[broze]

Vampyr, on 2014-November-19, 08:36, said:

And now that the EBU have started a grading scheme it is obvious how inaccurate rating systems can be.

What about the EBU scheme has demonstrated this?

[mycroft]

uday, the reason is that everyone wants to play pickup with someone slightly better than they are. Which of course is possible to do for everyone, every time :-)

I'm reminded of one of my locals' (on another site) discussion of "levels of 'can't play'" - truly, what I want is someone within one 'can't play' level of me either way, who doesn't think (or at least, won't act like) he's more than one 'can't play' level away from me (again, either way).

[Cthulhu D]

uday, on 2014-November-19, 12:20, said:

And why is *skill* a primary determinant, anyway? Sure, some people want it, others don't care.

Not that there are any good ways out there to measure skill.

Valve have a very good approach for a 5v5 team game that works well. Skill is best measured from bresults obviously.

Quote

It is my belief that what makes someone compatible is blah ( you may have a diff blah and in a perfect world, we'd use your blah for you, mine for me ) where blah is at least a function of niceness, language, willingness to stick around at the table, etc.

When I play at a real club, pard's skill isn't of great importance. Just as my (in)ability to produce a decent backhand doesn't detract from my pard's willingness to play tennis with me.

Indeed, and DOTA2 offers a good template of an approach that streams the userbaser into team rager and team normal people and puts the ragers with each other, meeting this objective.

Its a mostly solved problem - the issue for BBO is the language and system barriers which have not been resolved

[Vampyr]

broze, on 2014-November-19, 13:25, said:

What about the EBU scheme has demonstrated this?

Just looking up people and seeing that their grade relative to another player's grade does not match with how well they play relative to each other.

[broze]

Vampyr, on 2014-November-21, 08:44, said:

Just looking up people and seeing that their grade relative to another player's grade does not match with how well they play relative to each other.

Fair enough - but I would not expect my own impressions of how well people play to be better than that of a rating algorithm. Where I do agree with you is that fluctuations in people's grades seem so large that it is difficult to attach much meaning to them at any one moment.

[wzd3]

navahak, on 2014-November-18, 11:26, said:

... To improve partnership matching for help me find game it might help to have someway to match people based on systems they know about.

This is very much needed.

[helene_t]

wzd3, on 2014-December-20, 15:05, said:

This is very much needed.

We have a subforum here for finding partners. There, you can write a long post describing what you are looking for. It will always be much better than an automated service.

Suppose you could check a number of boxes to indicate which systems you know. Then you can be mathced with someone who knows one of your systems. But if that common system is Acol, SA or Precision, it is useless because of the many variants. SEF and WJ2005 would be OK but most people don't know any such well-defined system.

[OldPlayr]

Skills ranking is a tricky business.

"Help Me Find a Game" could at least match players who play the same bidding system. It would not be that difficult and the approach could range from a simple general bidding system check-box to a more elaborate process using bidding system and other conventions played to try for a better match.

Pairing people who play different bidding systems is just dumb! What are they supposed to do?

I find myself jumping back and forth from tables to "Help Me Find a Game". I get assigned, then check the partner's profile. If is is blank or a system that I do not play, I jump back and get assigned again. This is a waste of time for me, and cannot be too nice for the tables that I jump in and out of without paying a hand.

If I could simply get assigned to only partners who play SAYC, for example, I'd be much happier & willing to take my chances with varying skill levels. At least I'd know what bidding system we were trying to use

[TylerE]

Don't see why it's so hard. If you're unwilling to play with randoms, don't use "help me find a game".

[Vampyr]

OldPlayr, on 2015-January-01, 13:33, said:

If is is blank or a system that I do not play, I jump back and get assigned again.

I don't play on BBO so I don't know for sure, but I think that a blank profile means that they are willing to play whatever you like. Or they might prefer to use a standard BBO convention card rather than trying to reconcile bidding systems sketched in on different people's profiles. I'd give them a chance.

[jillybean]

Vampyr, on 2015-January-04, 22:14, said:

I don't play on BBO so I don't know for sure, but I think that a blank profile means that they are willing to play whatever you like. Or they might prefer to use a standard BBO convention card rather than trying to reconcile bidding systems sketched in on different people's profiles. I'd give them a chance.

Wrong.

Blank profiles most often means the person plays a basic, standard american and probably doesn't see a need to enter information.

[Bbradley62]

Or they're too lazy or unsophisticated to include details. They very well may be playing "house bridge", like my parents, and not even SAYC which includes difficult conventions like Stayman. It's probably not worth hanging around to determine why the profile is blank. But I agree that you shouldn't be using "take me to a table" if you don't want to play under these conditions.

[Bbradley62]

TylerE, on 2015-January-04, 21:30, said:

Don't see why it's so hard. If you're unwilling to play with randoms, don't use "help me find a game".

There's nothing wrong with "help me find a game" as long as your next choice is "list interesting tables" so that you can see tables with empty seats and preview profiles.

[wzd3]

OldPlayr, on 2015-January-01, 13:33, said:

"Help Me Find a Game" could at least match players who play the same bidding system. It would not be that difficult and the approach could range from a simple general bidding system check-box to a more elaborate process using bidding system and other conventions played to try for a better match.

It's really help me find a partner that's needed. Bidding system is a first level, and there are no tools to help. I have to click thru looking, and by the time I look the table is gone. So the jump in and look seems to be the only way. Really guys, this is computer based, it should be able to do that type of stuff for me!! For both bridge club and pairs tournaments.

For the main bridge club - it needs find a partner -then find a game for a pair. The way it works now is lazy, and from the many comments, frustrating. I know you can't deal with personalities, but computers should be good at filtering out bidding system matches. It's not that hard!

[helene_t]

If you play SA you can just pick a random. Should work with SEF also as long as you agree what 2d means.

If you play acol you can go to the acol club.

If you play precision you don't want to play with a random precision player unless you are Chinese.

If you play something exotic the computer wont find a match for you anyway.

System matching might work for wj and Chinese precision but I think a better solution would be to set up clubs for those like the acol club.

[scarletv]

I don't think it is a good idea to have separated areas for each system. It is not the same to play against opponents all playing a similar style or against a variety of styles. We should benefit when we meet different systems and styles at the tables.

When searching for a partner in the MBC it is really hell when you do not play with friends. I nearly gave it up, as it is frustrating.

Just imagine maybe 5 questions allowing multiple answers. And then giving the opportunity to set filters when searching a partner at tables or tourneys/teams for a good match. Nobody should be forced to answer but those that do will find that very helpful I am sure.

1) System combined with a self rating of knowledge (novice, beginner, intermediate, average, advanced, expert)
(SAYC, 2/1, SEF, ACOL, Precision, Polish club, other natural systems, other non-natural-systems)
2) Defense (attitude, count, discards)
3) Social game / competitive game / wishing to learn
4) Preference of gadgets (KISS / average / like it)
5) Available languages

[OldPlayr]

Any small search feature added to 'help me find a game' would be a help over the current completely random partnering.

Even a simple keyword match in profiles would be an improvement. Let me specify 'acol' or 'sayc' and get only partners with search hits in their profiles.

By no means perfect, but better than the nothing that we have today.

People would learn to put keyword in their profile to help, if they used HMFAP.

[barmar]

It's not random. It matches things in your profile that are easy to compare, such as country, skill level, masterpoint level (the number next to your name), whether you're following them.

Systems and conventions are hard to match because there's no standard way to write them, and people put them in different places (when people run out of room in the profile, they sometimes put it in the email field).

[OldPlayr]

What about simply an option to specify "Keyword in profile"? Better than nothing. Once in use, players might even develop the habit of modifying their profiles accordingly.

The current process is clearly not useful.

Certainly better than the useful advice above to "not use it".

[Dinarius]

I don't find the star rating system all that useful.

But, the self-determining 'Skill Level' is worse.

My BBO ideal would be that the software (BBO) would allocate a Skill Level to a player after, say, 100 boards completed.

This would be based on IMPs won and lost, obviously, and would take into account who he/she was playing with and playing against.

The Skill Level would then be adjusted up or down after every X number (say, 50) of boards thereafter. It would give a player something to aim for. Most importantly, it would remove the skill level allocation from the players themselves.

This is probably too complicated to implement. But, it would separate the wheat from the chaff pretty quickly.

There are far too many fake Experts and Advanced players.

D.