[Tomi2]

in a recent played training match the aim was to look for the form of some pairs to help a NPC selecting his optimal team of a grup of some pairs. several sessions of 16 boards were played - so far so good.

The NPC wanted to avoid boring deals so he did following:
he duplicated two sets of boards for each segment. then looked at board board #1s and took the more interesting. then he looked for the more interesting #2 etc. so the more boringboards were dropped.
Of course he is running the match, so he may use the boards he wants to use, but my question:

will the results provide you a FAIR ranking between the pairs? and 2nd question, if not, how could pairs adjust their methods/style playing in such a contest to improve the results?

I would think of two situations, where the fact the boards are selected could be important:
A: xx - xx - xxx - AKQxxx, w/w, partner opens 1NT (15-17), they pass

probably most will agree that this is a 3NT bid. Either the clubs are running or not. Even two small might be enough. you won't find out better - every bid you make gives them a chance to find a better defense. 5Clubs seems far away and a slam too

now lets imagine partner has all suits stopped and 3 easy tricks and clubs are running for 6 tricks and are not blocking... don't you think the other board, the NPC has refused to play, was more intersting?
isn't it far more likely that something is not breaking etc. than normal?
say one pair bids only 3 clubs here and scores 110 for +5 imps against the field that was -2 in 3NT. Normally you expect such a pair to win those 5 imps maybe 3 times out of 10 and 7 times they lose 6. in this tourney they might be favorites to win imps on that board. and if you know in advance, shouldnt you adjust your bidding / play?


B: you have a random 9 count, w/r 3rd seat...
normaly there is some x% that the deal gets passed out, y% that they have some partscore and z% that 4th hand has a good+ hand and they might hae game/slam
those percentage must change now, since the pass out would have been sorted out and the partscore maybe as well. if you think x=y=z = 33% on random hands then the pass-out percentage should change to 11%, that is the case both possible boards start with pass pass pass AND lead to a passout. to make it a partscore hand both boards must provide no game, that is 1-(2/3)^2... so game+ with opps will become 55%.
players who like to psyche in 3rd seat, open very light etc. to make life difficult for opps will more often be rewarded now.

any opinions on a competition like this?

[barmar]

I think it's kind of like the bidding contests that are in bridge magazines (e.g. Challenge the Champs). Usually the players bid like they normally would, but once in a while they do something unusual because they try to work out why the hand was included in the contest.

But in general, I think it can be difficult to know exactly what the criteria the organizer used to decide whether a hand is "interesting" or not. It's a game of bluff-and-double-bluff trying to work out what you should do because the hand was selected. Maybe he'll occasionally leave in some normal hands, just to confound players who try this logic.

[Hanoi5]

Probably the organizer shouldn't have said anything.

[barmar]

Hanoi5, on 2012-April-08, 12:18, said:

Probably the organizer shouldn't have said anything.

Right -- if the players don't know the game is rigged, they'll play normally and you should get useful info.

[mich-b]

Something like you describe is sometimes done in Israel, when deals are prepared for practice matches of national teams - most often juniors.
When selecting the more "interesting" deals , usually the organizer does not get into deep analysis , and just looks for hands where there is some distribution. So what often happens is that someone has a 4♠ opening , Passed out , and the hand turns out "boring" after all , when some 1NTs do require delicate play.
Having said that , this method of selecting hands is ok if practice time is limited , as long as everybody understand thats only a practice match.
I would suggest not using the results of this directly for selection. Though it's ok if the NPC just looks of how everybody plays, which mistakes have been made, and uses his observations in selection process. With this setting in mind , I would suggest to the players to forget the hand selection method , and play their best normal bridge to leave a good impression on the selector.
When this method is used around here - that works well , everybody plays normally and seriously , though results are not really recorded anywhere.
One last point : I would suggest not to do this too often , because players' "feel" for distributions may get skewed - they get used to wild distributions and expect them all the time...

[paulg]

As someone who does arrange practice matches and sessions for international pairs, I've always taken the approach of finding interesting sets of boards rather than interesting hands. So the pairs will play a set of hands from international competition and they can compare their performance with those that they consider peers (and I would consider superiors ).

This approach does mean some dead, or less interesting, boards but there is no suggestion that there is anything special about any particular hand. Perhaps the only inference the players can make is that they might get more competitive decisions than normal, but that is not a bad thing.

One of my friends, a well-known author, provides excellent write ups for his international pairs of the hands he selects for them. However his attempt to use par contest hands was an abject failure when none of the first team got any of them right!

[gnasher]

If it's a trial, it should use random hands. Removing dull hands is unfair on pairs who do well on dull hands. Suppose, for example, you remove a deal because NS have a combined 22-count and can make exactly 8 tricks in 1NT. That unfairly penalises the maestro who was going to bid 3NT and then make it. And it unfairly favours the idiot who was going to bid 3NT and go one down, or the other idiot who was going to overcall at the two-level on K9xxx and go for 500.

If it's a test of particular skills, it's sensible to provide set hands that test those skills, but you should be open about what you're testing.

If it's a practice match and you want to make it interesting, I like PaulG's approach. In fact, I've used one of his sets of hands and I thought it was excellent.

[barmar]

Agree with gnasher. The organizer's ideas of "interesting" and "boring" may not reflect the skills of the players.

[phil_20686]

Were you to produce "interesting hands" the only fair way to do it, to my mind, is simply to remove all the boring games. There a wide class of hands where ten tricks are cold in 3N, or 11 tricks in a major, and the defence might fight over an over trick, but by and large in any sensible field they will be flat at imps. If you remove these you increase the chance for players to display their skill without unduly influencing the hands.

This is similar to picking interesting sets from international matches, we think of them as interesting if there are few boring games!

As a rule its wrong to remove part score boards. There are no boring partscore boards imo, and it is an important part of how a pair builds pressure in a long match.