Source: 
New South Wales Bridge Association By Steve Hurley
Normal bidding using say Standard or Acol is an exchange of information to show and deny things. If you open 1
and partner bids 1
, by bidding 4 card suits up the line, partner has (initially) denied 4 Diamonds or 4 Hearts. I say initially because after you rebid say 2, partner may now bid 
or 
, showing a 4 card suit. And this now means that Spades are longer. The bidding then continues, to show and deny.
Relays take the exchange of information and turns it on its head. One person asks: the other person shows, all the way through the bidding. So the asker will take in the information, assess that looking at the two hands, then placing the contract. The asker forces continuation of the bidding by making the cheapest bid each time, then partner responds.
Such systems are usually big clubs (where opening 1 is 16+ HCPs, like Precision). If partner opens 1, responder bids 1 to show less than game going values (up to 9 HCPs), and all other bids show game forcing values (so the asker knows that at least game is on and can continue bidding to that level). So after 1:1, opener then rebids the cheapest bid (1, and responder describes their hand again.
I mentioned that the responder shows their hand, but not that 1 specifically shows Hearts. It could do, depending on the version of system you want to use. But many partnerships agree that this shows Spades. Why? So that when the asker asks again with 1, if Spades is the contract, the asker has not shown their hand but is instead playing the declarer.
Anyway, the principle of relay bidding is for one person to show suits first, then length/shortage, then number of Aces and Kings, then where these Aces and Kings are, and for slam bidding, even where the Queens and Jacks are. A major drawback with using relays is remembering all the steps. The original system drained people with their intensity. Then around the 70’s symmetric relay was devised so that whatever path you take, you arrive at a common point to show shapes, so you only need to remember a few structures.
Back to the problem at hand. How could symmetric relay deal with this hand:
| AKT9 AK65 AKQ74 |
J954 Q853 T85 65 |
The bidding would proceed:
1 (16+HCPs) 1 (less than 9 HCPs)
1 (cheapest bid) 1 (less than 5 HCPs)
2 (GF, BIG) 2 (balanced)
2NT (cheapest bid) 3 (two suits, same rank)
3 (cheapest bid) 3NT (doubleton , so 4432)
4 (cheapest bid) 4 (No king, but 1+ Queen)
4 (cheapest bid) 5 (Denial cue bids, denies Q, shows Q)
6 (to play) Pass
So the asker knows the exact shape, strength of the hand, and key cards. Consider also that at the point where opener can bid 4, asker knew the exact shape and strength. Now if responder had bid 4 showing not even a Queen, asker could not be able to bid 4 to play (as this would have been a relay and continuing the bidding). But asker was ready for this and is also happy to play in 4 instead and this would have been the final contract (as bidding 4 is NOT the next cheapest step). In other words the asker has to be careful about what the possible responses will be when getting towards what the final contract should be.
This looks pretty good, so why do many people not play it? It is memory intensive, takes time to bid that regular club events do not cater for, and a lot of work needs to go in to discuss the usual sequences as well as where there are interruptions (opposition bidding). And often the mere mention of symmetric relay at the table is like a red flag to a bull and the opposition interrupt by bidding aggressively.
But if you put the effort in, it is worth it.