Sandwich NT

[sceptic]

Example 1

 

P - 1 C - P - 1S

1NT ............................is this sandwich NT and if so what range and distribution do you play it with or what is standard

 

Example 2

 

1C - P - 1H - 1NT......... is this sandwich NT and if so what range and distribution do you play it with or what is standard

 

also what difference should there be between examples 1&2

 

 

Example 3

 

1C - P - 1H - X ........ if you use sandwich NT what would you ultimately play this as

 

Example 4

 

P - 1C - P - 1S

X ............................if you use sandwich NT what would you ultimately play this as

 

also what difference should there be between examples 3&4

 

and would you build all 4 instances into a partnership agreement?

 

also is there better ways to relay info about the unbid suits in the sequences ?

[pbleighton]

1 and 2 are sandwich NT, the range is anything less than a takeout double. Range and shape are really what you are comfortable with. I will do it with 4432 shape and a yarborough at unfavorable, but....

 

3 is a takeout double.

 

I have no bid for 4, though I'm sure you could come up with one, such as 5-4 with the 5 in the higher (or lower) raking suit.

 

I've been playing sandwich NT for about a year now, I think it's great.

 

Peter

[firmit]

Me and my partner have just agreed on Sandwich.

(1C)-p-(1H)- ?

X = exactly 4 spades with 4+ unbid minor or strong one-suiter, or strong balanced

1NT = 5spades + 4+ of unbid minor ( may even be bid if 1C promise 2+)

[awm]

In the first sequence, 1NT is unusual for the unbid suits. You could also call it sandwich notrump by a passed hand. This meaning is extremely standard, since there is no sensible natural meaning for the 1NT call (yes I know some people play it as 9-11 balanced, I think they're crazy myself). I would definitely assume "unbid suits" as the meaning for this call if undiscussed with a good partner.

 

The second sequence is the typical "conventional" sandwich 1NT showing the unbid suits. There seems to be a lot of controversy over whether this is a good convention or whether it would be better to play 1NT as natural in this auction. Obviously this is something to discuss with partner.

 

In the third sequence, double is takeout. The difference between this and the sandwich 1NT (if you play that convention) is that the double shows a better hand. Typically double is at least opening strength and usually 4-4 in the unbids; double could also be a "power" double with a very strong one-suiter or a slightly off-shape double (say 4-3 in the unbids) with a strong hand. This call suggests defensive values and partner is encouraged to penalize in some auctions when the opponents bid on without a great fit. The sandwich 1NT would show a weaker hand with more shape (at least 4-5 in the unbids and less than opening values).

 

In the fourth sequence (passed hand), a popular treatment is that double shows a preference for the higher suit (typically 5-4 in the unbids) whereas 1NT shows a preference for the lower suit (typically 4-5). Both generally show around 8-11 hcp (near maximum for a passed hand, the lower end of the range is often 5-5 or 4-6 respectively for the two calls).

[Free]

Don't we have enough of these threads yet? :P

[bid_em_up]

Example 1

 

P - 1 C - P - 1S

1NT ............................is this sandwich NT and if so what range and distribution do you play it with or what is standard

 

Example 2

 

1C - P - 1H - 1NT......... is this sandwich NT and if so what range and distribution do you play it with or what is standard

 

also what difference should there be between examples 1&2

 

 

Example 3

 

1C - P - 1H - X ........ if you use sandwich NT what would you ultimately play this as

 

Example 4

 

P - 1C - P - 1S

X ............................if you use sandwich NT what would you ultimately play this as

 

also what difference should there be between examples 3&4

 

and would you build all 4 instances into a partnership agreement?

 

also is there better ways to relay info about the unbid suits in the sequences ?

Case 1) definitely sandwich NT. You are already a passed hand. It should always be sandwich in this case.

 

Case 2) is only sandwich by agreement. It is possible for you to still hold the 15-17 NT hand in this case, and some people like to play it as such.

 

Sandwich NT usually is 5+/5+ in the other two suits, with weakish values. However, you can do it as in example two when holding extra values where you will still have options to get partner to choose a suit later. Alternatively, you could define 1N and 2N in such a way to show a weak 5-5 and a strongish 5-5, if you so desired.

 

Case 3) is takeout for the other two suits, usually 4-4 or 5-4 and 10+ hcp as you are not a passed hand.

 

Case 4) is same as case three, but limited by the fact you are a passed hand. Range is approximately 9-12. The HCP range needs to be a little stronger here than in the cases of a 1N overcall, because you do not have the "extra" distributional compensation that comes from being 5+/5+.

 

Yes, you should build (or at least discuss) all 4 sequences in any serious partnership.

 

There may be better ways to incorporate info regarding unbid suits, but I have yet to find them.

 

The main downside to using sandwich NT overcalls is you give up the natural 1N overcall showing 15-17 by a non-passed hand. I don't find this to be much of a problem, but other people seem to have difficulties with this.

[jtfanclub]

I have been considering using sandwich 1NT to show the unbid major and the BID minor.

 

Example: 1C-P-1H-1NT as spades (4) and clubs (5+).

 

The logic is obvious to me...there's no reason that I can't have a long club suit (I don't think opener having 3+ clubs is a big negative indicator of that), and I'll never be able to show them later if I don't show them now. We already have a perfectly good bid for spades and diamonds, and if I bid spades now and diamonds later it should be clear what's going on.

 

Does anybody on the planet play this? I'm sorry if I've asked before.

[Codo]

I think Sandwich NT is better to show different length not different strength compared to a take out double. I play it to show 4/5 in the unbid suits without a 5card major. I believe that this imporoves competitive bidding much more then knowing whether you are 5-10 or 11-15 (or whatever range you like).

 

So for your 4 cases:

1. 4 Hearts, 5 Diamonds

2. 4 Spade 5 Diamonds

3. 4 Spade 4 Diamonds

4. 4 Heart 4 Diamonds

 

If I would believe that there are better tools, I would use them. This way works quite well for me and is easy to rememberand to handle.

[sceptic]

Don't we have enough of these threads yet? 

 

if I thought so I would not have posted it :)

 

as bid em up has cleared up or at least made me aware of hand 1 & 2 which is where my main query was intended

 

and codos answer gives food for thought

[MFA]

By an unpassed hand the sandwich NT should be natural. The traditional assumption that opener always holds 12+ and responder always 6+, and we thus don't have a game, is too far off in modern bridge.

 

1NT in sandwich by a passed hand is of course unusual. In my system, 1NT is then a shape take-out while D is more balanced. Other distinctions are possible here.

 

@pbleighton

1 and 2 are sandwich NT, the range is anything less than a takeout double. Range and shape are really what you are comfortable with. I will do it with 4432 shape and a yarborough at unfavorable, but....

 

LOL, good luck with that :P

[jdonn]

By an unpassed hand the sandwich NT should be natural. The traditional assumption that opener always holds 12+ and responder always 6+, and we thus don't have a game, is too far off in modern bridge.

 

1NT in sandwich by a passed hand is of course unusual. In my system, 1NT is then a shape take-out while D is more balanced. Other distinctions are possible here.

 

@pbleighton

1 and 2 are sandwich NT, the range is anything less than a takeout double. Range and shape are really what you are comfortable with. I will do it with 4432 shape and a yarborough at unfavorable, but....

 

LOL, good luck with that :)

Ditto 100%, including sarcastic remark at end.

[1eyedjack]

By an unpassed hand the sandwich NT should be natural. The traditional assumption that opener always holds 12+ and responder always 6+, and we thus don't have a game, is too far off in modern bridge.

Your assumption that once in a blue moon we might have game on because both opponents might be subminimum for their bidding may be valid. I don't see that as sufficient justification for 1NT to be natural. Contrast that frequency with partner having a bust on a misfitting hand and your optimistic scenario is one that I for one would be happy systemically to forego.

[Tola18]

I have been considering using sandwich 1NT to show the unbid major and the BID minor.

 

Example: 1C-P-1H-1NT as spades (4) and clubs (5+).

 

The logic is obvious to me...there's no reason that I can't have a long club suit (I don't think opener having 3+ clubs is a big negative indicator of that), and I'll never be able to show them later if I don't show them now. We already have a perfectly good bid for spades and diamonds, and if I bid spades now and diamonds later it should be clear what's going on.

 

Does anybody on the planet play this?

I think this is an interesting idea, especially playing against systems when the minor may be less then 4 cards - and frequently is.

 

I for one will think about it.

[1eyedjack]

I have been considering using sandwich 1NT to show the unbid major and the BID minor.

 

Example: 1C-P-1H-1NT as spades (4) and clubs (5+).

 

The logic is obvious to me...there's no reason that I can't have a long club suit (I don't think opener having 3+ clubs is a big negative indicator of that), and I'll never be able to show them later if I don't show them now. We already have a perfectly good bid for spades and diamonds, and if I bid spades now and diamonds later it should be clear what's going on.

 

Does anybody on the planet play this? I'm sorry if I've asked before.

For quite a long time I have been playing something similar:

(1C)-P-(1?)-X = 3 suited takeout of Clubs

(1C)-P-(1?)-1N = 3 suited takeout of ?

(1C)-P-(1?)-2C = 2 suited takeout of C and ?

 

Not only does this capture the occasions when 1C is short and we have a fit there, but we also get to find our fit in "?" where responder only had 4 in them.

[jdonn]

By an unpassed hand the sandwich NT should be natural. The traditional assumption that opener always holds 12+ and responder always 6+, and we thus don't have a game, is too far off in modern bridge.

Your assumption that once in a blue moon we might have game on because both opponents might be subminimum for their bidding may be valid. I don't see that as sufficient justification for 1NT to be natural. Contrast that frequency with partner having a bust on a misfitting hand and your optimistic scenario is one that I for one would be happy systemically to forego.

You refer to it as someone else's assumption when he never said "once in a blue moon" and also never said the only goal is to reach game. You are putting words in the mouth of the other position to try and strengthen yours.

 

Aside from the (far greater than once in a blue moon) likelihood of a game, it is important to have 1NT available for the partscore battle. When you are that strong there is a real likelihood the hand belongs to your side. The sandwich notrump tells the opponents all about your shape when you are so weak that you are likely to end up on defense, and all on a hand that had other available bids (double, 2NT) to show the other suits. Unlike other two suited bids that may be weak, it is KNOWN to be weak yet has no preemptive value. Hardly a recipe for success.

 

Your point about the danger is valid, and I usually bump up the range for this 1NT to something like 16-19 or so.

[1eyedjack]

By an unpassed hand the sandwich NT should be natural. The traditional assumption that opener always holds 12+ and responder always 6+, and we thus don't have a game, is too far off in modern bridge.

Your assumption that once in a blue moon we might have game on because both opponents might be subminimum for their bidding may be valid. I don't see that as sufficient justification for 1NT to be natural. Contrast that frequency with partner having a bust on a misfitting hand and your optimistic scenario is one that I for one would be happy systemically to forego.

You refer to it as someone else's assumption when he never said "once in a blue moon" and also never said the only goal is to reach game. You are putting words in the mouth of the other position to try and strengthen yours.

 

Aside from the (far greater than once in a blue moon) likelihood of a game, it is important to have 1NT available for the partscore battle. When you are that strong there is a real likelihood the hand belongs to your side. The sandwich notrump tells the opponents all about your shape when you are so weak that you are likely to end up on defense, and all on a hand that had other available bids (double, 2NT) to show the other suits. Unlike other two suited bids that may be weak, it is KNOWN to be weak yet has no preemptive value. Hardly a recipe for success.

 

Your point about the danger is valid, and I usually bump up the range for this 1NT to something like 16-19 or so.

Now whose putting words in whose mouth? When I sandwich it is not "KNOWN to be weak". Also it is clear from my post that it is my opinion that the likelihood of game is that small, just is it IS in fact clear from the OP that I was quoting that it was the prospect of missing game that drives his argument for a natural 1N. The introduction of the partscore battle may or may not be a valid point, but it is exclusive to your post.

 

Even when I am weak, whether we are likely to win the battle to declare will largely be a function of partner. It is by no means certain that we will lose out on that purely because I personally, of our partnership, might lack the strength.

 

It is also far from clear to me that just because we have the balance of the values in the partscore battle (when that happens) that we are better off declaring than defending.

 

It is also far from clear to me that on those occasions that we do have game on the opponents will obtain a superior score by playing in their ultimately doubled partscore.

 

As regards what is the actual frequency of our having game on, I would be interested in some hard stats, if anyone out there has the resources for it.

[1eyedjack]

Couple of additional points.

 

I would be interested to know what proportion of the top players play a sandwith 1N as natural by a non-passed hand. Not sure how anyone would know the answer to that, but those who move in those circles may have a feel for it.

 

If you are going to play it as natural, and for fear of punitive action decide to keep it up to strength, say 16+, then you are going to cut out a significant proportion of the hands when you have game on (perhaps the majority), where the values are more evenly divided between your hand and that of partner. Perhaps nothing to be done about that.

 

As to contesting the partscore, even if I am weak and playing the classical 2-suited sandwich, I would expect to have reasonable prospects of winning the partscore battle after

(1C)-P-(1D)-1N

when I might have less anticipation after

(1H)-P-(1S)-1N

Perhaps if you are concerned about being outbid, you might want to distinguish between these (and other) situations when deciding on the merits of the balanced alternative.

[1eyedjack]

I had another thought:

 

It has long been my opinion that the propensity to make increasingly light opening bids (in 1st or 2nd seat) opposite a partner who makes increasingly light responses is a strategy destined to end in tears. I am not talking about systems specifically designed to make light openers, such as Moscito, which typically place a corresponding lower ceiling on the upper limit assigned to the opener.

 

Occasionally you may successfully talk the opponents out of game, and when both partners are minimum or subminimum for the bidding you will not overreach. And of course you have entered the race at a sprint in the partscore battle.

 

But as against that you will overreach by going down in high level partscores when game tries are refused (or punted games find a subminimum partner).

 

The increasing instances of this policy suggests empirically that it may on balance be the winning technique. But I am unconvinced of this, and I also suspect that its prevalence is "talked up". If it is losing strategy then there remains the question, in defending against such methods, how should you (if at all) adjust your own methods. Should you design your methods to cater for their both being rock bottom minimum? Or should you accept that on occasion their system works against you, secure in the knowledge that in the long term they will lose more than they gain? Not obvious, to me. There is a risk, in my view, that if you adjust your methods to cater for their best possible scenario, then you are playing into their hands and actively increasing the success rate of their methods.

[jdonn]

As I understand it, the convention is specifically for hands lighter than a double would be. If yours is not known to be weak, on what hands are you using it? I wouldn't mind some stronger alternative so much, but that such hands can already be shown with a double.

 

I agree, there is not much to be done when the cards are something like 12 and 13 and no player has convenient distribution to enter. Life stinks that way. It is simply a matter of deciding on the compromise between safety and enterprise.

 

I still disagree with what you are assuming the original poster meant. All he said was people bid with lighter hands these days than they used to. This is not (exclusively) an argument that we will likely have game, but just an argument that we will likely own the hand.

 

When you say on a given hand maybe we should defend when we have the balance of values, that is true but moot. On a given hand we might do better by passing out in 4th seat with 16 high card points also. The fact is when you have more values than the opponents, it is more likely than otherwise that you want to play the hand rather than defend it at a low level.

 

I don't know how you are doubling all these partscores the opponents will bid on these auctions. Are you doubling on balanced 17 and no particular length in any suit when it goes 1♣ p 1♥ p 1NT p p ?

[Guest Jlall]

♥

[mikeh]

I am late to the thread, because I thought this topic had been exhaustively canvassed not too long ago, but apparently not all recall that, or read it.

 

I am a big believer in the natural, strong 1N. I completely agree with those who argue that this wins because:

 

1. We may have game, despite their bidding. Certainly, against me and (I think) most experts, the auction 1minor [p] 1major to you could be your LHO opening on a shapely 11 and rho responding on a 3 or 4 count.

 

While some (leveljack, for one) think that this light opening, light response is asking for trouble, I disagree. 30 years ago, much expert bridge was devoted to disaster avoidance, so this kind of bidding rarely happened. Modern expert bridge is based, to a great extent, on disaster-creation: to getting in the face of the opps and creating opportunity for error. This is as a result of the large number of players who are very good constructive bidders... compared to the far lower number of players who have correspondingly good competitive methods/judgement.

 

2. We may well have a good partscore: partner's pass of the opening does not deny a 5 or 6 card suit, and he can introduce that suit comfortably after we show a balanced good hand

 

3. The sandwich NT is a duplication, to some degree, of the takeout double and, unless you play all cues natural, the cue. Many experts now play that the cue of either opp suit is natural, but I still play that the cue of opener's suit is takeout and the cue of rho's suit is natural. Furthermore, we have 2N available for extreme shape. And, as has been pointed out earlier by others, the use of the sandwich 1N for hands too weak or otherwise unsuited for other action more often helps the opps than it harms them.

[1eyedjack]

As I understand it, the convention is specifically for hands lighter than a double would be. If yours is not known to be weak, on what hands are you using it? I wouldn't mind some stronger alternative so much, but that such hands can already be shown with a double.

There is nothing in my opinions that is new or that I have made up myself. They are straight out of Payne and Amsbury's "TNT and Competitive Bidding" (ISBN 0 7134 25423 or 0 7134 25431). Rather a dated publication now, and perhaps predates the currently perceived propensity to light openers and light responders. As I said in an earlier post I prefer to distinguish double and 1N purely on shape, not on strength. The Law of Total Tricks has been largely discredited now, but it still has a modicum of validity. All that you say may be correct if your underlying system that relies on a very weak 1N intervention is the norm (when 2-suited), but you have yet to address the diversity of the cases where opponents have shown both minors v both majors, and you have not addressed the possibility that partner may have strength + fit to compensate for your weakness.

I agree, there is not much to be done when the cards are something like 12 and 13 and no player has convenient distribution to enter. Life stinks that way. It is simply a matter of deciding on the compromise between safety and enterprise.

There is I think also to be considered a general principle that the partner who has the shape has some obligation to act upon it, because his partner is assuredly "fixed" if you do not. He could be really quite strong, but be unable to act.

I still disagree with what you are assuming the original poster meant. All he said was people bid with lighter hands these days than they used to. This is not (exclusively) an argument that we will likely have game, but just an argument that we will likely own the hand.

Ok, let us agree to differ. I simply quote from his post, with emphasis added by me, "The traditional assumption that opener always holds 12+ and responder always 6+, ***and we thus don't have a game***, is too far off in modern bridge." The assumption may be off. Whether you should adjust your strategy to compensate is yet to be proven to my satisfaction. It may be so, but for the moment I think otherwise.

When you say on a given hand maybe we should defend when we have the balance of values, that is true but moot. On a given hand we might do better by passing out in 4th seat with 16 high card points also. The fact is when you have more values than the opponents, it is more likely than otherwise that you want to play the hand rather than defend it at a low level.

If it is true, why should it be moot? By passing out the hand in 4th seat you are comparing your expected plus score (from bidding) with zero (from passing it out). In our original situation you are comparing our expected plus score (from making a partscore) with our expected plus score (from taking a penalty). The analogies do not match. It may be that MP v IMP scoring makes a difference to the strategy here. I can certainly envisage that the argument for a natural 1NT is strengthened in a MP arena.

 

I have always been taught that whether you should choose to defend or declare, where passing it out for zero is not an option, should be predicated primarily on whether or not you are likely to have a fit, how big that fit is likely to be, and at what level you will be required to commit to buy it. If the hand is so misfitting that I prefer NT then if I have the balance of values I am not particularly concerned whether I am declaring NT or defending NT, subject to the caveat below.

I don't know how you are doubling all these partscores the opponents will bid on these auctions. Are you doubling on balanced 17 and no particular length in any suit when it goes 1♣ p 1♥ p 1NT p p ?

On this occasion I probably pass them out. If I happen to have game then I get stuffed. Am I going to change my system for that low frequency, instead risking coming in with 1NT? I don't think so. One factor I take into account: If I bid 1NT and it is right for the opponents to double, then they are in a position to do so with a considerable degree of accuracy.

[awm]

The discussion about the merits of sandwich NT was covered pretty extensively in a recent thread. Since people still seem to want to talk about it, I'll reiterate some of my points from before in favor of the convention.

 

(1) While it's true that people are opening/responding lighter and lighter, this doesn't mean that they usually have nothing for an opening and response. In fact the odds that our side has 24+ hcp are still pretty low even assuming they open all tens and respond all hands.

 

(2) In fact the light opening/response style often backfires when the opening side has no real fit. The issue is that while occasionally you "steal a game" when you have 14 hcp combined and no fit, much more often someone has a little bit of values and you end up playing 1NT with 17 hcp combined and no fit. Quite frequently this is worse than just passing and letting the other side bid their partial. The reason good players still bid/respond very light has to do with the fact that you have a fit on something like 85% of all hands (even more if responder has shape, and most people aren't actually responding on balanced zero-counts so much as shapely zero-counts). It's not clear that letting the opponents play the hand when they open/respond and have no real fit and less than half the high card points is necessarily a bad result.

 

(3) When fourth seat has the wrong shape for a takeout double, it tends to increase the odds that the opponents have flattish hands and/or no real fit. Note that most people don't open balanced ten counts, it's the shapely ten counts where they bid. The more cards I have in opener's minor the more likely he is flat. And if the opponents have a fit in responder's suit, often fourth hand will be short in that suit and be able to manage a double. The very fact that fourth hand has the right shape for a 1NT bid and the wrong shape for a double tends to imply that opponents being very light on fitted hands is less likely than it would otherwise be.

 

(4) In addition to increasing the odds that we have game (slightly) by opening/responding very light, the opponents also increase the odds that we want to compete for the partial (substantially) by opening/responding very light. Intermediate hands (say 8-11 high) with the unbid suits are going to be more frequent than strong (say 16-19 high) hands without great support for the unbids. If our goal is to compete for the partial we want to get in on as many hands as possible.

 

(5) An advantage to using sandwich is that it keeps the minimum values up for the takeout double. This sometimes allows us to penalize the opponents. Of course, the strong notrump also allows us to penalize, but here opponents have an easy out (just pass the 1NT) when they are bidding on garbage, and the value-showing double is going to be a lot more frequent than the big notrump.

 

(6) In any case, sandwich notrump is more appealing at MP scoring than IMP. I've won a ridiculous number of matchpoints by bidding sandwich 1NT on hands where other people couldn't get into the auction. Of course, at MPs the frequency of bids is extremely important. Perhaps a part of my contrasting philosophy with some of the best players on these forums has to do with a preference for MP scoring (as opposed to the apparently prevalent preference for IMPs).

[awm]

Here's another interesting thought.

 

If the auction goes 1X - Pass - 1Y, it'd be nice to be able to get in with an offensive-oriented hand with moderate values and support for both other suits. It would also be nice to be able to announce a strong hand with defensive values for a number of reasons, including avoiding missing games, penalizing opponents, and competing when partner is weak but with a long suit. Suppose we'd like to allocate 1NT and Double to these two hand types.

 

It makes a lot more sense for Double to show the strong hand. With the strong hand you will occasionally want to defend doubled, and starting with a double tends to pressure the opponents any time they don't have a fit in responder's Y suit. Now they are scrambling, partner knows you have a hand, they could be doubled anywhere and everywhere. Also, there exist some strong hands that are unbalanced, and people come in a lot more aggressively in this auction (especially at the one-level) than they used to. These days 1♣-P-1♥-1♠ doesn't show much more than a 1♠ direct overcall (which doesn't show much at all). So we might want to distniguish a "big hand with spades" from "garbage with spades" by using the initial double. Of course you could use 1NT for this, but presumably 1NT if a "strong hand" is supposed to be natural. In addition, using double on the strong hand may give us some safety (in an abstract sort of way) because most people use redouble as support. This makes it somewhat more difficult to penalize if we've stepped into an auction where we have no fit and the opponents have the values -- if we overcall a natural 1NT then either opponent can double to show values, whereas over a value-showing double it's not clear that either opponent can actually redouble to show a good hand (and even if they have that agreement they may not want to play 1Y redoubled).

 

So it seems reasonable to define 1NT as a "takeout double" and X as "I have a good hand." Now the question comes up: how strong is the double, and what kinds of hand can it be? It seems like with length in responder's suit Y it is relatively safe to pass. If the auction continues with 1NT from opener and two passes, double should show a "good balanced/semi-balanced hand with length in suit Y." If opener rebids 2X it's reasonable for double in balancing seat to be "takeout of X, normally with support for Y and good values." So perhaps the double is best defined as something like "general values; will not have substantial length in suit Y; typically 3/3 or better in the unbids or a strong one-suiter; frequently has some length in opener's suit X as well."

 

This looks a lot like "sandwich notrump and a takeout double" except that the range for the 1NT call is a bit wider (includes minimum opening hands), the takeout double is stronger (say 15+), and the takeout double need be only 3/3 in the unbids. The guarantee of 3/3 is nice because it lets partner run with a weak hand and four-card suit with some guarantee of safety (in fact he can also run to opener's suit in some situations). In some ways this is similar to playing "overcall structure 1NT/Double" except that the shape for both double and 1NT is better defined.

[blackshoe]

Tell you a story about Sandwich NT. Some years ago, playing with my then regular novice partner, we arrived at a table where the opponents were discussing this convention (it had apparently come up in the previous round, and they'd had a miscommunication). Second board of the round: 1C on my right, I pass, 1H on my left. Partner gets this look on her face, and bids 1NT. I alerted. RHO asked. I said "we haven't discussed it or agreed to play it, but given your previous discussion of Sandwich NT and the look on partner's face, I'd say she has the other two suits". They won the auction. Turned out partner had a 4=2=4=3 13 count, and yeah, she figured this "Sandwich NT" gadget was just perfect for the hand. :)