[1eyedjack]

I find it a common scenario as declarer that
1) I need to predict the location of a particular missing honour, without any fancy opportunities for discovery plays.
2) One opponent has indicated in the bidding that he holds the majority of the points, and for that reason is more likely to hold that card.
3) That same opponent has indicated in the bidding that he holds length in another suit, and is therefore more likely to be short(er) in the critical suit and for that reason less likely to hold that card.

The usual problem that I face here is which influence between 2 and 3 above has the greater effect. Assume that you have no other helpful clues from (eg) opening lead.

No doubt based on individual facts either might be more influential in any particular case. If an opponent has shown 12-14 HCP and there are only 12 missing HCP then, while acknowledging the possibility of a system deviation, it would take some extraordinary shape-showing bidding to overturn the conclusion that the knowledge of the HCP distribution would heavily influence the prediction. Perhaps by contrast if an opponent has shown a 9 card suit, and the missing card is in some other suit, knowledge of HCP division between the opponents would take second place.

In most real life cases you do not have the luxury of such clear-cut ranking of priorities, and I was wondering whether there are any techniques out there to help?

[Mbodell]

I agree this is hard.

Check out this article by Kit Woolsey for a recent example of similar reasoning: http://bridgewinners...flicting-clues/


You are right that the greater % of the points that one opponent has then the greater the chance that they hold a given high card. But since points aren't continuous, there may only be so many ways to arrange the points, so you can think how many hands can I construct that give them 12-14 without the key card, how many can I construct that give them 12-14 with the key card. Visualizing these hands help. It could well be that this lets you realize other things like "Hmm, if he doesn't have the spade K then he needs both the A and K of diamonds for his bid, but if he had those cards is it as likely he would have led a club?" etc. and find other clues that help. Also, if some other place in the hand you either really need a certain card somewhere to make, or else can afford to hedge and want to consider that you are safe if some other card is onside, so assume it is offside, you can place some of the other as yet unseen cards which can again influence your odds and choice of plays.

[JLOGIC]

Distribution is always more important than HCP unless they have showed a specific NT range or something.

Look at it this way, when deciding to overcall would you ever prefer a random Qx to a stiff? Would you ever care if you had Qxx instead of xx? Would a stiff K ever make you more likely to overcall? These random HCP do not matter to people when they're overcalling or bidding in a competitive auction. Sometimes it matters when they have opened, most people won't open an 8 count even if they'd open it with stiff K and an 11 count. But if you are not 100 % sure that they must have a card for their bid, go with the distributional clue/empty spaces.

[rwbarton]

JLOGIC, on 2012-June-27, 01:19, said:

Look at it this way, when deciding to overcall would you ever prefer a random Qx to a stiff? Would you ever care if you had Qxx instead of xx? Would a stiff K ever make you more likely to overcall? These random HCP do not matter to people when they're overcalling or bidding in a competitive auction. Sometimes it matters when they have opened, most people won't open an 8 count even if they'd open it with stiff K and an 11 count. But if you are not 100 % sure that they must have a card for their bid, go with the distributional clue/empty spaces.

But don't forget to crosscheck your picture of their partner's hand with the bidding as well. Usually this will tell you nothing but sometimes advancer's Qxxx vs. xxxx or Kxx vs. xxx could be the difference between pass and a simple raise or between a simple raise and a limit raise.

[SteveMoe]

Perhaps the issue isn't locating a specific card, rather finding the safest (highest probability) line to make your contract. Knowing winners and losers, etc, and accounting for danger hands and avoidance plays often makes finding a specific card less important. I like the approach Rodwell teaches in his recent book: Identify trick "packages" that will complete your contract and read shapes to determine the most likely "package". Whether you think lines of play or trick packages or both, being able to visualize the key card and shape distributions you need are at the core of getting it "right". If what you need is consistent with the evidence, go for it. If not, then reset and find one that is.

I would like to better understand what you mean by no ability to employ discovery. Perhaps there's a way to accomplish what Justin suggests with the existing information plus a few gatherable facts.