Help
Can any one explain at IMP's a contract of 3n/t +1 freely bid,despite for examle a known 5/4 fit in spades or hts scores the same as 4s or 4 hts making
620 V 630---is there a reason and does this apply to matchpoints, although i never play MP-is it possibly to do with a possible defence-that 4h/4sp is 100% and 3 n/t can be held to 9 tricks???
Scoring IMP's
#1
- Group: Full Members
- Posts: 638
- Joined: 2008-November-06
- Location:asia at present time now HK time
- Interests:Bridge- scuba-natural sex,no porn:)<br> Associate member I.B.P.A. workaholic
Posted 2010-January-22, 11:06
#2
- Group: Advanced Members
- Posts: 2,772
- Joined: 2006-January-30
Posted 2010-January-22, 11:19
10 pt differences to do not score at IMPs...period.
#3
- Limit bidder
-
- Group: Advanced Members
- Posts: 8,482
- Joined: 2004-November-02
- Gender:Female
- Location:England
- Interests:Bridge, classical music, skiing... but I spend more time earning a living than doing any of those
Posted 2010-January-23, 07:06
At matchpoints there is a huge difference between +630 and +620.
At IMPs a 10-point difference makes no difference at all.
As for why a pair might bid 3NT with a 5/4 spade fit, I can think of three reasons off-hand:
1. They thought they were playing matchpoints, or point a board (BAM)
2. They didn't realise they had a 5/4 spade fit, because they had forgotten their system, or misbid, or not seen the auction properly
3. They thought 3NT might be more likely to make. Perhaps the hand was something like this:
At IMPs a 10-point difference makes no difference at all.
As for why a pair might bid 3NT with a 5/4 spade fit, I can think of three reasons off-hand:
1. They thought they were playing matchpoints, or point a board (BAM)
2. They didn't realise they had a 5/4 spade fit, because they had forgotten their system, or misbid, or not seen the auction properly
3. They thought 3NT might be more likely to make. Perhaps the hand was something like this:
#4
- Group: Full Members
- Posts: 447
- Joined: 2007-January-26
- Location:York, England
- Interests:People, Places, Humour
Posted 2010-January-23, 15:00
At Imp scoring, you are rewarded for bidding games, especially vulnerable games, but it is not realy important which game you may choose to bid, except that you should normally choose the game contract which is most likely to make.
This might mean bidding 5-minor contracts more often, and 3NT contracts whenever you have sufficient points. The main point is that at Imps you look for safety first. "Bid boldly, Play safe" (Rixi Marcus) so bid most borderline games.
At Matchpoints, you must try to outscore your opponents who are bidding the same cards. The gain of 10 points can be huge if 3NT and 4♠/♥ are both making 10 tricks, but the major suit contract is preferred if you can score extra ruffing tricks, so shortages in dummy are essential. Bidding borderline games is not as important unless you think most others will be bidding it.
Tony
This might mean bidding 5-minor contracts more often, and 3NT contracts whenever you have sufficient points. The main point is that at Imps you look for safety first. "Bid boldly, Play safe" (Rixi Marcus) so bid most borderline games.
At Matchpoints, you must try to outscore your opponents who are bidding the same cards. The gain of 10 points can be huge if 3NT and 4♠/♥ are both making 10 tricks, but the major suit contract is preferred if you can score extra ruffing tricks, so shortages in dummy are essential. Bidding borderline games is not as important unless you think most others will be bidding it.
Tony
Hanging on in quiet desperation, is the English way (Pink Floyd)
#5
- Group: Full Members
- Posts: 612
- Joined: 2007-December-31
- Location:Bethesda, Maryland, USA
Posted 2010-January-23, 18:27
A real hand, playing IMPs:
xxx
AKJx
xxx
AKx
P-P-P-1NT
P-2D!-P-2H
P-3NT-P-?
As it happened partner was 5-5 so bidding 3NT was quite precipitous. But if he has a balanced 9 or 10 count, do you really think you can make 4H?
xxx
AKJx
xxx
AKx
P-P-P-1NT
P-2D!-P-2H
P-3NT-P-?
As it happened partner was 5-5 so bidding 3NT was quite precipitous. But if he has a balanced 9 or 10 count, do you really think you can make 4H?
"It is not enough to be a good player. You must also play well." -- Tarrasch
#6
- Group: Members
- Posts: 44
- Joined: 2008-November-08
Posted 2010-January-25, 05:32
The question is a little vague: are you asking why you might choose to avoid an 8-card major fit? At IMPs, Frances's example is a perfect one: there are 9 certain tricks, 4 top losers, and nothing useful to ruff, so a notrump contract will surely be better.
More generally, at IMPs it is better to bid the safer game, and at matchpoints the higher-scoring one. A lot of players read this as: at matchpoints avoid 5m, and consider avoiding 4M when 3NT might take as many tricks. But often the reverse is true: avoid 3N if 4M might take more tricks. For example: (sorry, can't make the diagram buttons work)
KJ98
JTxxx
x
KQx
AQT
9xx
AKxx
Axx
3NT is laydown, but you'll make only nine tricks on a minor-suit lead. In 4♠, you have several chances to make ten tricks. None of those chances are guaranteed, though, and if the opponents manage to defend well on an unfavorable layout you will go down. The odds are a bit too complicated for me to analyze, but it certainly looks better than 50%, so I'd want to be in 4♠ at pairs. At IMPs, however, 3NT is certainly best.
More generally, at IMPs it is better to bid the safer game, and at matchpoints the higher-scoring one. A lot of players read this as: at matchpoints avoid 5m, and consider avoiding 4M when 3NT might take as many tricks. But often the reverse is true: avoid 3N if 4M might take more tricks. For example: (sorry, can't make the diagram buttons work)
KJ98
JTxxx
x
KQx
AQT
9xx
AKxx
Axx
3NT is laydown, but you'll make only nine tricks on a minor-suit lead. In 4♠, you have several chances to make ten tricks. None of those chances are guaranteed, though, and if the opponents manage to defend well on an unfavorable layout you will go down. The odds are a bit too complicated for me to analyze, but it certainly looks better than 50%, so I'd want to be in 4♠ at pairs. At IMPs, however, 3NT is certainly best.
#7
- Follower of 655321
-
- Group: Members
- Posts: 3,293
- Joined: 2008-December-05
- Interests:drinking, women, bridge...what else?
Posted 2010-January-25, 09:15
9 tricks are easier than 10
#8
- Brain Farts and Actual Farts Increasing with Age
-
- Group: Advanced Members
- Posts: 9,586
- Joined: 2005-September-21
- Gender:Male
- Location:Lima, Allen County, North-West-Central Ohio, USA
- Interests:www.limadbc.blogspot.com editor/contributor
Posted 2010-January-25, 10:45
The "asnwer" is that this is how we keep score.
If there was some bizarre bonus for something completely absurd, we might all shoot for that. For example, if the score for a contract 3♣ making exactly four somehow resulted in a bonus of 500, then people might bid 3♣ more frequently, with interesting declaring and defense problems arising, like intentional trick loss and counters to the same.
Actually, that concept might have some interesting permutations, such as if there was an additional penalty (like in some games) for "underbidding." If overtricks cost more than lost opportunity, but actually resulted in negative points, then the defense could actually launch a counter-intuitive lose-tricks defense, and Opener as well, with counters all over the place.
Suppose, for example, that 3♣ making four yielded +110 (for making 3♣), but -50 for the overtrick, netting +60. Two overtricks might reduce an additional -100, for a net of -40. The defense, then, could switch tactics to a sabotage defense.
Strange situations might develop, where Declarer gets caught in a situation where he can never get the exact 9 tricks contracted for but will either take 10 on one line or face only 8 with the counter-strike. Of course, he'd usually opt the overtrick, unless he missed the counter-strike option.
If there was some bizarre bonus for something completely absurd, we might all shoot for that. For example, if the score for a contract 3♣ making exactly four somehow resulted in a bonus of 500, then people might bid 3♣ more frequently, with interesting declaring and defense problems arising, like intentional trick loss and counters to the same.
Actually, that concept might have some interesting permutations, such as if there was an additional penalty (like in some games) for "underbidding." If overtricks cost more than lost opportunity, but actually resulted in negative points, then the defense could actually launch a counter-intuitive lose-tricks defense, and Opener as well, with counters all over the place.
Suppose, for example, that 3♣ making four yielded +110 (for making 3♣), but -50 for the overtrick, netting +60. Two overtricks might reduce an additional -100, for a net of -40. The defense, then, could switch tactics to a sabotage defense.
Strange situations might develop, where Declarer gets caught in a situation where he can never get the exact 9 tricks contracted for but will either take 10 on one line or face only 8 with the counter-strike. Of course, he'd usually opt the overtrick, unless he missed the counter-strike option.
"Gibberish in, gibberish out. A trial judge, three sets of lawyers, and now three appellate judges cannot agree on what this law means. And we ask police officers, prosecutors, defense lawyers, and citizens to enforce or abide by it? The legislature continues to write unreadable statutes. Gibberish should not be enforced as law."
-P.J. Painter.
-P.J. Painter.
#9
- Group: Full Members
- Posts: 447
- Joined: 2007-January-26
- Location:York, England
- Interests:People, Places, Humour
Posted 2010-January-25, 11:06
kenrexford, on Jan 25 2010, 05:45 PM, said:
The "answer" is that this is how we keep score.
Suppose, for example, that 3♣ making four yielded +110 (for making 3♣), but -50 for the overtrick, netting +60. Two overtricks might reduce an additional -100, for a net of -40. The defense, then, could switch tactics to a sabotage defense.
Suppose, for example, that 3♣ making four yielded +110 (for making 3♣), but -50 for the overtrick, netting +60. Two overtricks might reduce an additional -100, for a net of -40. The defense, then, could switch tactics to a sabotage defense.
Love the idea, can you patent it?
On the theme of scoring, I would like to see larger slam bonuses and a higher score for minor suit contracts
If clubs and diamonds scored 25 points per odd trick, the game would be vastly improved
Tony
Hanging on in quiet desperation, is the English way (Pink Floyd)
#10
- Group: Members
- Posts: 16
- Joined: 2005-July-26
Posted 2010-January-30, 17:06
NEW SCORING METHOD PROPOSAL: Board-Weighted Percent Scoring.
Now that computers make scoring so much easier, why don't we combine the nice features of both MP and IMP scoring - more explicitly let's have a scoring system for pairs that gives a %, without declarers being clobbered for giving up the chance of an overtrick to ensure the contract makes. The system I propose is particularly useful for duplicate events with small numbers of tables, where matchpoints are almost meaningless.
Here are the details:
For all NS pairs, subtract the lowest NS score from theirs.
For all EW pairs, subtract the lowest EW score from theirs.
This will make the lowest scoring NS and EW pairs get 0.
Divide these adjusted scores by the highest adjusted score in their direction, and multiply by 100, to get a %. Call this P. For those concerned about possible division by zero (for flat boards), define P = 0 in these cases, though this problem goes away naturally, as it turns out (see next-but-1 paragraph).
As it stands, this unfairly penalises pairs who make game but don't get an overtrick, if every other pair foregoes the safety play and makes the overtrick. So to fix this, we define a 'board weight', as follows:
Find the range of NS scores (highest minus lowest), convert this 'score' to IMPs, then multiply this by the number of other results (i.e. 1 less than the number of results for the board). This is the board's weight - call this W. Clearly, a flat board has a weight of zero. Some may object to this, but if everyone gets the same score on a board, why should it be considered in assessing someone's score? On some boards, any sensible line results in the same number of tricks; does scoring the same as everyone else on such a board mean you are a 50% player? I would say it means nothing. This also avoids worries about defining P for flat boards.
A pair's final % is the sum of P x W for each of the boards they play, divided by the sum of the weights of the boards they played.
Now a quick point about IMPs. To make it easy to score without a computer, IMPs were defined as a step function of points scored, but with a computer we can reward scores much more accurately (as BBO does) by using decimals. There may be a formula for this somewhere, but I couldn't find it, so I made my own, which stays pretty close to the step function, though it slightly overvalues low scores.
IMPs = Log(1+Score/2)^2.6
I've written an Excel file that can be used to score events in this way. If anyone wants a copy of it, send me an email at benbeever@hotmail.com
The main problem with this method, as I see it, is how best to desensitize the scoring to extreme, freak scores. IMPing the scores before calculating their % is one way. Capping all scores at 500 pts above the 2nd placed score for their direction is another.
Now that computers make scoring so much easier, why don't we combine the nice features of both MP and IMP scoring - more explicitly let's have a scoring system for pairs that gives a %, without declarers being clobbered for giving up the chance of an overtrick to ensure the contract makes. The system I propose is particularly useful for duplicate events with small numbers of tables, where matchpoints are almost meaningless.
Here are the details:
For all NS pairs, subtract the lowest NS score from theirs.
For all EW pairs, subtract the lowest EW score from theirs.
This will make the lowest scoring NS and EW pairs get 0.
Divide these adjusted scores by the highest adjusted score in their direction, and multiply by 100, to get a %. Call this P. For those concerned about possible division by zero (for flat boards), define P = 0 in these cases, though this problem goes away naturally, as it turns out (see next-but-1 paragraph).
As it stands, this unfairly penalises pairs who make game but don't get an overtrick, if every other pair foregoes the safety play and makes the overtrick. So to fix this, we define a 'board weight', as follows:
Find the range of NS scores (highest minus lowest), convert this 'score' to IMPs, then multiply this by the number of other results (i.e. 1 less than the number of results for the board). This is the board's weight - call this W. Clearly, a flat board has a weight of zero. Some may object to this, but if everyone gets the same score on a board, why should it be considered in assessing someone's score? On some boards, any sensible line results in the same number of tricks; does scoring the same as everyone else on such a board mean you are a 50% player? I would say it means nothing. This also avoids worries about defining P for flat boards.
A pair's final % is the sum of P x W for each of the boards they play, divided by the sum of the weights of the boards they played.
Now a quick point about IMPs. To make it easy to score without a computer, IMPs were defined as a step function of points scored, but with a computer we can reward scores much more accurately (as BBO does) by using decimals. There may be a formula for this somewhere, but I couldn't find it, so I made my own, which stays pretty close to the step function, though it slightly overvalues low scores.
IMPs = Log(1+Score/2)^2.6
I've written an Excel file that can be used to score events in this way. If anyone wants a copy of it, send me an email at benbeever@hotmail.com
The main problem with this method, as I see it, is how best to desensitize the scoring to extreme, freak scores. IMPing the scores before calculating their % is one way. Capping all scores at 500 pts above the 2nd placed score for their direction is another.
#11
- Duchess of Malaprop
-
- Group: Full Members
- Posts: 722
- Joined: 2006-March-09
- Gender:Not Telling
- Interests:i am not interested
Posted 2010-January-30, 17:35
Beever
Could you elaborate on the impact of this scoring method on the bidding of a few of the major systems, like sa, precision or 2/1, for example.
These systems pefrom better or worse for different types of scoring.
Could you elaborate on the impact of this scoring method on the bidding of a few of the major systems, like sa, precision or 2/1, for example.
These systems pefrom better or worse for different types of scoring.
Free, on 2011-May-10, 03:57, said:
Babalu just wanted a shoulder to cry on, is that too much to ask for?
#12
- The Abbess
-
- Group: Advanced Members
- Posts: 17,397
- Joined: 2004-April-22
- Gender:Female
- Location:Odense, Denmark
- Interests:History, languages
Posted 2010-January-30, 17:57
Lu, this has obviously no impact on bidding systems. It may have impact on how aggressive you should bid, e.g. you may need only 9.3 rather than 9.4 HCPs to raise 1NT to 3NT. But the definition of your 1NT opening will not change.
The system would be identical to MP if there are only two tables, and it would approach IMPs (in terms of strategy) when the number of tables becomes very large. For a moderate number of tables, it would be somewhere in between.
The system would be identical to MP if there are only two tables, and it would approach IMPs (in terms of strategy) when the number of tables becomes very large. For a moderate number of tables, it would be somewhere in between.
The world would be such a happy place, if only everyone played Acol :) --- TramTicket
#13
- Group: Members
- Posts: 16
- Joined: 2005-July-26
Posted 2010-January-30, 18:29
Helene - although the % scores for each pair would be the same as MPs when there are only 2 tables, the big improvement would be that a board where say 1 NS pair were in no-trumps (and scored 10 points more than the other pair in hearts/spades) would be worth relatively little, compared to the board where 1 NS pair made game whilst the other NS pair went off. This therefore mimics IMPs scoring even with a low number of tables.
#14
- Group: Advanced Members
- Posts: 1,488
- Joined: 2006-March-23
- Gender:Male
Posted 2010-January-30, 19:10
BnBeever, on Jan 30 2010, 07:29 PM, said:
Helene - although the % scores for each pair would be the same as MPs when there are only 2 tables, the big improvement would be that a board where say 1 NS pair were in no-trumps (and scored 10 points more than the other pair in hearts/spades) would be worth relatively little, compared to the board where 1 NS pair made game whilst the other NS pair went off. This therefore mimics IMPs scoring even with a low number of tables.
Why is this an improvement?
The artist formerly known as jlall
#15
- Duchess of Malaprop
-
- Group: Full Members
- Posts: 722
- Joined: 2006-March-09
- Gender:Not Telling
- Interests:i am not interested
Posted 2010-January-30, 19:18
helene_t, on Jan 30 2010, 06:57 PM, said:
Lu, this has obviously no impact on bidding systems. It may have impact on how aggressive you should bid, e.g. you may need only 9.3 rather than 9.4 HCPs to raise 1NT to 3NT. But the definition of your 1NT opening will not change.
The system would be identical to MP if there are only two tables, and it would approach IMPs (in terms of strategy) when the number of tables becomes very large. For a moderate number of tables, it would be somewhere in between.
The system would be identical to MP if there are only two tables, and it would approach IMPs (in terms of strategy) when the number of tables becomes very large. For a moderate number of tables, it would be somewhere in between.
i beg to disagree helene
acol and precision systems do work better in imps
2/1 and ks better at matchpoints
It also has an impact in defensive bidding agreements, if you read mike lawrences books on competitive bidding he always elaborates on different actions depending on scoring
i am not ken rexford, so i want to open a book and learn from that
Free, on 2011-May-10, 03:57, said:
Babalu just wanted a shoulder to cry on, is that too much to ask for?
#16
- Duchess of Malaprop
-
- Group: Full Members
- Posts: 722
- Joined: 2006-March-09
- Gender:Not Telling
- Interests:i am not interested
Posted 2010-January-30, 19:22
BnBeever, on Jan 30 2010, 07:29 PM, said:
Helene - although the % scores for each pair would be the same as MPs when there are only 2 tables, the big improvement would be that a board where say 1 NS pair were in no-trumps (and scored 10 points more than the other pair in hearts/spades) would be worth relatively little, compared to the board where 1 NS pair made game whilst the other NS pair went off. This therefore mimics IMPs scoring even with a low number of tables.
then if this thing is similar to imps scoring, then i think i will pass
i much prefer matchpoints games, i prefer the bidding, i prefer the defense, i prefer the card play
People who like imps have a lot more opportunities to play online that i have
me want matchpoints
Free, on 2011-May-10, 03:57, said:
Babalu just wanted a shoulder to cry on, is that too much to ask for?
#17
- Group: Members
- Posts: 16
- Joined: 2005-July-26
Posted 2010-January-30, 20:12
PhantomSac, on Jan 30 2010, 08:10 PM, said:
BnBeever, on Jan 30 2010, 07:29 PM, said:
Helene - although the % scores for each pair would be the same as MPs when there are only 2 tables, the big improvement would be that a board where say 1 NS pair were in no-trumps (and scored 10 points more than the other pair in hearts/spades) would be worth relatively little, compared to the board where 1 NS pair made game whilst the other NS pair went off. This therefore mimics IMPs scoring even with a low number of tables.
Why is this an improvement?
Well of course everything is a matter of taste, but it seems that 90% of BBO players would agree it is an improvement, as roughly this proportion play for IMPs rather than MPs.
#18
- Group: Full Members
- Posts: 130
- Joined: 2009-December-26
Posted 2010-January-30, 22:49
This reminds me of my proposal from a few (OK, many) years ago.
Each pair's score is the IMP score vs. a zero par, so +110 gets you 3IMP. Of course if the board is flat then every NS gets +3 and every EW gets -3, so there is no overall effect. Equally, for any flat board, regardless of result, there is no overall effect, provided that one compares NS with NS and EW with EW.
However, if any pair has bid and made a small slam, the par score becomes the slam bonus (+500 or +750) instead of zero. If any pair has bid and made a grand slam, the par score becomes the grand slam bonus (+1000 or +1500) instead of zero or the small slam par.
In those extreme and rare cases where both sides have bid and made a small slam, then the par goes back to zero. Ditto for both sides making a grand slam (in your dreams). If one side bids and makes a small slam and the other side bids and makes a grand slam, then the par becomes the small slam bonus for the grand slam side.
The point of this scoring system is two fold. First, the effect of those weird sets where someone has been dropped in a redoubled cue bid is minimized. The two pairs affected have a strange result, but it doesn't change other effects much. Second, there is still a big incentive to bid and make a slam. Suppose one pair bids and makes a small slam in hearts, not vul. Then, that pair scores +980 and the par becomes +500. So, all the pairs in game score -20, or -1IMP. The slamming pair wins 10 IMP. If that one pair goes down, then they score -50 or -2IMP, while the game pairs score +450 or +10IMP. So, bidding a slam on a finesse is more or less 50-50, which is what my intuition says is right. (Of course, if someone bids and makes a small slam on three finesses and a suit break, when the sensible pairs are in game, then the slam bidders still win 10 IMP. This doesn't eliminate luck.)
Each pair's score is the IMP score vs. a zero par, so +110 gets you 3IMP. Of course if the board is flat then every NS gets +3 and every EW gets -3, so there is no overall effect. Equally, for any flat board, regardless of result, there is no overall effect, provided that one compares NS with NS and EW with EW.
However, if any pair has bid and made a small slam, the par score becomes the slam bonus (+500 or +750) instead of zero. If any pair has bid and made a grand slam, the par score becomes the grand slam bonus (+1000 or +1500) instead of zero or the small slam par.
In those extreme and rare cases where both sides have bid and made a small slam, then the par goes back to zero. Ditto for both sides making a grand slam (in your dreams). If one side bids and makes a small slam and the other side bids and makes a grand slam, then the par becomes the small slam bonus for the grand slam side.
The point of this scoring system is two fold. First, the effect of those weird sets where someone has been dropped in a redoubled cue bid is minimized. The two pairs affected have a strange result, but it doesn't change other effects much. Second, there is still a big incentive to bid and make a slam. Suppose one pair bids and makes a small slam in hearts, not vul. Then, that pair scores +980 and the par becomes +500. So, all the pairs in game score -20, or -1IMP. The slamming pair wins 10 IMP. If that one pair goes down, then they score -50 or -2IMP, while the game pairs score +450 or +10IMP. So, bidding a slam on a finesse is more or less 50-50, which is what my intuition says is right. (Of course, if someone bids and makes a small slam on three finesses and a suit break, when the sensible pairs are in game, then the slam bidders still win 10 IMP. This doesn't eliminate luck.)
codo said:
It is a fact that most people here write as if their opinion is a dogmatic fact.
eugene hung said:
My opinion is that this ought to win the award for best self-referential quote of the new year.
#19
- Group: Advanced Members
- Posts: 10,611
- Joined: 2009-September-15
- Gender:Female
- Location:London
Posted 2010-January-31, 03:35
BnBeever, on Jan 31 2010, 03:12 AM, said:
PhantomSac, on Jan 30 2010, 08:10 PM, said:
BnBeever, on Jan 30 2010, 07:29 PM, said:
Helene - although the % scores for each pair would be the same as MPs when there are only 2 tables, the big improvement would be that a board where say 1 NS pair were in no-trumps (and scored 10 points more than the other pair in hearts/spades) would be worth relatively little, compared to the board where 1 NS pair made game whilst the other NS pair went off. This therefore mimics IMPs scoring even with a low number of tables.
Why is this an improvement?
Well of course everything is a matter of taste, but it seems that 90% of BBO players would agree it is an improvement, as roughly this proportion play for IMPs rather than MPs.
Why aet up a system of matchpoints that mimic IMPs, rather than just playing IMPs?
I know not with what weapons World War III will be fought, but World War IV will be fought with sticks and stones -- Albert Einstein
#20
- Group: Advanced Members
- Posts: 10,611
- Joined: 2009-September-15
- Gender:Female
- Location:London
Posted 2010-January-31, 03:40
Dirk Kuijt, on Jan 31 2010, 05:49 AM, said:
So, bidding a slam on a finesse is more or less 50-50, which is what my intuition says is right.
What do you mean by "what your intuition says is right"? You don't need to rely on intuition; the percentage you need for slam can be calculated exactly at various combinations of form of scoring and vulnerability.
I know not with what weapons World War III will be fought, but World War IV will be fought with sticks and stones -- Albert Einstein
2NT - 3NT
P