[mycroft]

"The most dangerous convention in Bridge". Yeah, any system where pros (at the time, he's a high level TD now) need a bid in their system for "I forgot again, partner; I wanted to play 3NT" is not something I'm willing to play.

[pescetom]

Stephen Tu, on 2021-April-09, 09:12, said:

Probably more problematic than the forget issue is giving the opps a lead directing double or not inference on a very high frequency action.

The forget issue is alread problematic enough, and not just for the partnership concerned. Any TD will assure you that 1N-2N "transfer to diamonds or maybe a balanced invite" is second only to 1M-3♣ "more or less certain 2-suiter or maybe clubs". And an artificial 2N-3N (or 1N-4N thanks to 2♠ Range Ask) is more of the same, just lower frequency.

[Cyberyeti]

Stephen Tu, on 2021-April-09, 09:12, said:

Probably more problematic than the forget issue is giving the opps a lead directing double or not inference on a very high frequency action.

I used to play a different system for about 5 years where in other auctions if you wanted to bid 3N you had to bid 3♠ and let partner bid it, we NEVER had an issue with this.

[pilowsky]

This seems to be a common structure in Bridge - Don't say what you want, bid the strain at the level EXACTLY one level below where you want it to be played and make your partner suffer the consequences of your decision.

[Zelandakh]

pilowsky, on 2021-April-09, 14:27, said:

This seems to be a common structure in Bridge - Don't say what you want, bid the strain at the level EXACTLY one level below where you want it to be played and make your partner suffer the consequences of your decision.

Bidding either one below (transfers), one above (skip bids) or the first step (relay) is nearly always more efficient than pure natural bidding from a theoretical point of view. Ideally, any time you bid a real suit it would be non-forcing with better hands using an artificial call but outside of pure codified systems it is not completely simple to achieve this.

[pilowsky]

Zelandakh, on 2021-April-09, 19:00, said:

Bidding either one below (transfers), one above (skip bids) or the first step (relay) is nearly always more efficient than pure natural bidding from a theoretical point of view. Ideally, any time you bid a real suit it would be non-forcing with better hands using an artificial call but outside of pure codified systems it is not completely simple to achieve this.

That is certainly what is taught but without any mechanistic explanation of why it is so.
One good reason for using them is that the good players do, and they will mercilessly mock if you don't.


I am not saying transfers are wrong; In fact, as I become used to it, it seems to allow me to make sense of most bidding systems.

For me, though, the interesting part of your response is "...more efficient than pure natural bidding from a theoretical point of view."


My question is: Why is it more efficient, and how much more efficient is it? What is this theory you speak of?


Amongst the arguments for transfer systems are that "the lead comes around to the strong hand" and that the "strong hand remains hidden" as two of the most important reasons.


Both of these arguments may be reasonable, but has it been tested?


If so, how valuable is it? 1%, 5%, 10%, more, less?


When I try (in a very primitive way) to see whether or not a contract is makeable from the "weak" or "strong" side, there appears to be no significant difference.


My understanding is that transfers became part of Bridge lore before computer simulation could test them as worthwhile or not.
Has computer simulation validated their utility?

[Zelandakh]

pilowsky, on 2021-April-09, 19:17, said:

My question is: Why is it more efficient, and how much more efficient is it? What is this theory you speak of?

To take your last question first, I am talking about system design theory.

On the first part, what is important is how many different auctions are available below the safe limit for a given pair of hands. Each auction represents a range of hands so the more auctions you have, the tighter your ranges becomes and the more accurate your bidding. The simple truth is that forcing calls tend to create more auctions than non-forcing ones and obviously a first step relay has many more auctions than a 3rd step natural call.

How many more depends on your safe limit and how efficiently you have the rest of your structure. In general though you can calculate it using the Fibonacci sequence. When I am designing bidding systems seriously, I tend to spend a fair amount of time working out how many sequences I have and how best to match hands to create some sort of optimisation. When I post relay structures on BBF and comment that I "did it quickly" or words to that effect, it usually means that I did not go through this optimisation process but am just posting my first idea to demonstrate the viability of the concept.

[Stephen Tu]

Siding issues for transfers is really only a very slim part of the advantage. The main thing is you get to pack more hand types into one bid, you get more sequences than if you mostly had natural NF calls available.

If you play 1nt-2H as natural, for example, you pretty much have to make it either signoff, invitational, or forcing. The signoff hands are pretty important and common, so if 2H is natural, throughout bridge history people always played this as weak NF. So now you need to put the forcing hands and invitational hands somewhere else, and might have to give up sequences such as 1nt-3H to be natural and forcing. And might not have any clean way to just invite with hearts and play 2nt without fit & without max opener.

But if you put most of the primary heart hands in 2d (other than texas xfer hands), otoh, now one bid (2d) can cater to all strength ranges, since stronger hands have another chance to bid since opener isn't passing the transfer ever. Then you get back your 1nt-2s and 1nt-3h for other purposes. The only thing you give up is to be able to play 2d to play, which players gladly did since good opps would often find 2M competition over 2d, and there is so much upside in being able to bid major suit hands accurately since game is more frequent and successfully outbidding the opps for partials is more frequent.

[pilowsky]

Thanks - very helpful.

[ThomasRush]

Stephen Tu, on 2021-April-09, 19:43, said:

[...] But if you put most of the primary heart hands in 2d (other than texas xfer hands), otoh, now one bid (2d) can cater to all strength ranges, since stronger hands have another chance to bid since opener isn't passing the transfer ever. Then you get back your 1nt-2s and 1nt-3h for other purposes. [...]

Yes -- these are the reasons for playing transfers.
In addition, if you play transfers, you can start a slam sequence at the 3-level -- very hard to do after 1N - 3!H or 1N - 3!S!

[Zelandakh]

ThomasRush, on 2021-April-11, 22:05, said:

Yes -- these are the reasons for playing transfers.
In addition, if you play transfers, you can start a slam sequence at the 3-level -- very hard to do after 1N - 3!H or 1N - 3!S!

Not so difficult:

1NT - 3♥ = 6+♥ SI
==
3♠ = decline slam try (Frivolous)
3NT = serious spade cue (of your preferred type)
4m = serious cue (of your preferred type)
--

1NT - 3♠ = 6+♠ SI
==
3NT = decline slam try (Frivolous)
4m/♥ = serious cue (of your preferred type)
--

This can easily played in conjunction with standard transfers.

[pescetom]

pilowsky, on 2021-April-09, 19:17, said:

If so, how valuable is it? 1%, 5%, 10%, more, less?


When I try (in a very primitive way) to see whether or not a contract is makeable from the "weak" or "strong" side, there appears to be no significant difference.


My understanding is that transfers became part of Bridge lore before computer simulation could test them as worthwhile or not.
Has computer simulation validated their utility?

I just ran a primitive dealer script to see if a conventional major suit transfer over 1NT gains tricks or not.
Discussion about the actual script in that thread and not here, please.
But the tentative conclusion is that this transfer is gaining a trick (occasionally two) 5.9% of times and costing a trick 2.6%.
Which is less than one might expect, but enough to make it worthwhile for most of us I think.
Of course a Jacoby/Texas transfer is not the only occasion for a "rightsiding" gain, even in a simple mainstream system.