Why is matchpoints different from IMPs

[inquiry]

I propose this thread to highlight the differences between duplicate bridge when played at Matchpoints (MP) versus IMPs. To play on line (unless you play only at at total point table), you must learn the basic differences between matchpoint and imp scoring. Ideally, hands will be shown in it that highlight how you would do something differently at Matchpoints than you would at imps (bidding or play) and explain why.

 

As an introduction to the topic, the difference between the MP and IMPs affect both bidding and play. At imps, the strategy can be summarized pretty simply as stretch to bid any vulnerable game that has almost any chance to make, and (almost) always play to make your contract, taking the safest line of play available (at imps doubled, you might decide to play safely for down one rahter than risk down four or five, but that is an exception). This "stretch" to bid vulnerable games means you will bid games with less than 50% to make. How much less? Roughly 1 in 3. This is related to the scoring table at imps. A vul game (lets say 4S) is worth 620, while if you bid a part-score you get 170. The difference, 450, is worth 10 imps. But if you bid the partscore and the game goes down one, the result would be -100 versus 140, or a 240 imp difference, with is 6 imps. So you win 10 imps when right and lose 6 imps when wrong. So if the game has ~37.5% chance of making, you break even (in the long run) by bidding it.

 

Unlike IMPs where the size of the score is important (which is why you bid close games), at MP you win a single matchpoint for each pair you beat, and 1/2 a point for each pair you tie. NOTE" At imps the size of the difference between you and the opponent is important, but at MP, you win or lose 1 point depending on whether you scored more or less than your opponent, the SIZE of the difference at MP is irrelevant.

 

This oddity, that beating someone by 2300 points or 10 points on a hand is worth exactly the same (1 MP), which makes Matchpoint strategy very different from imp strategy. For example, people play in NT more ofter at matchpoint than imps (NT scores more), people seldom play 5C or 5D at matchpoint, prefering riskier 3NT or 4M on 4-3 fit, as they score more (if they make).

 

Ben

[Free]

Exactly, and I have 1 more thing to add: if you play a contract in imps, you'll try to make it with the highest probability, overtricks are only important when you have the extra luxe. But in mps, you'll try to make as much tricks as possible (most of the time), whatever your contract is. Only exceptions here are when you bid a contract that you think nobody will bid (example 6S), and you have a 100% sure way to make it, you won't go for that +1.

 

This means that at imps, you should make your contract most of the time if it is possible, maybe with overtricks, but in mps, you'll make more overtricks and sometimes go down in a 100% sure contract.

This can be the case if you play in NT and you have a length 6-3 with A and K, and no other entry to take these tricks. You can do 2 things: cash A and K and hope the cards are divided evenly (then you have 6 tricks), or give away a trick and make 5 tricks in that suit. In imps, you'll allways give away that trick (unless you need it), but at mps you might just cash and go for it...

[inquiry]

Here is a hand from the second B/I Lounge that would be played differently at imps and matchpoints (note hands rotated for display).

 

void

86543

KQT94

K98

KQ A97642

J T97

J7652 82

AQJT6 K98

 

JT852

AKQ2

A

542

 

NS get into 4H after West overcalled 2NT (unusual NT). Opening lead is Diamond 5 to the ACE. South cashes the HEART ACE, both follow. At imps, south has 10 sure tricks. 5H in dummy, AKQ of diamonds, an eventual club ruff comes to 9 tricks. And, this is important, the Diamond-Ten can be given to the diamond JACK, setting up the diamond 9 for the tenth trick. That is how I would play it at imps.

 

Now at Matchpoints, you can't afford to play the hand that way. With the clue from the bidding, you can play WEST for club ace, and 2-1-5-5. If you do that, you can make the all important 11th trick. The strategy is simple, play the 2NT bidder for the club ACE, so that the club king will be your 10th trick, and now you play for the 11th. Pull trumps, ruff a spade, cash one top diamond, ruff a diamond, and lead a low club towards dummy in this position...

 

void

x

QT

Kxx

Q

void

Jx

AQJ

JTx

void

void

xxx

 

On a low club, West has two losing options. If he wins the Club ACE and exits the spade QUEEN. North ruffs, cashes the club king, and throws WEST in with the last club to force a diamond lead from JX into dummies QT. If he ruffs and leads another club. Dummy wins, and cashes the HEART EIGHT. If WEST throws away a Diamond, both diamonds are good. If he throws away a club,dummies club is good, so he will be forced to throw away the spade. Now, a low spade puts WEST on lead for a diamond away from JX to QT again. And if WEST ducks the first club? You win the club KING, and once again cash the last heart. West is place back into the same position of having to keep two diamonds, so you can throw him in with a club.

 

Thus, at matchpoints, you RISK your contract (if EAST has the CLUB ACE) by ruffing the diamond NINE instead of using it to set up the TEN. All this in a effort to earn an extra 30 points (650 versus 620). You would NEVER risk that at imps, but at matchpoints that extra tricks could be worth as much a bidding at grand slam.... as you would beat every NS player who fails to find the squeeze/throw-in/endplay.

 

This hand is a perfect example of why some people LOVE matchpoints or hate it. You risk your contract (by ignoring the 100% line), to find a fun squeeze/endplay for an overtrick. The bridge technique to win this hand at matchpoints doesn't make a lot of sense to traditional imp players, who would see the 100% line and take it to make their contract. Neither play is "right" or "wrong", at one game you should play this hand one way, at another game the other way...all due to differences in the "scoring".

 

Ben

[paulg]

Ben's explanations are excellent, but another point when playing matchpoints (MPs) is that it is not necessary to bid thin games to win.

 

This is a difficult lesson to learn (especially online where everyone overbids) but will pay dividends in the long run. Again it is because the SIZE of the difference at MP is irrelevant that this is true.

 

For example, suppose you bid and make nine tricks in a 3NT contract when you only have a combined 23 points. You will beat all the other pairs and score 100% (a top) on the board. However you will see that only 20% of the field made 9 tricks in 1NT, so you would have scored 90% anyway even if you'd have stopped in 1NT and made your 9 tricks.

 

Imagine though that you had only made 8 tricks. In 3NT now you score 0% on the board, whereas in 1NT you would have scored, say, 50%.

 

So bidding thin games, and thin slams, is a poor proposition.

 

regards

 

Paul

[Free]

And actually, there's another difference. If you play imps and you help your opponents to get a heavy score, it's VERY DIFFICULT to get that back. But with mp's, one or two 0's isn't that bad...

[pbleighton]

What about defense at matchpoints.

 

I tend to play more passively at MPs, figuring if most of the field is letting the opps make 4H there is no sense trying for a low percentage heroic play to beat the contract, which will usually wind up handing the opps an overtrick.

 

Is this right?

 

Warning - be careful of encouraging my overly passive tendencies on defense ;D

 

Peter

[luis]

What about defense at matchpoints.

 

I tend to play more passively at MPs, figuring if most of the field is letting the opps make 4H there is no sense trying for a low percentage heroic play to beat the contract, which will usually wind up handing the opps an overtrick.

 

Is this right?

 

Warning - be careful of encouraging my overly passive tendencies on defense ;D

 

Peter

 

There're no Merrimacks at MPs :-) that's true but I wouldn't use the term "passive" to define the defensive strategy at MPs.

First of all you must put your side in a cash-out situation many times, whenever there's danger of being overtricked you should try to cash out some times giving away a remote chance to defeat the contract "if..." this is agressive defense not passive.

And you have most opportunities to false card since declarers are more greedy there're many falsecards that may induce a player to think that there're more tricks to him available than expected resulting in a play favorable to the defense.

[inquiry]

What about defense at matchpoints.

 

I tend to play more passively at MPs, figuring if most of the field is letting the opps make 4H there is no sense trying for a low percentage heroic play to beat the contract, which will usually wind up handing the opps an overtrick.

 

Is this right?

 

Warning - be careful of encouraging my overly passive tendencies on defense ;D

 

In the long run, passive defense at Matchpoint is a winner. Now, if dummy hits and it is obvious that the opponents are in an unusual contract (dummy overbid twice to reach game, or dummy passed short of game when it looks like everyone would bid game), you will want to revisit this idea.

 

Let's take the first case. They are in a game that you think no one else will bid. Now, you had better be sure to set them. This calls for aggressive defense, especially if you can see key suits are breaking favorably for declearer.

 

Now lets take the paradoxical situation. They stayed out of game you think everyone esle bid. Defend like crazy. If they just make and everyone else goes down, you lose anyway. If everyone else makes game, your defense doesn't matter. But if game isn't making (say down one), you had better set their part-score (does this make sense)?

 

Sometimes they will play 3NT when 4S looks obvioius on a 4-4 fit. Here you evaluate how many tricks are available to you on defense against a spade contract, and do whatever you can to win one more against the notrump contract. If there is 10 tricks in both, you are on your way to a horrible score. It is situations like this that requires judgement on how to plan your defense. In all these cases, at imps, you simply play to set the contract, so as you can see matchpoints is a much harder game to play.

 

Ben

[inquiry]

Is matchpoint different? Here is a hand to illustrate that it really is...

 

 

Dummy

S-Kxxxx

H-Ax

D-xx

C-A98x

You RHO YOU RHO Pard

S-AQJx 1S 2C 2H P

H-Q 3H P 4H all P

D-K72

C-KQTxx

 

T1. C7 to ACE-x-x

T2. Dx-7-Q-x

T3. DA-x-x-K!

 

Here I tossed the diamond king in tempo (no thinking).

South now lead a diamond and ruffed with the ACE. Ruffed a spade, and ruffed a second diamond low, which I can now overruff, and we get a club and a trump promotion. South held (he should toss a club on a spade ACE in here)...

 

S-VOID

H-KJT8xxx

D-AQxx

C-xx

 

Everyone made 5, 6 or 7 tricks in hearts but at my table, where we held them to 10 tricks. At MP, this would have been a HUGE sucess... one that would give us a big boost. At imps, it was a little more than one imp better than we could have done if we let them make 6 I think (a few missed game). So at MP, this would be exciting hand involving a trump promotion and being able to score the heart Queen (if he ruffs low on first round of diamonds, he can make all 13 tricks. 2D ruff, 2D, 1S, 1C and 7 hearts. So this little nothing hand (if they fail to bid slam) becomes quite excting at MP (no one bid slam, and in fact, 4 out of 16 didn't bid game).

[Free]

The way I usually defend in imps is very "wishfull thinking". In some games it's clear that they 'll make their contract except if partner has that card, so I'll give him that card and play for it.

 

In mps, you don't want to give away too many tricks, so we play as normal as possible, think if partner hasn't got the card we want him to have,...

[paulg]

David Burn is one of the top players in the UK and one of our most amusing commentators on the game. At our largest event - the Summer Congress - he is the most popular contributor to the "Ask the Experts" session and he frequently muses about how to play matchpoints in a large field. His observations include:

 

1. You never score badly with a positive score

 

2. Never bid slams, as you will score 75% just making 12 tricks in game

 

3. Although 5C/5D may score badly with 3NT making, bidding on and going one down in 6C/6D scores a lot worse - see point 1.

 

regards

 

paul

[inquiry]

Matchpoints

Dealer East

Vuln. NS

 

S J32

H A765

D KQJ8

C J9

 

S 965 S QT874

H J982 H KQ

D A9 D T6

C A853 C K742

 

S AK

H T43

D 75432

C QT6

 

 

Here is another hand that illustrates why Matchpoints and IMPS are really different. This is a hand from a matchpoint event. At most tables, the bidding went pass-pass-pass to north. There is a rule for this kind of hand, named the RULE of 15. If in fourth seat, count your hcp and add that to the legnth in spades. If the total is not 15 DO NOT OPEN. The idea is that SPADES is the master suit and if the opponents outbid you in spades, you have to go one level higher to buy the contract. This rule doesn't mean that if you can count up to 15 you should open. Here north has an ugly 12 hcp and only 3 spades, just adds to 15, so pass has a lot of merit to it. In fact, I would "evaluate" this 12 hcp hand to worth only 11hcp, and I would add 11 to 3 and come to 14 and pass. But if north does open 1D, he soon will have another problem. East will bid 1S, South will raise diamonds and West has enough to raise spades. So what will north do over 2S. If he passes, he is sure to get a bad score on this hand (see below). If he bids on, he has a chance to greatly improve his score or make it marginally worse. Let's examine the common results on this hand.

 

PASS OUT - 0 68%

2D making +90 84%

2SEW making 2 -110 39%

2SEW making 3 -140 26%

2/3D down 1 -100 50%

3D down 2 -200 6%

3Dx down 2 -500 2%

3SEW down 1 +50 77%

2/3SEW down 2 +100 88%

 

Even 2D for NS can go down, losing 2C, 2H, and a heart ruff (south can't overuff East diamond Ten on the fourth round of hearts). So 2D down one, is about average. But 3D down two would be disaster. Fortunately for NS, most defenders will mishandle the suit (must lead heart early, unblock hearts, cash both clubs, then lead two rounds of hearts). Two Spades for EW should always make,and like 2D, it requires very precise defense to beat 3S. In 3S, you have to start with a diamond (maybe not so hard to find) to prevent East from discarding his diamond loser on the heart JACK. Even one round of trumps before leading a diamond will not work because the heart TEN falls on the third round. You need to knock out the diamond ACE then club AK so you can score all your tricks on defense.

 

Again. At Imps, this would be a hand you would play and forget about, since it is the big hands (slams, games, especially vul) that win imp events. But at Matchpoints, this hand is worth just as much as a grand slam. And you can never give up on a hand. so at matchpoint, every hand, and every trick is important. That is why matchpoint is more tiring on a good player, and why some players love matchpoints (every hand requires intense thinking), and why some prefer imps (spend energy on the big hands).

 

Ben

[inquiry]

David Burn is one of the top players in the UK and one of our most amusing commentators on the game. At our largest event - the Summer Congress - he is the most popular contributor to the "Ask the Experts" session and he frequently muses about how to play matchpoints in a large field. His observations include:

 

1. You never score badly with a positive score

 

2. Never bid slams, as you will score 75% just making 12 tricks in game

 

3. Although 5C/5D may score badly with 3NT making, bidding on and going one down in 6C/6D scores a lot worse - see point 1.

 

regards

 

paul

 

As a general rule, these rules 1 and 2 are not that useful. For instance, in todays www.bridgebase.pl matchpoint tourney,with 102 tables if you made 12 tricks in 4H on this hand, you scored just 32%.

 

Dealer North

Vuln. NS

 

S 72

H AKQJ8

D A87

C A85

 

S AJT98543 S KQ6

H 62 H T7

D T54 D Q632

C C J972

 

S void

H 9543

D KJ9

C KQT643

 

and NS players who were +100 earned 3% scores (so much for you never score badly with a plus score).

 

So if you are reading this threat, take this advice with a LARGE grain of salt.

 

Ben

[Free]

I also don't agree with the rule "never bid slams". I allways try to bid slams, because I like to play them, and in pairs it's just going for the 100%. The rule of writing at your side is correct on the other hand. In the short run it isn't (like this game above) but in the long run you'll see that if you wrote a lot at your side, you allways get a good result.

[luis]

David Burn is one of the top players in the UK and one of our most amusing commentators on the game. At our largest event - the Summer Congress - he is the most popular contributor to the "Ask the Experts" session and he frequently muses about how to play matchpoints in a large field. His observations include:

 

1. You never score badly with a positive score

 

2. Never bid slams, as you will score 75% just making 12 tricks in game

 

3. Although 5C/5D may score badly with 3NT making, bidding on and going one down in 6C/6D scores a lot worse - see point 1.

 

regards

 

paul

 

Disagree with 1

Disagree with 2

Disagree with 3

 

I'm sure this is either missquoted or incomplete or was said as a joke since it's clear that the three tips are absolutely wrong. Just terrible.

 

"5c at matchpoints is forcing" (Luis)

[inquiry]

I also don't agree with the rule "never bid slams". I allways try to bid slams, because I like to play them, and in pairs it's just going for the 100%. The rule of writing at your side is correct on the other hand. In the short run it isn't (like this game above) but in the long run you'll see that if you wrote a lot at your side, you allways get a good result.

 

As a general rule, all rules should be ignored, even this one. :D

 

I appreciate the advice paulg tried to provide, but I think we need to be a little more careful in stating "rules" for beginners who might lack the experience to know that these are not only greatly oversimplified, I disagree in principle with all three of them. Luis's 5C's is forcing is much closer to the truth at matchponts (see below).

 

The slam advice can be proven to be wrong on "obvious" hands. I think the small slam advice might apply at matchpoints for beginners/intermediate players where they are stretching to bid the slam. The reason being if those slam go down, maybe by lack of card playing technique, they are sure to get zero or near zero. For the expert and advanced player, this slam advise is utter nonsense.

 

The 5C versus 6C advice is only acceptable if you have sound reason to believe that 3NT will have problems. So if you bypassed 3NT in a slam attempt, bitter experience has shown that 5C/5D really is not a good contract (aka, see Luis's rule above). If you bypassed 3NT out of fear of an unstopped suit, then by all means feel freet to stay in 5C/5D (but see my first matchpoint rule below... bid 3NT more often).

 

And while in the long run, plus scores are better than minus scores, but it will become obvioius during the play that some plus scores will nto be good enough, in which case, you must try really hard to get a bigger plus score. Take a look at this thread http://forums.bridgebase.com/in...msg7293#msg7293

where I said over the opponents 3D bid I would double if they are VUL to try for +200, since I "knew" I could be +110 or 140. And if they were not vul, I would bid 3H. Why, because I know from the auction that +50 or +100 will NOT be a good score for me. As you see from ron's reply, 3D would be down one, so if they were vul, my proposed dbl would be right, and 3H can make, so if they are not vul, my bidding on would be right. At imps, this hand would not have required much thought at all. 3D down one or 3H making would be almost the same.

 

Having SLAMMED giving "rules, I will now give two. My machpoint advise is to bid 3NT much more often than you are use to doing (again this is for matchpoints).

 

However, probably the best advise I can give is for imps. Where a good beginner rule is never bid grand slam. The reason for this is if you are wrong, you lose the game bonus and the small slam bonus. So grand slam making versus small slam, wins a few extra imps (13 vul, 11 nonvul), but if the grand slam goes down, you lose tons of imps. The odds are such that, say a 70% vulnerable grand slam should not be bid. Here is a table showing why....

 

Assume you know from expert bidding that you have 70% chance to make 7NT, otherwise you will make 6NT and you are vul... Also for the sake of argument, lets assume your oppoents bid 3NT, 6NT or 7NT, lets look at the results.

 

A) If you bid 6NT.

 

Make 6 make 7

opp bid +1440 +1470

3NT you win you win 13 imps for +750

6NT it is a push 0 imps

7NT you with win 17 imps or lose or lose 13

 

If you bid 7NT, you are either -100 or +2220, if the opponents"

 

Bid 3NT, you either lose 13 imps or gain 17 imps

Bid 6NT, you either lose 17 imps or gain 13 imps

Bid 7NT, you break even

 

So if the odds are 70% you make the grand slam, you can begin estimating the result if you bid it.

 

If you bid the grand slam, and they stop in game, you risk your sure 13 imp gain in an effort to gain just 4 more imps (17 versus 13). 0.7 x 4 versus 0.3 x -13, shows the odds are against bidding the slam, as the net expectation is you will lose on average 1.1 imp in this situation.

 

The odds are much better for bidding the grand slam if you think your opponents will bid the small slam. Now, you will break even if you bid the small slam, and if you bid the grand slam, you will lose 17 imps if you are wrong and gain 13 imps if you are right. Here the odds actually favor bidding the slam, but only if you are certain your opponents will reach the small slam. Playing against good opposition, go ahead and bid the grand slam. Playing in a large event like an online cross-imp event, or against random opponents in team game, it is best to stop just is small slam. Nothing is worse than bidding grand slam down one when your opponents didn't reach the small slam.

 

PS... even my never bid GRAND SLAM at imps rule is not sound. If you know you have 13 tricks or if you are sure your opponents will find the small slam, this brings the grand slam back into the picture.

 

Ben

[inquiry]

Another matchpoint hand.

Dealer South

Vuln. All

 

S AJ974

H AQ

D 983

C 976

 

S KT S Q3

H J863 H T54

D Q754 D AKT

C AJ2 C Q8543

 

S 8652

H K972

D J62

C KT

 

North will open 1S in third seat and South will raise to 2S. At imps, EW will probably stay out of the auction (although if it goes PP1SP; 2SPP east might well balance back in at imps).

 

The best way for EW to enter the bidding, at matchpoints, is for WEST, who passed originally, to double 2S for takeout. AT imps, this would be risky, as north is unlimited, but at matchpoints you have to fight for the hand. EW get to 3C and score very well indeed. And if NS bid to 3S, EW can set it two with good defense of a wonderful score. On this hand, EW scores of

-110 was 11%, +100 was 50%, +110 was 65%, +130 was 74%, and +200 was 90%.

 

So expectations for WEST actions over 2S...

 

Pass either 11% or 50% (depends upon quality of defense)

 

DBL either 65% or 74% for 3D or 50% to 90% for defending 3S (or 90+ if 3Sx)

 

Again, this is a hand no one would show if it occured at imps, and EW would not be under such pressure at that form of scoring.

 

ben

[paulg]

David Burn is one of the top players in the UK and one of our most amusing commentators on the game. At our largest event - the Summer Congress - he is the most popular contributor to the "Ask the Experts" session and he frequently muses about how to play matchpoints in a large field. His observations include:

 

1. You never score badly with a positive score

 

2. Never bid slams, as you will score 75% just making 12 tricks in game

 

3. Although 5C/5D may score badly with 3NT making, bidding on and going one down in 6C/6D scores a lot worse - see point 1.

 

regards

 

paul

 

Once again I get panned in this forum for giving advice that I think is particularly appropriate for B/Is, perhaps slightly deserved since I could have put the comments into a bit more context (and it is difficult to convey the serious but funny way that DALB presents this stuff).

 

However I would say that my comments (and David's) were definitely targeted at the B/I audience - playing matchpoints in a large field of B/Is is very different to the game that most adv/exp players see.

 

Anyhow, point 1 was in the context of not stretching to bid contracts like thin games, and protecting your positive score. Making a game try and then going down at the three level is a good example of poor matchpoint tactics.

 

The slam was a bit tongue-in-cheek - it is probably true though that beginners would score more if they never bid (or looked for) slams but then what fun would the game be :D

 

Point 3 was about if you have missed 3NT, then it's likely that other B/Is will have missed it too. I believe 5C is forcing is an expert rule, not a B/I rule.

 

Like Ben commented, bidding 3NT more often was also a theme - "you'll never be alone in 3NT" :)

 

Cheers

 

paul