Three-Card Limit Raises that Promise Four-Card Support I asked my expert
panel: What are your thoughts on three-card limit raises in the following
auctions? 1 - pass - 3, 1 - pass -3 or if playing Bergen raises,
one-of-a-major - pass - 3 (artificial four-card limit raise). This is assuming
that you could also show three-card limit raises using some other method. What
do you think of bidding 3 holding AxxxxKxxxxKJx over 1?What do you think
of bidding 3 holding xxAxxKxxxxKJx over 1?
I’m a proponent for making unbalanced three-card limit raises. First we’ll
hear from the defense, i.e. those who agree with me and think that a player who
makes a direct three-card limit raise is not guilty of conspiracy to make a bad
I’ll give you a simple lead problem. You hold 432KQJ10A32432. The
auction goes by the opponents 1 - 3 - 4. Is there anyone who wouldn’t lead
the K? Now, suppose the auction went 1 - 1NT - 2 - 3 - 4? Is there anyone
who wouldn’t lead a trump? Too much information is bad for the offense and good
for the defense. It’s much easier to defend against 1 - 1NT - two-any - 3 - 4
than 1 - 3 -4.
I’ll give you a bidding problem.You hold 4AQ1065A54K543. The auction
goes 1 - Pass - 1NT. Wouldn’t you bid 2 without much thought? Now, suppose the
auction goes 1 - Pass 3? It’s scary bidding 4.
There are experts who are witnesses for the defense.
---I have no problems whatsoever making a three-card
limit raise if the hand looks right for it. Partner will find out how many
trumps I have when I put the dummy down. The only thing wrong with a three-card
limit raise is that it commits us to playing in the major -- 3NT is no longer in
the picture. I play semi-forcing 1NT response, so I will never respond 1NT with
three-card support and a singleton. With three-card support and a doubleton, it
is a judgment call. I would probably choose a limit raise on the hands you gave.
---The advantage: It makes it more difficult for the
opponents to come in; the disadvantage: if they do come in, it distorts the
bidding and makes it more difficult for partner to evaluate his offense and
defense. Both your examples have shortness in the other major -- should that be
a criterion? I recall that I've only done it one time in the last several years
(and I had a singleton). I also recall that it was done against me once or
I admit that if you make a three-card limit raise, you can’t play 3NT. I
don’t make a three-card limit raise on all hands. I need either a singleton or a
small doubleton. I need good trumps. Opposite a 1 opening bid, I don’t make a
three-card limit raise, if I have four hearts. I don’t want to miss my 4-4 heart
fit. Holding AKxKJxxxxxxxx, I bid 1NT and then bid 3 unless partner bids
2. Holding AKxxxKJxxxxxx, I bid 3. If you’re not playing two-over-one, it
makes the auction much easier to bid 3 holding AxxxxKJxxxQJx, then bidding
2 and then supporting spades.
---I promise four but could make an exception with the
hands you listed.
---If we have ways to distinguish between three and
four trumps, then why would we ever pretend to have a four trump raise with a
balanced hand and only three trump? As an illustration, I tend to accept a limit
raise with any 5-4-3-1 hand. Thus, with
Qxx, I would bid 4
, but would pass after 1
- 1NT - 2
. Put this hand opposite your
example (with or without the
KJx or Kxx
Game is quite marginal with the
J in dummy, and very poor without it. In
KJx or Kxx
Game is nearly cold with the
J in dummy, and excellent without it. As a second
example, consider this opening hand -
AQ. Opposite your example
(with or without the
J), we want to be in 3NT, which is a much better game than
. We will never get to notrump when partner pretends to hold four trumps. At
IMPs, 5-3-3-2 hands, opposite a balanced, three-card raise, should often play in
3NT. Several published studies, based on long simulations, found that bidding
3NT blindly rather than four-of-the-major, will win out substantially on such
hands. By the way, this is one reason that Puppet Stayman is such an abused and
misunderstood convention. Puppet Stayman is designed to describe the eventual
dummy, keeping declarer's shape undisclosed to the defenders. It is not supposed
to unearth a 5-3 fit. It is acceptable to pretend to hold four trumps on some
unbalanced limit raises. A hand like
Kxxx is quite suitable for the
. The singleton makes up for the trump shortage, and 3NT is no
longer a target game. Moreover, 3
is a good tactical shot, making it harder for
the defenders to enter intelligently with hands like
AQJxx. In fact, I would bid 1
with that hand, playing a club
system. One final note: Playing Bergen raises, one can happily play
one-of-a-major - 3
as a three- or four-card limit raise, with a 3
asking. I would much rather declare 4
after the auction 1
than the revealing 1
- 1NT - 2
Playing semi-forcing 1NT, there are some hands with three-card support that
you know don’t belong in 1NT. Playing forcing 1NT, you don’t have to worry about
---If your 1NT bid is forcing then I believe all
direct or Bergen raises should have four-card support. The reason is that it
makes slam bidding easier when there is no doubt about trump length, which often
is key to the hand. If your 1NT is non-forcing (up to 12 points), then I think
you can add in unbalanced limit raises (to your jump limit raise bid) such as
Jxxxx over 1
because you too often will go down in a 1NT
contract and make at least three-of-the-major. Over a two-under Bergen limit
raise you may be able to ask with the in-between step. Partner signs off with
three (then next step asks shortness) and partner starts cuebidding with a
four-card limit raise. I don't consider the example hands unbalanced (no
shortness) and would bid 1NT if non-forcing or forcing. I certainly might go
down in 1NT, but that is life. Also, this strategy tends to be much better at
IMPs because playing 1NT with eight spades rates to be wrong with the example
hands as too often I will make 90 or 120 vs. 140. So to be real detailed, you
may actually change the strategy depending on the scoring.
There are other ways to make three-card limit raises other than going
through 1NT. One way is to use a jump to 3 to show a three- or four-card limit
raise. If opener cares how many trumps responder has, he can ask with 3.
---I'd make a three-card limit raise only when
holding a side singleton. That's a hand I'd hate to hear partner pass my
semi-forcing 1NT. Ideally I'll hold three strong cards in the suit and my
singleton will be in the other major, e.g.
Holding three strong trumps is mandatory. Being on the top of limit raise
strength can make up for the lack of a fourth trump.
---I do not mind it nearly as much as my
experienced students do.
Marty Bergen of Bergen raises says it’s OK to make three-card limit raises.
---Since these are non-competitive auctions I think
that a healthy three-card limit raise, both balanced and with fewer high cards,
unbalanced, should be included with four-card limit raises; this further allows
the partnership to bid 1NT semi-forcing.
---I don't play Bergen raises. I will make a
direct three-card limit raise if I'm unbalanced, because 1NT followed by
three-of-the-major tends to be balanced, no shortness. I would bid 1NT followed
by three-of-the major on both of your example hands, even though they're very
---When playing forcing 1NT, I much prefer a limit
raise to promise four-card support. When playing a semi-forcing 1NT by a passed
hand, the limit raise must contain the three-card raises with shortness.
---I might consider a three-card limit raise with
three good cards in the suit over 1
since I play 1NT as semi-forcing, but never
after a 1
opening. I would start with 1NT and follow up with 3
---My strong feeling is to bid a forcing 1NT and then
jump in partner's major at the next opportunity to show a three-card limit
raise. It is just too important to guarantee four-card support for a direct
limit raise especially in competitive auctions. Admittedly some three-card
limit raises are better than others, but playing constructive raises, most
three-card limit raises are on the higher end of the continuum anyway. If I were
to make a direct three-card limit raise, I would have more sympathy for doing it
with a hand with a singleton, which makes the hand almost as valuable as having
---I would never do it with three spades and a
limit raise as long as I can do two things: make a forcing 1NT bid, and if
perchance partner actually jumped over that forcing 1NT bid, I could then show I
had a three-card limit raise and not just a delayed noise raise. With hearts,
and if I had only two spades, I might choose to get it up to the three-level
ASAP and make it harder for them to bid.
Now, if I’m not playing forcing 1NT with partner, then it's a whole other story.
I refuse to languish in 1NT with an eight-card major suit fit and the majority
of the high cards, and have to settle for an IMP or two loss figuring the other
table got to spades, or worse yet at matchpoints, figuring it's a near-zero. In
addition there are a lot of hands that partner will like his hand more when he
hears I have three-card support. He might even bid a minor that makes my hand so
much more attractive that I might just shoot out game instead of stopping at the
three-level--none of which might happen over a semi-forcing 1NT bid.
There's also the problem on the long end, when partner jumps, you have to be
able to let him know you had a three-card limit raise so he now knows he's
trying for slam and not just choice of games.
Recap: per your directions, there was some other way to show the three-card
limit raise, so I'll trust that. With spades, nope. With hearts and only a
doubleton spade, especially if partner opened in first seat, so LHO is an
unpassed hand, I will try to get it up there faster so he has to risk 3
---I've come around to thinking that direct
three-card limit raises are OK almost all the time. Partner usually won't care
how many trumps you have, as long as you have limit raise values of some kind.
Where you can get hurt is if opener gets slammish; then the lack of the fourth
trump can be critical. I especially like the direct three-card limit raises when
I have distribution, like
xxxx for example. This way going through
1NT first always shows real high-card values and implies a flatter hand, and if
partner passes 1NT, using semi-forcing 1NT, you'll be in an OK spot.
If 1NT followed by three-of-a-major shows a balanced hand, opener should
suggest 3NT more often.
---I'm probably the wrong person to ask, as I do
make three-card limit raises. I make them through 2NT, which I play as a limit
raise or better. It’s the only way I have to show a limit raise, whether three
or four cards. Having said that, I perhaps play single raises a little wider
range than most, ranging up to a quacky balanced ten-count with three-card
support, e.g. holding
Qxx I would raise 1
. I would probably
only raise 1
also. I would not dream of doing anything other than making
a limit raise with either of your example hands.
The Law of Total Tricks guru (Larry Cohen) doesn’t mind three-card limit
---I used to teach (as did most) that a limit raise
such as one-of-a-major – three-of-a-major should promise four-card support.
However, I have changed my tune. I think it is okay to make the bid with only
three-card support if other actions are less palatable. This is especially so if
you are playing 1NT as "semi-forcing" or "not forcing."
Now we’ll hear from the prosecution. There are at least two reasons why a
limit raise should show four trumps. If opener is interested in slam opposite a
limit raise, he’ll be very disappointed when dummy comes down with three trumps.
Another reason is Q5432 is a reasonable holding opposite four trumps. With that
holding, opener will be very disappointed when dummy comes down with three
trumps. However, if opener has strong trumps and is not interested in slam, that
won’t be a problem. Some experts use 2NT as a three-card limit raise. Knowing
for sure that responder has only three trumps solves the above two problems.
---I would never offer an uncompetitive limit
raise with only three-card support and, pursuant to your question of a few years
ago, would not be caught dead using Bergen. If using a semi-forcing 1NT, I would
agree that some three-card limit raises are going to be necessary. I see the
Italians use 2NT as a three-card limit raise and doubt they would make up a
crappy convention, so there might be merit there. However, any jump raise method
I use will always be defined as to how many trumps I have (exception above
---You need to state whether a forcing or
semi-forcing 1NT is in place, since it makes a big difference. Note that BART
(an artificial forcing two-diamond bid used in this sequence: 1
- 1NT - 2
) or the equivalent matters too. In either case, if forcing 1NT is in play, I
would not make the limit raise--especially not if I’m playing Bergen raises with
---If I have other methods to show a three-card
limit raise, I would not make a limit raise with this. Sometimes knowing partner
has four-card support makes a big difference, especially when opener is light on
high card points but big on distribution.
---I avoid a three-card limit raise in those
sequences you presented. The fourth trump is too important. If I do it, it's
because I don't want to bid a weak suit, which is misleading if slam is in the
air. Also, I will have a maximum with three good trumps.
---I think it is useful to distinguish three- and
four-card raises for partner's evaluation, so would not jump to the three-level
and make what is normally a four-card raise on the given hands. Note that, over
a limit raise, partner will usually bid game with a fairly minimum hand with a
singleton (minor suit card), and that will not fit well with the given hands.
I'd be more inclined to do it on a hand like
Jxxx where partner's
minor suit singleton is worth more, and it is good to get the support
in immediately, though it might lead to us bidding 5
over a 4
bid by the
opponents when we belong in 5
---I think it is very important to distinguish three-
and four-card raises because knowing the minimum number of trumps is
important. The difference between eight or nine is huge. I suppose you really
mean in the question that you do not have other methods to show three-card limit
raises, because I would not show four when I have only three cards in partner’s
suit. It would create havoc in bidding games, and the winner and loser count is
off by one. It is also rare that after an initial 1NT bid by responder, assuming
that it could be passed, that a game would be missed, although we can always
create the fitting hands to make the three-card limit raise effective; but it is
a guess at that point. Besides, the hands partner will pass will be 12-13 point
flat hands and game is unlikely.
---Since I use forcing 1NT, I don’t need to jump to
the three-level to make a three-card limit raise. In your sequences, the fourth
trump is frequently crucial. Also, on some hands, I would bid a game over a
minor response to 1NT by opener. As for Bergen raises, I like 3
for the limit
raise and always with four or more cards in opener’s major.
---The presence of a fourth trump in responder's hand
can make a very significant difference in opener's hand evaluation. I would not
dream of using a limit raise as showing three or more trumps. Eddie
Wold---Never! I think it is much too important to know about the fourth trump.
---Don't know, never tried it. It has some definite
advantages--mostly of a pre-emptive nature. I believe in identifying a fit as
soon as possible after its discovery, and it would satisfy that criterion. It's
hard for the opponents to come into the auction over 1
. By the same token,
it may be difficult to show your heart fit later if LHO has a pre-emptive hand.
Philosophically, it does not fit the Bergen mold. Bergen is heavily invested in
“The Law of Total Tricks” and determining the number of trump each side
holds. That can be critical in a competitive auction--or even in uncontested
auctions that involve the decision to try for game or slam. Max Hardy used to
approve of three-card raises if responder had a singleton. There is also a
treatment, which I have occasionally played, wherein when partner opens a major
and the opponent overcalls at the two-level, a 2NT response shows specifically a
three-card limit raise. But there is no system that I've played that calls for
systemic three-card limit raises. I think, if I played it, I would like a
guarantee that the 2NT bidder does not have four of the other major.
---For the past five years I have played in an
occasional partnership with P.O. Sundelin. P.O. has some great ideas about
bidding theory, one of which is that responder bids 2NT with any three+ -card
raise in a (major) suit with limit raise values or greater. Opener can then ask
for length and strength. In this way, when opener signs off in three-of-a-major
or simply bids four-of-the-major, the defense is left in the dark as to dummy's
trump length and its strength. Essentially, they don't know whether to attack or
to go passive with a trump lead. I love the method, but also play in other
partnerships, which delineate three- and four-card trump length, as well as
strength, making it easier on the defense. So, no, I am not now a big fan of the
definition you suggest. Nevertheless, I play 1NT forcing one round with some
partners, and three-of-a-major after that shows the three-card limit raise, an
immediate 2NT shows the strong four-card raise and three-of-the-major or
three-of-a-minor shows the four-card limit raise.
---Never, it’s too important to differentiate
between eight- or nine-card fits, especially in competitive situations, and
considering it’s so easy to show the difference even in most competitive
---NEVER under any circumstances. I feel strongly
about this one, since I will always have alternative ways of showing
an equivalent strength three-card raise.
---In my curmudgeonly opinion, "three-card limit
raise" is an oxymoron.
---If it is within a partnership's parameters to
make jump limit raises with three-card support, then the examples given fit that
criterion nicely. I am, however, a believer in “The Law of Total Tricks,” and
feel that knowing that one's combined total of trumps adds up to nine+ is
extremely important for determining whether one has a game, or whether it is
right to bid higher when the opponents interfere.
---My primary rule is that there are no
"absolutes", but this one is close to absolute -- these raises require four+
trumps. Sometimes, partner will be counting tricks, and lying about a trump in
these auctions, or a Jacoby auction, is asking for trouble. I would NEVER make
a limit raise without a singleton, and even then, it takes a very rare hand with
no other palatable choice. Your examples don't come close -- you either hope
that a semi-forcing 1NT bid isn't passed (and if it is, you likely have enough
power to survive it), or force to game.
---I’m much more comfortable
with one-of-a-major – three-of-a-major promising four-card support and using 3
for other purposes, usually natural and non-forcing invitational. “The
Law of Total Tricks” seems to work for me, especially when part scores are
involved. With three-card support, bidding 1NT forcing works well for me because
partner’s rebid can help me evaluate what level to play at and partner will
realize that. For example, on the two hands you gave, a 2
bid would get a
four-of-a-major bid from me. I’m not a great theorist, but this seems to work
The Law of Total Tricks is used only in competitive auctions. You can play
7 in a 4-4 spade fit. If it goes 1 - 3, the Law is not involved.
---I am 100% against it. Often the crux of the
hand will be the fourth trump. I have a rule that I always accept a limit raise
with a singleton (as opener). This would be ridiculous if I could not count on
four or more-card support. I also, similarly believe that you cannot bid Jacoby
2NT with only three-card (even the AKQ) in support.
All the facts have been presented and it’s time for you the bridge playing
jury to make your decision.When you hold AKxxxKJxxxxxx and bid 3 over
partner’s 1 opener, you’re voting not guilty.If you go through 1NT first,
you’re voting guilty. I know it’s illegal to tamper with the jury; however, I
strongly suggest that you vote not guilty. You’ll probably get rewarded with