Tuesday, February 23, 2010

Points, schmoints

This posting is actually a proposal for a way that the ACBL could improve the masterpoints scheme and the corresponding ranks such that decent players who will never achieve Grand Life Master status, would have something meaningful to work for. I'm not wedded to the specifics of the proposal (one precious metal is much like another) and the particular point values I suggest are quite arbitrary. Let me begin with a little introduction.

Who cares about masterpoints? Non-life-masters care a lot. Experts couldn't care less. Other players are somewhere in between, although it's probably not too far from the truth to say that nobody who has achieved gold life master status, cares much about their points. When was the last time you heard a really good player looking at the recap sheet and saying "we got 13 points!"? Come to think of it, when was the last time you saw a really good player even at the recap sheet?

I can't really say what motivates expert players, but I'm sure isn't masterpoints. That wasn't always true – there was a time of course when masterpoints really meant something. At the other end of the spectrum, I know quite a few players who regard masterpoints as a tax: the more they earn the sooner they will be forced to play in flight A. Neither of these constituencies gains much from the masterpoints scheme.

Leaving aside for the moment that an expert player might be being rewarded by getting paid, I think that what motivates such players, apart from the simple love of the game, is winning. Winning is the only path to recognition (and respect) by their peers.

The conclusion that I draw from this is that, if the ACBL really wants a way of recognizing achievement beyond life master, they should come up with a new system of points which recognize winning only. Platinum points, as currently defined, are OK but they are only available at NABCs and not everyone can afford to go to such tournaments. Moreover, fractional platinum points are awarded for placing quite distantly in an NABC+ pairs event.

This month's ACBL bulletin has yet more letters bemoaning the failings of the current masterpoints system. Here's my suggestion for another type of points: titanium, awarded only for winning open events at sectionals, regionals, and of course Nationals. No points for second place, no points for flight B. No points for winning a bracketed KO (unless it's the top bracket). Just winning. Two-session events would score at least double the points for a single-session event. Some exceptions might be made, such as winning a section (26 pairs) at an NABC+ event such as the LM pairs because that's probably at least as difficult as winning a two-session local sectional. Or winning a KO match in the Spingold or Vanderbilt: because of seeding there are no easy wins. I think there should probably be some minimum table requirements too. For example, a 10-table sectional event should probably not count (or maybe the award would be reduced to 1). Perhaps "consolation" events should not count either (that might rule out all single-session regional events).

Here are some possible titanium point values (these are just suggestions):
event typeaward
single-session sectional* 2
double-session sectional 5
single-session regional 4
double-session regional 10
NABC+ section top as now (7)
KO (Spingold/Vanderbilt) win 15
NABC+ as now (typically 120)

* including STAC games provided that there are at least two clubs competing and at least 20 tables.

As you can see, the titanium awards would be about the same, numerically, as you would receive in platinum points. But they would not be interchangeable.

The requirements for silver, gold, diamond, etc. life master rankings could be amended to include a certain number of such points. For example (again, these are merely suggestions):

rankrequirement
silver LM 1000 points, inc. 1 titanium
gold LM 2500 points, inc. 10 titanium and at least one two-session sectional win
diamond LM 5000 points, inc. 25 titanium and at least one two-session regional win
platinum LM 10000 points, inc. 50 titanium and at least two two-session regional wins

This would give people who have achieved life master status something meaningful to work for. For those of us who will never reach the rank of Grand LM, there is really nothing to work for beyond LM. Just more and more relatively meaningless points.

I think this would give back meaning to the idea of masterpoints, something more like in the "good old days" that we newer players keep hearing about.

Sunday, February 21, 2010

Swiss roll

Ever noticed how quickly your team can slide out of the money in a Swiss?  My teammates and I were in the B1 bracket in the Watertown Swiss and were leading after six matches.  We had 81 VPs and 5 wins.  Even though we won 16 VPs in the last match, we were (almost) out of the money.

That's because we got blitzed by the eventual winners in the 7th match.  Two of the key decisions were slams.  Our opponents bid 6 missing an Ace as well as KJxx of trumps.  But everything came home.  Our teammates bid a 69% slam but it was off.

The point I'm trying to make is that in a Swiss, particularly using the 20-point VP scale, you cannot afford to lose by more than about 12 imps in any match and still expect to win.  You have to win some VPs in every match.  The 20 point scale is not forgiving of blitzes.  The winners did have one match in which they lost by 19 imps, but they steadily made up ground in all of the other matches.

In the flight A event, the winners actually lost 3 matches, but they never lost by more than 10 imps.  There were teams that won 5-and-a-half and 6 matches, respectively and ended up 4/5 and 7th respectively.  This does seem a little zany.  Most probably this could be "corrected" by employing the 30-point scale, which emphasizes winning, more like the "old" style of 0,1/4,1/2,3/4,1 point.

Using the 30-point scale, we would have been a clear 4th in our flight, rather than 5th equal.  Counting wins as in the old days, we would have been 2nd equal.

There's nothing wrong with the 20-point scale.  But I do like the 30-point scale – it's more true to the origins of the Swiss format.

Wednesday, February 17, 2010

Minor Suit Stayman

I recall reading an article about Minor Suit Stayman after 1NT (or 2NT) in the Connecticut Bridge Association's The Kibitzer.  I knew that I liked the scheme but I couldn't remember the details.  Now I've found it: it's by well-known Connecticut expert Harold Feldheim and was published in the February 2007 edition: 4-4 Fit in Any Suit.

The general idea is that 3♠ is a relay to 3NT after which responder describes her hand further: 4♣ and 4 show a good suit, while 4 and 4♠ are splinter bids showing both minor suits.  4NT is a balanced hand with both minors.  He doesn't say whether the 4-of-a-minor bids should have any particular holding in the suit, or in the other minor suit, but presumably at least a five-card suit with two top honors is about right, and enough of the other minor that it wouldn't be better to start with Stayman or a Jacoby transfer.

In his example, South has ♠64 KT65 AQ96 ♣AK9.  The bidding proceeds 1NT – 3♠ – 3NT – 4NT – 6.  Most pairs were in 4NT making five, while some brave souls struggled in 6NT going down one.  But our heroes got to the making slam.  As he says, it doesn't come up often, but when it does, try not to chortle!

There is another way to get to this excellent slam which many pairs would not manage, even though the treatment is, nominally, "standard". Suppose that responder bids 4NT (as many did).  Assuming South decides to accept (not automatic perhaps, but reasonable), then exactly how should South accept?  The answer is that he bids 5 of his lowest ranking four-card suit, in this case diamonds and if North doesn't have diamonds, he bids his lowest ranking four-bagger until they run out of suits in which case they can opt for 6NT (although I suppose there might then be a case for stopping in 5NT).  In this case North, with four diamonds can raise to six right away.

Tuesday, February 16, 2010

Redoubling with three trumps

Why is that books and teachers tell us to redouble with a good hand and three of partner's major suit after RHO doubles?  What is the justification for this bit of nonsense?

The redouble should "imply no fit" and that means fewer than three major suit trumps.  How will partner be able to make a confident penalty double if he has a lurking suspicion that you might have three of his trumps?

Let's take an example. You hold ♠K84 K7 QT42 ♣QJ54, partner opens one heart and RHO doubles.  This is a clear redouble.  If partner has a decent spade holding (or if LHO bids a minor) we will double whatever they bid.  If LHO bids 1♠ and it comes back around, we can bid 1NT.

But what if we hold this similar but significantly different hand: ♠K84 K97 QT42 ♣QJ4?  If partner opens 1 and RHO doubles, we should pass.  Yes, pass!  We would normally bid a forcing 1NT and then jump in hearts.  So, in other words, we're willing to bid up to 3 without hearing of any extras that partner might hold.  If we pass now, and 3 is still available, we will bid it when it's next our turn.  Only if the bidding is already at the level of 3♠, will we be inconvenienced. In that case, we will double or bid NT according to the vulnerability.  Partner will know that we have exactly three hearts because we didn't redouble earlier and yet obviously have a good hand.  He will be very well placed to know what to do.

But I hear you saying "how will partner know we have a good hand when it's his second turn to bid?"  He won't.  But we'll enlighten him at our next turn.

Isn't it better for him to know your strength right away?  No.  It's much more important to tell partner about fit (or lack thereof) right away.  Strength can be shown later.

If you redouble holding two or three trumps, partner will know your strength right away but he still won't know what to do when your LHO jumps to 3♠.  He can't really double (penalty) with ♠Q5 AQJ642 K8 ♣AT5 because his spades aren't great and you might have three trumps.  But how can he bid 4 when you might have zero, one or two hearts?

So, does it seem odd to be passing with 10hcp opposite partner's opening?  Maybe, but the auction isn't over.  In the extremely rare event that they all pass, your partner will be playing 1X. That's not game, you say?  Indeed it isn't.  But let's see what our score will be if we can actually make 10 tricks?  Vulnerable, it will be 760 (like the sound of that?) and non-vulnerable, it will be 460 (not quite so good at IMPs but a very fine score at matchpoints).  But that isn't going to happen.  Someone is going to bid again, usually LHO.  If he passes, it means that the trumps are breaking 5-0 and this time you probably were never going to make ten tricks anyway.  But if you make your contract (perhaps with an overtrick) you're going to be well ahead.

BTW, since our redouble message now is only "no fit", we can actually do it with fewer than 10 points (if we compete later, we have at least 10).  But suppose we pick up this hand:  ♠KT84 J752 QT42 ♣4 and hear it go 1♣ double, aren't we likely to be better off in a different suit?  If LHO passes (he might because he might have long clubs), partner can bid another suit.  Our redouble has just performed double-duty as an S-O-S redouble.  If partner has decent clubs and no other suit, he can sit for it.  We probably shouldn't take this to extremes and do it with a complete Yarborough in case partner sits but can't make 1♣.  OTOH, if partner opens 1♠, and our hand is ♠4 J752 QT42 ♣KT84, we may already be in our best spot and LHO is very unlikely to pass it out anyway.

Monday, February 15, 2010

The world's worst bids?...

... or a Tale of Two Three-Diamond bids.

Some people just never know when to shut up, it seems. Unfortunately, I'm one of them. I'll come to that later. Both of these candidates for world's worst bid were perpetrated in the A/X pairs at the Cromwell regional.

A well known expert from Connecticut came to our table with this hand: ♠84 K8 9765432 ♣Q4. At love all, my partner opened on his right with 1♣. Our hero bid 3 which I doubled (negative). I was holding a 4333 11-point hand including AJT and was planning to bid 3NT unless my partner bid spades (in which case 4♠) [we play weak no-trumps so there was a reasonable chance that partner had a balanced 15-17, failing that a hand with a good club suit]. To my surprise, partner left the double in. We were destined to score 1100 but we managed to butcher the defense so badly that we only took 500.  Declarer observed drolly that at least he'd be beating the pairs in 2X down 4!  But it was not to be.  500 for us was a clear top as our best possible normal result would have been 460.

Now we come to the other candidate for world's worst bid.  I picked up ♠J82 95 AQT8654 ♣7 vulnerable against not.  My LHO opened 2♣ (strong, artificial), partner passed and RHO bid 3♣ (a positive, game-forcing bid).  You guessed it: I bid 3.  This was doubled (penalty) on my left and all passed.  Unfortunately, partner's dummy had good and bad news.  The good news was that, as expected, he had a bad hand: four points to be precise.  The bad news was that his four points were the A :(  Together with my other red ace, we could set any slam by the opponents (there were no voids).  Had that ace been in one of the opponents hands, where it was supposed to be, we would have been down 800 but there would have been lots of 990s at the other tables.  -500 got us the fine score of one matchpoint.  So, were there two other people sharing the same result?  No.  One of the other pairs in our direction managed to go to bed with two aces defending 3NT!

Was my bid really so terrible?  My partner thought so.  I'm not so sure.  I would be interested to know if any other players in my seat made it.  If they did, they got away with it, perhaps with LHO bidding 3NT (he held K2).

There were lots of ways that it could win.  First, my partner might not have had that heart ace.  Alternatively, the K might have been on my right.  Or perhaps they would eschew notrump knowing they had only one diamond stopper, bid 5♣ and be beaten by our two aces and diamond ruff.  Or, as mentioned above, LHO might simply bid 3NT and we'd be off the hook and partner would get off to a good opening lead.  Despite our two aces and obvious diamond ruff, there were two 480s the other way and one 490 (as well as the 520 already mentioned).

The most dangerous bids are the ones where RHO is limited in some way and where LHO's double is for penalties.  This setting allows LHO to make an accurate judgment.  But in this case both LHO and RHO were unlimited and, moreover, only one of them had thus far bid a real suit.  A small slam, or even a grand slam, were still very real possibilities at the time of my bid.

That's my story, and I'm sticking to it.

A den of iniquity?

Oh the iniquities (or should that be inequities) of the ACBL matchpoint scoring system!  I've been thinking for a long while now how silly it is that placings can often be decided by fractions of a matchpoint.

Because a table in Saturday's A/X pairs at Cromwell managed to report a strange result on a board (one pair was given a "no-play"), the factoring process kicked in for that one board.  This robbed Peter Matthews and I of a 2nd place tie and a whopping 2.2 masterpoints!  On the board in question, we were +100 for 7.63 matchpoints.  Our rivals were +80 for 4.92.  In other words, on that board we beat them by 2.71 mps.  For the other 51 boards, their score was 352.5 while ours was 349.5 mps.

Now, let's look at the effect of the various scores at the other table on this board:

scoretheir mps our mps their total our total our position
50 or lower5.58358357.53rd
8058357.5357.5tied 2/3
904.58357357.52nd
1004.57.5357357tied 2/3
110 or higher4.57357356.53rd

So in two out of three scenarios, we would be third, in one scenario would we be second, while in the other two scenarios, we would be tied.  But when it is so close, why don't they simply admit that the two pairs can't be reliably separated and call them both 2nd equal?  Then they could split the masterpoint awards by an appropriate ratio.  The awards in this case were 17.59 for 2nd place and 13.19 for 3rd place.  If there was a dead tie, each pair would get 15.39 points.  A more equitable sharing of masterpoint awards for the circumstances described here would be 15.83 and 14.95 (this assumes, rather arbitrarily, that all five scenarios are equally likely).

In actual fact, we do know the result at the other table because the pair sitting in the opposite direction were credited with -100.  Had that result stood for both pairs at the table, we would have been tied for second.  I don't know why the pair in our direction was given a no-play but presumably they were somehow damaged by the events at the table in a way that couldn't be redressed.  At the same time, their opponents were given the table result which was, for them, below average (37.5%).  And by the way, the pair that got the "NP" result were well out of the money and so any result for them would have made no difference at all.

Ah well, it's the bridge that's the thing, not the masterpoints!  Who cares about masterpoints?  But it does seem silly that placings should be determined by such uncertainties.

Wednesday, February 3, 2010

DSIP when we know we have a fit

The question has recently come up: which is better when we know we have at least an eight-card fit?  Penalty doubles (as recommended by Robson-Segal and as generally played by most players) or two-way (DSIP) doubles?

You won't be surprised to hear that I believe two-way doubles are best, although generally I agree with everything that Robson-Segal say in their excellent book Partnership Bidding: the Contested Auction.  But rather than simply go on gut feel, I decided to put it to the (theoretical) test.  I developed a spreadsheet which examines the question in an objective manner (I hope).

The result is that DSIP is better by an average of about half an imp or so.  I'll try to show the pertinent results here:



Unfortunately, Google Docs can't handle the spreadsheet but let me know if you'd like a copy.  Here's the basic idea:
  • Scenario: Equal vulnerability. S (dealer) has a reasonably balanced hand with five spades and opens 1♠, N responds 2♠, E bids 3♣
  • South's options are: pass, or double (penalty in R-S case, two-way in DSIP case);  minimum point counts are S: 12, N: 6.
  • North may or may not have four spades.  His options are pass or double (same as South).
There are several observations to make.  Firstly, we all agree that when a 9-card fit is known, doubles should be for penalty (9-card fits are frequently diagnosed with an initial jump, otherwise if a two-way double occurs, then the partner with extra length will go back to the known fit suit, and thereafter doubles will be for penalty).

Secondly, if either North or South had a distributional hand, there would be no problem (especially if the partnership is playing good-bad 2NT).

Thirdly, these doubles are not game tries per se (although the "maximal" double would be a game try if RHO had bid 3).  The primary purpose of these doubles is to compete, despite not having a distributional hand.  Nevertheless, since they show extras, they will result in a game being bid sometimes.  In the spreadsheet, this occurs once, where N/S are both short in clubs and both have extras.

The essential premise of the DSIP method is that you can compete or penalize when either player has both extras and shortness in the enemy suit.  I have defined extras as having a Queen over the minimum.  Shortness is typically a small doubleton.  When playing penalty doubles, either player has to have both extras and a "stack" in the enemy suit.  The player sitting over the length might be able to make a penalty double with a stack but no extras but the player under the length may not feel comfortable doubling even with extras.  In three cases (two where North makes a penalty double and one where he passes his partner's two-way double) I have adjusted for this by saying that half the time he will do something else that is not so lucrative.

Finally, an example (a minor modification of a hand from last night's IMP game).  Vulnerable vs not, you pick up in fourth seat:  ♠AQ42 AQ543 K8 ♣62.  After three passes, you open 1 and partner bids 2.  RHO now comes in with 3 ♣.  You have shortness in clubs, a K extra and so you double.  You weren't thinking of bidding 3♠, were you?  Partner has made a slightly conservative raise with ♠JT9 JT76 AQ97 ♣Q9.  Still, he has a full 8 count outside clubs (and it looks like the opponents have a 9-card fit) so he bids 4.  This is good for 3.86 imps.

What it comes down to is this: which is the more likely holding in the enemy suit?  Shortness or a stack?