I led the knave of spades and had to overtake. Still on lead I was getting cocky. If North won this she would have to make a decision whether to lead into my 'main' suit hearts (I was typically 1444 but this 1-3 was possible) or trust her partner's signal at trick one. Trick three, on the top spade, South discarded. North won and returned the spade nine. I ducked and she played a club, down one – not bad really. Still 11 IMPs away though.

One thing about the constant changing of the system. It is inevitable.

If you play a natural system in partnership your post mortems will be like this.

'Partner, after we agreed hearts you bid clubs again; I think you should have a better suit than AJxxx to stress them'.

Stress, emphasis, negative inferences, fit cards in partner's suit, second round cues. These are the vocabulary of a good natural discussion. You can do virtually none of that with an artificial system.

You can't judge better – you have to design better instead.

That means better siding or re-ordering the sequences for efficiency - so you can ask earlier or on more of the important hands. Or different types of ask – revealing pattern is a cakewalk compared to discovering honour cards and there's much more work to be done here.

Once you go digital you lose tolerance. Some years after acquiring my turntable I capitulated about getting a CD player. A like-minded friend asked me how many of my records I couldn't play. One was the answer, though a good clean fixed that. He had half a dozen unplayable CDs. The single big glitch is easier to recover from in analogue. In digital you can't remedy, there are no instructions that rectify a fault, everything means something new. Nothing says, "is it all right?".

5) Brighton Swiss Teams 2001

It was wrong at 3♥ of course. The actual sequence to show 3-4-3-3 has 2♠ over 1♠ then 3♥. As you can imagine I was very suspicious – the opponents had twelve hearts and we had bid the suit twice artificially without so much as a peep from them. I imagined partner had some 4441 but which? Eventually, after a useless mark-time 4NT, I plumped for clubs, at least I had the minor honours if it turned out to be a 5-1 fit. As you can see, spades is slightly the better slam but clubs are not so far behind, much better than 6NT.

North led a spade and I quickly decided to play simply. Obviously the only danger was a spade ruff, so I went up with the queen, nine from South, and led a club for a finesse. It held. I cashed the ace and conceded just one trick when the king didn't fall. Making, lose 2 IMPs. Only then did it start to go really sour. North held the trump king (and had ducked). North held four spades and could have given her partner a ruff.

They didn't like our auction (I wasn't so thrilled about it myself). North hadn't won the first trump and continued spades because my asking bid for the king of that suit had persuaded her that I didn't have it and South did. A nod to opponents who believed such a smooth duck was possible. They asked for a ruling. There was no intimation that any of our explanations were incorrect (they weren't) so this was slightly surprising to me. We explained the sorry auction to a director. He took notes and said he would leave the ruling with a colleague before departing. Unfortunately he was unable to do so and we got to tell the embarrassing story a second time.

The director ruled in our favour. Three hearts appeared to be a mis-bid not a mis explanation and although I had fielded this in some sense with four no-trump, it was with authorised information (void opposite alleged singleton). Also, no good had come to us from this – we had after all reached an inferior contract. Just one that was curiously hard to beat. They appealed. I was really surprised.

We gave our sequence the third time. It didn't make me feel any prouder. Everyone wanted to know what I bid 4NT for. I wish I had said "I was always going to bid 6♣, I was just trying to muddy the waters". But I didn't, I told the truth. I didn't know – I just hoped something good would come about and partner would bid a suit he had. But we all know that's not going to happen don't we.

The appeals committee ruled 'no damage'. Overturning the director's reading of no infringement. Effectively this meant that when the nine spot appeared at trick one from her partner, the opening leader should have worked it out. Surprisingly the opponents were happy with this. The director wasn't, when the appeals are published I shall read, learn and stop being surprised.

The episode does raise some interesting points. Playing this system the tempo is completely different. True, some pathways are now sufficiently reinforced that they can be conducted at pace but the halts are still there. Often the relayer will just make next-call bids until the information points are reached and only then stop to think and re-create the hand opposite. Also a whole chunk of bridge visualisation needs to cut in. Many natural auctions bid on general values and are the better for it, rarely is there enough information to picture exact hands. But that is essentially what you have to do from an early stage in the relay system – especially one such as ours where strength information comes to light relatively late.

Well it's new technology and the development pace is brisk. It is fun to play and be at the cutting edge – maybe some day everyone will bid like this. After all, who would have thought CDs would catch on. According to the advert for Linn Turntables, when you go to listen to them (take your own music to all demos by the way), see how your feet react. Start tapping? You're enjoying it – that's the only real test.

Anyway, if you're interested I know where you can get the latest stuff – on the web naturally. I understand that's the next big thing.

References

Some Other Time by June Tabor is a collection of standards, Porter, Gershwin, Cahn, etc., a fine diversion from her normal folk repertoire.

Symmetric Relays were invented in New Zealand. Stephen Burgess, Roy Kerr and Paul Marston are foremost among those accredited, either for the design or its popularisation. The Moscito System (Major Orientated Strong Club) was probably where the ideas first saw the light of day.

High Fidelity by Nick Hornby is a very good read.

High Fidelity directed by Stephen Frears, starring John Cusack, Iben Hjejle is a good film – I even forgive it moving the book's N. London setting to Chicago.

I get all my hi-fi from the Cornflake Shop, Windmill St, London but Suffolk has more good outlets now. Sadly the guy with the family car aphorism didn't make a go of it on Ipswich's Norwich Road. He was spot on though.