Tuesday, April 21, 2009

At the cross roads

The play sometimes progresses to a point where two options remain equally enticing. The evidence to date does not decide the case conclusively. Yet the professional player must take responsibility for his choice. Terence Reese has famously argued "There is always a clue."

For example, an opponent has overcalled 2 Clubs and we need to figure out whether they hold

or

The decision is not entirely logical yet it is not completely random either.

What usually happens after the post-mortem, however, is that the person in the hot seat, their partner and anyone else who knows the full hand concludes that "he should have got that one right."

For now, we are no longer grappling with the issue of assigning appropriate weights to all the subtle inferences. Subconsciously, either in an apologetic attempt at treating bridge as a purely logical pursuit or in an indulgent ego stroke for the capacity of our powers of deduction, we tend to favour the succesful choice. The weights that we struggled to appropriately apply to various factors at the table now appear to be clearer.

There is a term for this phenomenon, hindsight bias, that I came across in this brilliant book. Some links for the author and detail of his work can be found here, here and here.


Meanwhile, back on the subject, try solving a defensive problem





Trick one goes spade queen ducked all round. Trick two is a spade to the king ducked again leaving us on play. Is there a clue or will we eventually succumb to the hindsight bias ?

7 comments:

  1. Obviously there's no clue in partner's spade spot, as you didn't identify it :).

    I'd switch to the heart 10 because it's sexy. Maybe that's the wrong use of Rodwell's energy? But it could be right. Partner has overcalled on rubbish: if declarer has a 12-count, then partner has 5, so there is only room for a queen, and I will make it the heart queen.

    ReplyDelete
  2. I have to say, i'm with Bill 100%

    ReplyDelete
  3. The way a reporter presents a hand does not change the earlier decision about what card to play at the table.

    Partner knows that I will know exactly what all her four spade pips are when my SK wins. Thus if partner plays the lowest spade pip, I switch to CJ, playing partner for K109(x) or K108(x) in clubs. Otherwise, I play H10. I must trust my partner - otherwise it is my error.

    None of the above applies if I know that partner is unaware of the significance of his (sic) spade spots, in which case H10 is the best shot.
    Peter Gill

    ReplyDelete
  4. Clearly declarer has junk and partner is even junkier.

    Switching H10 needs the least from partner (ie HQ, rather than excellent clubs). If partner isn't telling me anything with his spade plays, I do that.

    I was wondering if I could pull anything from shape assumptions, but I don't think so. That is, what is South's likely shape? Can I use that to figure out which honours he is more likely to hold? I figure his most likely shape is 3424, maybe 3433, and the two candidate suits are hearts & clubs, so ... nuffin.

    Maybe looking at all the hands, the neutral umpire will ask "would your partner really overcall 1S with QJ10xx Qxx xx xxx, or is QJ10xx xxx xx K109 the only realistic possibility (that leads to a beat)"? I'm going to call them both rubbish and rate them more or less equally likely.

    ReplyDelete
  5. I am in love with Peter's answer. Why?
    First because it ties with the work I have been doing on focus style and so on. Trust Peter to think of all the relevant cues and not to get lost. :) But also because of his second ahswer. Like Kasparov said when asked why he made a particular move - and others before and after him I think - it smells right. with the KT8x lookig at A9, if HT is not the answer what is?
    That the stuff champion champions are made off.
    Cath

    ReplyDelete
  6. On the topic of biases let me point you to one of my favourite wikipedia pages:

    http://en.wikipedia.org/wiki/List_of_cognitive_biases

    Mike

    ReplyDelete
  7. Interestingly, the club switch doesn't require that much from partner: the KT will be enough, and maybe even K9 if declarer makes a strange play from QT8x on our switch to a club. However, we do need to return the 7, to hopefully discourage a club continuation should declarer duck partner's T(9); partner can then switch to a heart. If declarer wins the club in dummy and plays a diamond, I can eventually win and play a spade and later get to partner's hand with a club. Returning a club pip also avoids an overtrick if partner has barren clubs headed by the king.

    Anyway, that ignores the information from the spade spot, but is interesting nonetheless. Also, what spot would partner return with Qxx/Kxx in the rounded suits? 2nd lowest? Or if we have H JT8x and partner has Kxx/KTx? Qxx/KT8?

    ReplyDelete