Tuesday, March 26, 2013

Never - ever - say die

In looking at my posts and drafts, I am a bit shocked to see that I haven't published in three weeks! Apologies to my regular readers. As always, I have many inchoate articles but none have seemed ready. So, at the risk of being accused of too much self-congratulation or just too much fluff, I offer this reminder that we can turn things around -- especially in the relative randomness of a robot tournament.

This is something of a follow-up to my earlier Never say die article from last year and another in my series on robot bridge (see previous posting). The basic premise is that, no matter how bad things seem in a session of bridge, you should never (ever) give up.

Admittedly, this is a far cry from doing well in a real-person two-session event, but the message applies even to a lowly 12-board robot tournament on Bridge Base Online.

I like to think of myself as not being a quitter, but there are times when, having invested only a dollar and a few minutes of my time, I am tempted to give up on one of these tournaments if I start badly. But I don't give up. The only time I give up is if I make a mis-mouse which is unfortunately all-too-frequent with my MacMouse (I've mentioned this here before). If, instead, I simply make a bad play or I take a losing gamble, I simply resolve to do better in the remaining boards.

The other day, I played such a tournament (matchpoints), and my early efforts did not augur well. On the first board, I was slated to go down 2 in 1NT but managed hold myself to down 3. That was a zero, as everyone else was able to take more than four tricks. On board 2, we failed to compete to the two-level on our vulnerable 7-card fit and allowed the opponents to properly take their 9 tricks in diamonds (36%). Then in 3NT, I didn't appreciate the likelihood of a K being either singleton (a "kingleton?") or doubleton when I was missing six cards to the KJ in the suit (30%). Then, my opponents defended accurately in another 1NT hand and I went down 2 for -200. Half the field got softer defense and only went down one (23%). So, after four boards I had a rollicking 22.16%. Was this an appropriate time to quit? Hell, no!

I came back with a 64%, 90%, 98%, 60%, 92%, 63%, 88% and finally 82%. These scores were good enough to get me up to 60.4% and a fourth-place finish.

I was particularly happy with this hand, which earned the 98%, especially because I so often don't take advantage of potential ruffs in dummy:

After the switch to the Q, I sensed something a little fishy, like a 1-3 split but it seemed that no harm could come of another round of trumps (plus, these Robots can be tricky). The auction suggested that my RHO wouldn't be able to overruff any club or diamond ruffs I might want to take. So I resisted my normal urge to wipe the enemy trumps out, and set about doing some ruffing. I didn't have a lot of hope for the the diamond finesse, given the auction, but it worked!

Twelve tricks are cold on the deal, but of the other 21 heart contract declarers, only two managed to take all the tricks, and one of those wasn't even in game!

BTW, did you spot that the robot sitting East tried to confuse me by playing his 9 and 7 before the 4? The robots like to do that in case there's a Rabbit at the wheel who might forget about the lowly 4. In this case, he just made it too easy.

Monday, March 4, 2013

Can you win this robot tournament - solutions

Last time, we looked at several problems in a 12-board robot tournament. Top was 25.

Board 1: ♠AT AK98 Q764 ♣K87. Partner invited game in hearts after his Stayman bid had been doubled by RHO. If you bid 4, you earned 16 or 23 if you found the (doubleton) trump queen (I didn't). If you passed you were below average with 5.5 or 10.5.

Running score after two: 36 (72%).

Board 3: ♠K72 KJ72 ♣AKJ64. You opened 1♣ and partner bid 1. If you were really disciplined and resisted the reverse, to rebid your nice clubs, you earned a healthy 19 (+110), assuming you didn't butcher the play. If you reversed, you received 13 for –50. If you rebid 1NT then you were getting either a top or a bottom depending on your play and the robots' defense.

Running score after six: 92.5 (61.7%).

Board 7: ♠A932 K95 AQT6 ♣KJ. You opened 1NT, partner bid 2♠ (Minor Stayman), you rebid 3 and the robot bid 3NT. If you made a move (I chose 4), you would have ended up with a 24.5 (only one other table bid the cold 6). I reasoned that the KJ were useful cards in clubs (they were) and that the robot wouldn't have bid 2♠ unless slam was reasonable.

I mentioned in the exposition of this little problem that not all humans responded 3 to the Minor Stayman bid. I can only guess that those who bid 3♣ didn't bother to read the explanation of the call. Some simply rebid 3NT, essentially denying a four-card minor. Presumably they didn't read the explanation either. I have some sympathy for these folks. The explanations are often quite inaccurate, at least as far as strength is concerned. But when they say they have a particular shape, that is usually true. [Just today, I opened 2NT, next player bid 4 and my robot doubled; the explanation was "biddable diamonds" – he had a small singleton!]

Running score after seven: 117 (67%).

Board 8:

I probably just got very lucky but I bid 4♣ (splinter) which was worth another 24.5 when 11 tricks were there for the taking. If you did anything else (or if you didn't double at your second turn) you most probably scored a 13.5.

Running score 152 (67.5%).

Board 10: ♠K982 T62 K6 ♣AQ32. East dealt and opened 1. Did you double? If you did, then you won this board in a somewhat unusual way: you wrong-sided the resulting 3NT contract by the opponents. You will have suffered a twinge of dread when it went redouble on your left, but your partner was there with his 1♠ preference bid. Most probably you would have ended up with 18.225 for down one but I got really lucky when declarer went wrong and down two for 23.45 mps.

Board 11:

As you can see, if you took the spade finesse (or you just cashed the top two spades at some point), then you made ten tricks like most other declarers for 14 mps. The danger of a diamond switch was a lot less than it might have been. Somebody managed to take 11 tricks when West got pseudo-squeezed and pitched a spade.

BTW, I always look for opportunities to sow seeds of confusion in the defenders. I therefore took the club "finesse" at trick 3 to suggest to East that his partner might have the Q.

Running score: 189.5 (69%).

Board 12: ♠J943 A76 A8 ♣AQ97. After East opened 1 in third seat, double was the call most likely to succeed. A few bid 1NT which strikes me as being absurd (15 hcp vulnerable, with only Ax in the bid suit?). Some of those got lucky making 120 or even 150 when LHO didn't lead his partner's suit. But most got their just deserts with –100, –200 or even –300.

After your double, BTW, it went 1 by partner and 2 by East. Would you have doubled again? If you did, partner would have competed to 3♣ for down one and a shade over 7 mps. If you were disciplined and passed (or even if you raised to 2), West would compete to 2♠ going down one (12.5 as in my case) or, if you were lucky, two (18.75).

Final score: 202 (67.3%), good enough for 1st place.

I hope you've enjoyed this little quiz. It isn't hard to win a robot duplicate and I don't claim that it reflects any great brilliance on the part of the player. Like any duplicate, it needs luck, some gifts and some reasonable care not to end up in unusual, no-play contracts.

What makes a robot duplicate a particularly good event to use for a bit of fun like this is that the robots are fairly predictable. If there was an alternative play/call that some other human chose, we can see what the robots' reaction(s) would likely have been if you had done the same thing.

Sunday, March 3, 2013

Can you win this robot tournament?

Continuing my series on robot play, I offer this little diversion.

Any time I can get in a normal contract at a robot tournament (i.e. Bridge Base Online, ACBL Robot tournament), I figure I have a good chance for a good score, especially now that the human plays his robot partner's declarations. Those of you who saw my last blog (The hitchhiker's guide to bridge – part 1) will know the kind of thing I mean where we are not in a normal contract.

The trouble is, it's rare when playing with the robots to put together twelve normal contracts. In this tournament, however, I managed to get to what I consider normal contracts on every hand and, with some careful though not brilliant declarer play I managed to come out on top.

I was fortunate that Leo LaSota wasn't playing in this tournament! He just topped 10,000 BBO points. Congratulations, Leo!

There are 26 humans in this event so that top on a board is 25.

Board 1: ♠AT AK98 Q764 ♣K87. You open 1NT and partner bids 2♣, Stayman. RHO doubles and you bid 2. Partner raises to 3. Your call? I should add, BTW, that in all but one of the 26 auctions, the bidding to this point is identical.

On board 2, you end up defending 3♠ for down one. The only odd thing about this is that your robot partner does some strange carding in the diamond suit which results in some human defenders going astray. Through no great brilliance on your part, you manage to snag 20 mps.

Board 3: ♠K72 KJ72 ♣AKJ64. You open 1♣ and partner bids 1. Opps are silent. Your call?

We'll skip over boards 4 and 5 which were reasonably normal major suit game contracts (going down in the second one when North decided that his 9-count with a singleton in my original major suit was worth an invitation), earning 16 and 10.5 mps respectively. And we'll skip the next one (6) too where we defended a normal 3♣ contract earning an 11.

Board 7: ♠A932 K95 AQT6 ♣KJ. You, like almost all of the humans, open 1NT and partner bids 2♠, described as "Minor Stayman 4+♣; 4+; 10+ total points." You dutifully bid 3 (many didn't) and the robot now bids 3NT. Your call?

Board 8:

Your call?

On board 9, you open 1♣ and your robot partner bids 1♠ with five spades to the ten, three hearts to the queen, four diamonds to the K and two small clubs. You struggle (remember the human has to play all hands now) to go down only one but succumb to a two trick set for 10.5 matchpoints. The play for down one which would score 21.5 is, to me at least, not only a double-dummy play but one that I would call triple-dummy. The play is likely to be found only by a computer playing double-dummy.

Board 10: ♠K982 T62 K6 ♣AQ32 (all vulnerable). East deals and opens 1. Your call?
[I originally showed the wrong hand here - my apologies]

Board 11:

Everyone is in the same contract and the lead is always the same. Clearly time to maximize your tricks. What's your general plan? Will you take the ♠ finesse at the risk of going down when they switch to a diamond?

Well, you've managed to lead the field for the last several boards so all you have to do now on the last one is to avoid an average minus or a bottom.

Board 12: ♠J943 A76 A8 ♣AQ97 (we are vulnerable). West deals and East opens 1 in third seat. Your call?

Answers to follow tomorrow.