[kenfree]

/**In bridge, the most important criterion in partner selection, in my humble opinion, is bidding system. If i am committed to bidding 2/1, I don't want a SAYC or (God forbid) a Precision bidder, but in order to figure out what they bid I have to go their profile, and then I STILL may not know. But even so, there is not time enough for all this navigation when trying to join a table in main bridge, where you only have a few milliseconds to request a seat at an open table. If the player logo (meaning the name as it shows up at a table, or on list of players seeking partners at a tourney) has some kind of visible identification that would allow other players to know on sight whether they bid a particular system or not it would be SO MUCH EASIER, AND FACILITATE MORE SUCCESSFUL PAIRINGS.

In part, I think this problem is related to another problem that I am quite surprised no one else mentions in this forum, but the amount of space allotted to explain one's various bidding preferences and conventions, etc. I strongly suggest that the current approach be scrapped in favor of a much larger profile area, perhaps as a popup, that would have checklists for more common conventions, and a large space for players to write in the more obscure stuff. Really, as great as bbo is, the problem of finding a suitable partner remains a real challenge, and it is a purely technical problem, which could be solved by redesigning the way in which players are able to communicate on their profile more (and more uniform) information about their bidding and carding predilections.

Thanks for listening.

Ken

[mgoetze]

I think you were looking for the "Find a Partner/Teacher!" forum, 2 forums below this one.

[kenfree]

Why would that forum be more appropriate? I am making a constructive suggestion for changing the software, one that is badly needed. I hope I can at least get a reply from Fred on this one.....

mgoetze, on 2011-July-07, 11:43, said:

I think you were looking for the "Find a Partner/Teacher!" forum, 2 forums below this one.

[matmat]

kenfree, on 2011-July-07, 12:45, said:

Why would that forum be more appropriate? I am making a constructive suggestion for changing the software, one that is badly needed. I hope I can at least get a reply from Fred on this one.....

I think what mgoetze is trying to say is that you should look for a suitable (semi)regular partner online instead of trying to play pickup. Between the inadequacy of the system, lack of clarity in skill level explanations, and dishonesty or ignorance of the general BBO player in their self-evaluation and system knowledge it is no surprise that one would find playing with a random partner frustrating.

Your suggestion is not without merit, but there are some problems with it. What about people who play many different systems with many different partners? what about people who can play many systems but have no preference and can adapt? do they *have* to fill out a form? what about someone who doesn't want to play with a person who has as strong feelings about system as you do? I, for instance, really dislike playing with anyone who is a stickler for system.


---
fixed a username reference.

This post has been edited by matmat: 2011-July-09, 08:28

[Vampyr]

I think that someone who wants to play with a pickup partner should expect to use one of BBO's default convention cards. I don't know what these are, probably Acol, SA, 2/1 GF... Precision, Polish Club or others? It might be suitable for a player's profile to display which general systems he is willing to play; the default CC can supply the details. But maybe there should be a few more default cards if popular systems aren't available as such.

[helene_t]

People can specify their favorite systems on their profile if they find it important.

I suppose it could be useful when using the "take me to a table" function to select by system played by partner. But most people who would use that function probably don't play any systems at all, although they may claim to be playing SAYC.

[cmsjw5]

I fully support Kenfree's idea of a visual system-identifier. It would make finding a compatible partner so much easier and quicker.

I find little merit in the counter-arguments:
Matmat's questions ("what about people who . .") are not a problem at all. No-one need be forced to choose a single system, nor be forced to 'publish' it, nor be obliged to take any notice of other players' preferences. The suggestion is simply to provide better information in a convenient and efficient way, and those who choose can use it, others can ignore it.
I would use it, because I have the same problem as Ken. And yes, I do have a regular partner, but he isn't always available when I want to play. And in any case, playing with different people is part of the fun of Bridge (provided you can understand each other).

Helene_t seems to miss the point that checking the profile is too slow - by the time the profile appears the table is often full, as I have found far too often. But I think the point about the 'take me to a table' function is underestimated: I agree that probably only people who don't play (or play all) systems would use it now, - but surely that is because the outcome is random, and therefore useless to those who do play one. If we had a system-id scheme and the 'tmtat' logic used it to match partners, then it would probably be used by far more people, because it would be much more useful.

I don't know how many systems there are, but the obvious suggestion would be to assign a colour to each one and use that for the background colour of the player name in the table list. Possibly a problem for the colour-blind, but even shades of grey would be better than nothing.

[Vampyr]

cmsjw5, on 2011-July-09, 06:05, said:

I don't know how many systems there are, but the obvious suggestion would be to assign a colour to each one and use that for the background colour of the player name in the table list.

This would not work, because better players will be happy with four or five systems. Would you assign a colour to each combination?

[kaltstart]

Of course, one needn´t use such a feature, but I don´t think that you wouldn´t be happy with a feature that only scans the profiles for systems, for example. There are so many brands of systems so that you probably won´t be able to play the system without further discussions.

Just a few examples: Precision seems to be a popular system in Bulgaria. Many of them play it with strong nt (15-17) and polish 2 openers, though. I haven´t met those who play it a la Wei yet. Some write Acol in their profile, yet playing 5 card majors, despite having agreed to play Acol (there is a brand of Acol, using 5443 opening scheme, by the way, too). There are probably lots of different brands of 2/1, well, furthermore if you see 2/1 as approach not as a certain system.

Those problems could be solved if the search is linked to a cc for example that specifies also further bidding. But I doubt that this would work in practice, since people would have to opt for those cc`s, I guess. Another feature, "I´m looking for a game", with a special colour on the nick button seems to be used rarely by people.

[cmsjw5]

Vampyr, on 2011-July-09, 15:48, said:

This would not work, because better players will be happy with four or five systems. Would you assign a colour to each combination?

To say that it would not work is far too sweeping a statement. It would work perfectly well for the vast majority of players with one principal system. And one colour could be used to represent 'any' or 'multiple' if that is a requirement. Even if a player was limited to choosing one system to publish, that is still a major advance over publishing no information at all. The aim is not perfection, just something which is better, and works most of the time.

[Vampyr]

kaltstart, on 2011-July-09, 16:07, said:

There are so many brands of systems so that you probably won´t be able to play the system without further discussions.

...


But I doubt that this would work in practice, since people would have to opt for those cc`s, I guess.

Pickup partners will not be having much discussion. If BBO does not have sufficient default convention cards for, eg, different varieties of Precision, then there could be a way for players to submit cards.

[matmat]

cmsjw5, on 2011-July-09, 16:10, said:

To say that it would not work is far too sweeping a statement. It would work perfectly well for the vast majority of players with one principal system. And one colour could be used to represent 'any' or 'multiple' if that is a requirement. Even if a player was limited to choosing one system to publish, that is still a major advance over publishing no information at all. The aim is not perfection, just something which is better, and works most of the time.

Specifying a general approach is insufficient and trying to institute a system that divides players by some very vague categories is a bad idea, in my opinion.
---
Two experts whose profiles were colored ceramic off-white sit down together, both drawn by the color of their profiles indicating that they play american standard. First board comes up, opponents are silent, 1♦-1♠; 1NT-2♦; all pass
"P! How can you pass a forcing bid?!"
"I'm sorry, i play nmf."
"But two-way new minor forcing is now standard among experts!"
--------------
One of the experts opens a 15-17 NT. lho bids 2s, responder doubles which is passed back to doubler. Well, in his part of america the standard meaning here is penalties, so he passes. The software records a -470 or whatever.
"you're no expert! you idiot! how can you pass my takeout double?!"
"takeout double? no real expert plays that as takeout, the standard meaning is penalties!!!"
--------------
but hey! they were both playing the same system!

[matmat]

Vampyr, on 2011-July-09, 16:16, said:

Pickup partners will not be having much discussion. If BBO does not have sufficient default convention cards for, eg, different varieties of Precision, then there could be a way for players to submit cards.

bbo has a CC system. you can even set your favorite card. presumably it wouldn't be that hard to sort on the general preferred treatment of players that is listed in those, but to mandate that everyone specify what they are willing or capable of playing is, in my view, a bad idea and impossible to implement.

[cmsjw5]

matmat, on 2011-July-09, 16:23, said:

Specifying a general approach is insufficient . . .

Why ? It solves the problem as originally stated - which is simply to find someone who plays the same general system, not exactly the same set of conventions and treatments. Clearly you would use the profile to check the details once the partnership was established.

It is obvious that anything more complex (in terms of matching criteria) would be too ambitious, and would probably return too few hits anyway. But that was not the suggestion. Simple systems are generally best. And to repeat the point, we are only looking for an indication of the system used, not a perfect match-making scheme.

[hrothgar]

I've always wondered whether some combination of classification and clustering algorithms could be applied to this problem.

Imagine the following:

You get presented with a list of 20 or so auctions.
You get to chose your bid playing whatever your preferred methods are

At the end of the exercise, you're presented with a class that you belong to.
You're also told that you're compatible with classes X, Y, and Z and incompatible with classes A, B, C

These sorts of methods get used all the time with online dating sites and the like
There's no real reason why the algorithms that EHarmony and the like use can't be modified for this type of use case...

[helene_t]

hrothgar, on 2011-July-09, 16:57, said:

You get presented with a list of 20 or so auctions.
You get to chose your bid playing whatever your preferred methods are

At the end of the exercise, you're presented with a class that you belong to.

Rather than 20 hands I think it would make sense to base such an analysis of your entire bridgebrowser history.

It is not trivial. If I sometimes open 1NT with 12 points and sometimes with 17 it could be because:
- I am not playing seriously and/or I can't count
- I play variable notrump in most partnerships
- I play strong nt in some partnerships and weak in others

But it would be an interesting excercise. It is quite possible that one could estimate which systems you make an effort to play, how good you are at each of them, and how often you deviate from the system and your skill level (dumping, psyching, playing drunk).

Something else: it would be nice if people could search compatible partners based on objective rating as well as which stock convention cards they are willing to play. If the menu of stock convention cards is too small then maybe an url to an description of the system could work, but people would find themselves orphaned if they chose a different url from the one other players of the same system use.

[hrothgar]

helene_t, on 2011-July-09, 17:26, said:

Rather than 20 hands I think it would make sense to base such an analysis of your entire bridgebrowser history.

It is not trivial. If I sometimes open 1NT with 12 points and sometimes with 17 it could be because:
- I am not playing seriously and/or I can't count
- I play variable notrump in most partnerships
- I play strong nt in some partnerships and weak in others

But it would be an interesting excercise. It is quite possible that one could estimate which systems you make an effort to play, how good you are at each of them, and how often you deviate from the system and your skill level (dumping, psyching, playing drunk).

Something else: it would be nice if people could search compatible partners based on objective rating as well as which stock convention cards they are willing to play. If the menu of stock convention cards is too small then maybe an url to an description of the system could work, but people would find themselves orphaned if they chose a different url from the one other players of the same system use.

I had considered a system like you described that would use your entire history, however, I think that this would run into trouble when people play different systems with different partners.
(You might be able to design a system that could discern this, but lets walk before we run)

Here's my reason to go with 20 questions

1. We can instruct everyone to specify a base bidding system to start and then have one set of questions for Acol, another for WJ2000, a third for Standard American
2. if we chose the right 20 questions, we should be able to get a lot of discriminating power with a small number of questions (for example, present someone with a 3=5=3=2 16 count in first seat and see if they open 1NT or 1H)\
3. I thought that the system would be most useful for new users who didn't have a bidding history (or established partnerships)

[Phil]

If I spell Staymen, er Stayman incorrectly, will people think worse of me?

[matmat]

Phil, on 2011-July-09, 18:59, said:

If I spell Staymen, er Stayman incorrectly, will people think worse of me?

no, they'll just think you're a flower.

[Icon U]

BBO is a great site. If I have a complaint it is that it is awkward to find a "suitable" partner easily (and that so many come and go from the table so quickly during play - presumably for the very same reason).

An abbreviation after your name showing standard (maybe C,D,H,S) and then your systems played would make partner choice from the table list fairly simple. I believe it would also stop so much diving in and out.