Version 1.48f - please post feedback and suggestions here

[1eyedjack]

People don't really know whether the country is 0.7 or 0.4 important to them

 

In which case they could try both options (given the choice), and stick with whichever setting gives, for them, the more accurate result.

 

 

Speaking personally, I think that I would be able to have a fighting stab at it. At least if I knew what other factors are applied and with what default weighting.

[Antrax]

Why on earth not use objective measures, instead? The only thing that might work better is asking people to rate how compatible the partner was, instead of just looking at IMPs won. Other than that, this seems like a computer's job through and through.

[Oceanss]

I don't understand what criteria makes it possible that some player, using more than one nickname, would see others (even regular pards) with just half star with one name, and 3+ stars with another (having same or very alike friends/enemy lists etc).

[helene_t]

I don't understand what criteria makes it possible that some player, using more than one nickname, would see others (even regular pards) with just half star with one name, and 3+ stars with another (having same or very alike friends/enemy lists etc).

Even if you have the same friends with both your nicks, some people may have marked only one of your nicks as a friend.

 

Also, the recent IMP average may be different for the two nicks.

[1eyedjack]

Why on earth not use objective measures, instead?

 

The problem with use of objective measures is that the objective test must itself necessarily be arbitrary. It is fundamentally a subjective decision by BBO what factors and what weighting are applied. How are they better qualified to judge them than the user?

 

I have no objection in principle to objective measures. Nor to subjective measures. Whatever gets the job done best. Compatibility being a fundamentally subjective quality, I just thought that it called for a subjective measure to test it.

 

The use of subjective measures has one big thing going for it: If the user tweaks the parameters and makes matters worse, he only has himself to blame. Nor do you deny the user an opportunity to revert to objective measures just because subjective measures are available. He would always have the opportunity to reset to default settings.

[Antrax]

The way I see it, a compatible partner means one of two things: someone you have an enjoyable game with or someone you do well with. The second is fairly easy to measure and noise (in the form of non-uniform level of the opposition) will cancel itself out over the long run. The former can be addressed by a simple survey in the end, or just measuring how many hands a partnership played together, and whether that partnership played together more than once. So, despite the strangeness of this claim considering the English meaning of the words, I do think this is something that can be objectively measured. Human input can/should come in the form "do you prefer winning or having a good time?", and probably most people would reply "both".

 

If the user tweaks the parameters and makes matters worse, he only has himself to blame

Have you ever worked with customers? In my experience being in the right isn't worth as much as one would expect.

Nor do you deny the user an opportunity to revert to objective measures just because subjective measures are available. He would always have the opportunity to reset to default settings.

Users can just ignore the compatibility rating. Nobody's forcing anyone to play with who the computer finds compatible.

[barmar]

The way I see it, a compatible partner means one of two things: someone you have an enjoyable game with or someone you do well with. The second is fairly easy to measure and noise (in the form of non-uniform level of the opposition) will cancel itself out over the long run. The former can be addressed by a simple survey in the end, or just measuring how many hands a partnership played together, and whether that partnership played together more than once.

The primary goal of the compatibility rating is to help players assess people they have NOT played with before. If you've played together, you're in a much better position to judge how well you enjoyed the experience than we could be, and we already provide the tools to help you keep track of this (marking them as friend/enemy, putting notes in the profile).

Users can just ignore the compatibility rating. Nobody's forcing anyone to play with who the computer finds compatible.

Exactly. Joining Match.com doesn't mean you can't ask friends to set you up on dates.

[1eyedjack]

Joining Match.com doesn't mean you can't ask friends to set you up on dates.

 

I have no knowledge of match.com, but I would be surprised if it did not allow the member some input into their preferences in a partner.

[FM75]

Why on earth not use objective measures, instead? The only thing that might work better is asking people to rate how compatible the partner was, instead of just looking at IMPs won. Other than that, this seems like a computer's job through and through.

 

According to everything that BBO personnel have posted, the rating system is based upon data. They don't have people making judgments on the "starriness" of a pair. That makes it completely "objective" to my way of thinking. The algorithm takes a number of measurements and calculates a number between 0 and 5.

 

They could do it subjectively, by crowd-sourcing, I suppose. They could show a friend list, country, and whatever objective measures they have and ask someone to rate between 0 and 5. That would be subjective.

[Antrax]

The primary goal of the compatibility rating is to help players assess people they have NOT played with before. If you've played together, you're in a much better position to judge how well you enjoyed the experience than we could be, and we already provide the tools to help you keep track of this (marking them as friend/enemy, putting notes in the profile).

This is about how to tune the learning algorithm. It says you and I are great. We play, then you tell it "he's a miserable bastard and I hated every moment of it", so it corrects the weight to not assign people like me to you again.

That's just an easy way to describe, in reality you can just mine HMFAG users to tune the algorithm.

[FM75]

This is about how to tune the learning algorithm. It says you and I are great. We play, then you tell it "he's a miserable bastard and I hated every moment of it", so it corrects the weight to not assign people like me to you again.

That's just an easy way to describe, in reality you can just mine HMFAG users to tune the algorithm.

Now adding an automated learning component is a very fine idea!

[1eyedjack]

This is about how to tune the learning algorithm. It says you and I are great. We play, then you tell it "he's a miserable bastard and I hated every moment of it", so it corrects the weight to not assign people like me to you again.

That's just an easy way to describe, in reality you can just mine HMFAG users to tune the algorithm.

 

And of the myriad components that were used to build up the compatibility rating in the first place for that player, how is the server to know which was fatally maladjusted?

 

I, on the other hand, might be aware that it is the fact that the miserable git would not speak English that caused so much suffering.

 

Who, then, is better placed to correct the algorithm?

 

 

 

[u4eni4ka]

I'm experiencing a strange .. "problem" - when I click on a nick FROM THE CHAT AREA - I see 3 stars for compatibility, when I click on THE SAME nick from the partnership desk I see 1 and a half.....

[Antrax]

And of the myriad components that were used to build up the compatibility rating in the first place for that player, how is the server to know which was fatally maladjusted?

You take the derivative of the error relative to the weight and go in the direction opposite the gradient. This is known as Gradient Descent.

It's not impossible from a technical perspective, it just might not be the best use of the common BBO development efforts.

[barmar]

I'm experiencing a strange .. "problem" - when I click on a nick FROM THE CHAT AREA - I see 3 stars for compatibility, when I click on THE SAME nick from the partnership desk I see 1 and a half.....

 

In the Partnership Desk, we take note of whether it's an ACBL or non-ACBL tourney. If it's ACBL, we compare your ACBL masterpoints; if it's non-ACBL, we compare your BBO masterpoints.

 

Outside of the PD, we consider you more compatible if either of them are similar.

[BriddlesBo]

For those of us who spend most of our time teaching, it would likely diminish our "compatability score" since we can be construed as being TOO critical with all our constant critiques/suggestions/observations at all of the tables we are working with.

[barmar]

For those of us who spend most of our time teaching, it would likely diminish our "compatability score" since we can be construed as being TOO critical with all our constant critiques/suggestions/observations at all of the tables we are working with.

 

OK, now that you've posted this twice (not to mention the duplicate post in the other thread that I deleted), I'm not sure what you mean by this. Compatibility is not a single score for a player, it's a comparison between two players -- how many statistics they have in common. How do you think critical comments affect this?

[mink]

Compatibility for me is composed of several unrelated factors:

 

0. Friendliness: There is no objective way to measure this, so I would stick to the Friend/Enemy lists. As this is already shown, there is no need to factor it into the compatibility.

 

1. Completion rate: This should be a factor, even though it is indicated separately.

 

2. Do we have a language in common? - This would require the possibility for each player to state the languages he is able to use for communication.

 

3. Knowledge of Bidding systems and conventions: There should be a list of well described bidding systems and conventions somewhere in BBO, and each player should be able to mark each system and convention with "unknown", "known", "used before", or "like". The relative amount of matches should be an important factor for compatibility.

 

4. Carding: Similar, but shorter list.

 

5. Declarer play: For each board played as declarer, and for each card played, determine if the player chose the optimal double dummy action or not. Use this to compute a % value "double dummy action chosen". The smaller the absolute value of the difference of this figure is for 2 players, the more compatible they are. I think this should be the most important factor.

 

6. Defending: same as declarer play, but it should be taken into account if the defenders shared common carding methods or not.

 

It could even be implemented for each user to assign a weight to factors 1-6.

 

Factors that should be not used for the compatibility are: Nationality, self-rating, masterpoints, ratio of number of friends/enemies, IP address (as this depends on where I happen to stay currently, or can even be totally meaningless when a satellite connection or a vpn-tunnel is used.)

 

How often the par contract or better was reached should also not be taken into account, because this is to much dependent on partner and opps.

 

Having available the %-Values for declarer play and defending might also help to identify cheaters.

 

Karl

[lycier]

Today I can't upload Lin file in "My Hand Folder",I don't know the reasons,but I can do last month.

[barmar]

Today I can't upload Lin file in "My Hand Folder",I don't know the reasons,but I can do last month.

We haven't changed this. I just tried and it worked.

[1eyedjack]

We haven't changed this. I just tried and it worked.

 

Is there a limit on allocated storage space?

[barmar]

There's no limit on the amount of saved hands. There's one player who has more than 21,000 hands! But it's only 10MB. I remember when that was a lot of disk space (my first hard drive, about 30 years ago, was either 10 or 20MB), but these day it's nothing. All the saved hands only take up about 250MB.

[1eyedjack]

Minor bug. Possibly reported before. In any case you are probably already aware. Anyway, just in case ...

 

My screen is divided into quadrants:

 

Results or who is online, top right.

Last played hand diagram, bottom right

Chat window, bottom left

Everything else, top left.

 

OK, I can move the chat over to the right by dragging the divider to the bottom of the screen, and that would in fact solve the problem. Except for the fact that I want the chat where it is, bottom left.

 

The problem is that the top left quadrant has no vertical scroll bar. Some of the windows displayed within it do (ie the news) but the whole quadrant does not.

 

So, on the main introduction screen, if I want to click on (say) the "Convention Cards" button under "My BBO", and I mean here NOT from the menu bar but from the button in the top left quadrant, the only way that I can get to see the button on which to click is by dragging the chat window divider down. Otherwise it is concealed behind the chat window.

[1eyedjack]

If I play in a tourney and complete the last hand before other tables, then I am unable to chat to tourney, even though chat to tourney is otherwise (ie prior to completion) generally allowed in the tourney settings (and I suspect still available to windows downloaded client players after completion of final hand).

[LisaJB]

I don't think language, location or common friends should be included in compatibility. One of the great things about BBO is playing with people from all over the world. In fact, I preferred finding a table based upon the contracts that people play and how they rate themselves.