[mycroft]

Quote

Experts should be a very small group indeed. [...] I suspect that 1 out of 4 players labelled as expert in the BBO is probably close to or actually really is an expert.

"Results of a recent poll have just come out, and it's confirmed: 90% of bridge players are better than their partners."

I have a rating I'm quite proud of, but doesn't fit into the list BBO provides: "Experts play with me...again". I is not an EEEEXpert (as one of my teachers, a multiple-time Junior Internationalist, would put it), but I do classify myself as Advanced, and that's what I think I am. It may just be that I lack the guts-driving will to be an expert that makes me not an expert.

Frankly, I don't care who I'm playing with, unless I'm playing dank/susieq (I'm going to lose, but I'd rather lose because they beat me, or because of suboptimal play on my part, not because of moronic play by random "expert") - as long as it's a pleasant table.

Hmm, that's a new "special conjugation": My play was suboptimal; yours was wrong; partner's was moronic...

I can enjoy the high-class, stressful game where the slightest mistake will be costly, and I am the weakest player at the table. I can enjoy playing in a room full of beginners, who are just spending half-an-hour or so pushing cards. I do not enjoy people sniping at me, or at their partner, or un-asked-for teaching (especially of me, and especially of me by an opponent!), or ...

Rudeness and griping at partner seem to flourish in a place where ratings could be affected. As the quality of the atmosphere (and the tolerance for everything, from dead-standard Kitchen bridge to forcing Pass and beyond - often at the same table!) is what keeps me at BBO, anything that is likely to degrade it, even if it improves the quality of the bridge that I can play, is a negative in my book.

Not that anybody cares :-)
Michael.
--
"Multiple exclamation points are a sure sign of a diseased mind." - Terry Pratchett (as Rincewind).

[Free]

I'm against a system which rates players. There are a few reasons for that:
- the amount of cheaters will rise, which is not that funny. I enjoy to play against fair people, I don't care if I get a cheater against me now and again, but I would hate it if I'd get a cheater at at least 1 table out of 2.
- to find a rating system which shows exactly what level of bridgeplayer you are is not easy. There will allways be players which are overrated and others which are underrated, so it doesn't make that big of a differense.
- with the ratings, you will get more discrimination, because people can't hide the fact that they suck. Now they still have the possibility to make their rating private, or several intermediate ratings.

And ATP ranking is the rating for playing 1 year of tennis. If they play better on a tourney than last year, they get more points, if they play worse, they lose points. If you want to try a similar system in bridge, go ahead! I won't work because of many reasons...

[bhugi]

I would not support rating system in BBO. It harms the friendly atmosphere there.
Cheating, discriminating, criticizing, without tolerance, etc. All sorts of problems may arise and become a problemr, because people will then take it serious about the points they get for their rating.

Diretors and Hosts are volunteeners who serve us for fun. If the atomosphere is no longer friendly and fun, but being too similiar to the hunting world for live or dead, that will decrease the attractiveness of BBO. =)

Barry Bhugi

[paulhar]

After reading the comments on ratings, I would like to put my two cents in. A while back, I remember reading somewhere that the BB administrators would have implemented ratings if they thought that the only reason people would want them was to check their own improvement, but they were afraid that players would be more concerned about other players' ratings. My solution is simple: simply do not display other people's ratings! There is no incentive to cheat if nobody can see your rating but you.

There have been mentions that setting minimum ratings on tables is bad news. I heartily disagree. Let me digress by stating what happens almost every time I play on Bridge Base. I start at a congenial table, but one of the players (almost always an opponent) must leave. A new player, usually an expert according to profile, clicks on, and within a couple of hands, it becomes obvious to all, and especially to his partner, that bridge as we know it is no longer being played. Then we go through a succession of players, each filling the seat opposite that player for 1 or 2 hands, realizing that they just can't take it, and leaving, usually sending a private message to myself and my partner with the reason.

A minimum rating for a table (which IMHO should be forced to be moderately lower than the table setter's rating - it's hardly fair to only accept players better than yourself) would stop that nonsense. Coming to such a table, there would be an expectation on the level of bridge being played, and we would avoid having a constant string of opponents leaving upset. Since everybody knows their own rating, they would know in advance whether they would be auto-rejected from joining the table. If I was to play with one of my favorite partners with a low rating, I would have no problem setting the minimum to his/her level.

People complain that pros wouldn't play with their clients because it would lower their rating. Nonsense! The client presumably has a low rating (I'm assuming chess-type ratings which go down if you don't achieve your expected performance with the given opponents and partner), so the expectation of the pair is low to begin with. If the pro gets reasonable results with this client, he will have a high ranking to show for it.

Of course, you could always play at an unrated table if you didn't buy that argument. Nevermind the argument that people would always want to play for 'points' - you're not playing for points, and your rating is as likely to decrease as it is to increase, unless you're rapidly improving! So a lot of people might want to play at an unrated table, perhaps if they are tired and don't think they're bringing their best game to the table.

I don't understand the people that say that ratings are based on how long you've played. Master points might be, but chesslike ratings that are based on expected performace are only based on how long you've played to the extent that (1) you improve by playing longer, and (2) playing longer implies playing whth some prior partners again and knowing their style. However, I woud presume the the actual skill of the player is much more a factor than the two stated above.

I would be quite happy to see ratings that only the player can see for the reasons cited above - and would be happy to hear from anybody who can tell me what would be wrong with such a system. Feel free to contact me at paulhar@juno.com.

Thank you for a fine program - Paul Harrington

[sceptic]

Perhaps you could change the rating system t something like

beginner, very little knowledge, but a nice person

beginner, very little knowledge and I have haemeroids so I wont sit at the table very long

intermediate, I would like to play the game seriously and learn but I make stupid mistakes, BUT I am kind to my children and I like people

Advanced, I consider I am reasonable at the game, I make the odd error and I am learning advanced techniques, yet I am not really qualified to teach.

Expert, nice person

Expert, complete jerk

Expert and I only want to play with Experts and above

Expert, I like to play socially, when I want a serious game I can log in another ID with serious games only on it

Expert, I am not really an expert, I just want to ruin it for other people

World Class, i.e nices ones like Rado, Tobice Stanned (to name a few)

and last but not least PRIVATE, well not really sure why anyone wants to be that secretive as no one needs to know who you are anyway

Ratings for playing ability don't really matter, it is better if we give ratings for manners stability and ability to be nice to people

[DrTodd13]

I would be in favor of a ratings system so long as there would be absolutely no way for people to see each other's ratings. I don't care how complicated the ratings system is so long as it is an accurate indicator. I don't have to understand it. Do people really believe that most people understand even the lehman system?

Todd

[McBruce]

Message in Bridge Base Forums, April 2006:

A few years ago, following a discussion on these forums, we decided to implement a rating system with the proviso that only players could ever see their own rating. We allowed tables and even tournaments with minimum and maximum rating levels. It seemed the best way to ensure that the worst aspects of rating systems were avoided.

We are now aware that ratings conscious people are logging on with kibitzer-bots, which keep track of who is at what table and send this information to a server which estimates people's ratings based on which tables they have been spotted at, combined with a look at a sample of the deals they have played recently and their results. Ratings are still kept private on BBO. But on several independant ratings-servers, you can find out the ratings of others, with an accuracy that is so good that there are few significant errors.

I don't think this is what we wanted two years ago. The question is: what do we do about it?

[red dwarf]

i am a member of a couple of other clubs(although i rarely play in either nowadays) that use ratings and the problems are many.

1)ratings seem to encourage "those inclined" to cheat to do so, and a few members have been caught and dismissed from one club for doing so.

2) you get strong partnerships waiting for less experienced members to start a table and join them so as to have an afternoon of "bunny bashing". this is one of the biggest bugbears

3) you are much less likely to have strong players partner weaker players, that just can't be good for the game

4) it is a regular occurance for a strong partnership to sit and wait for over an hour to get any ops. If ratings where to be implemented here i would bet the amount of tables asking for adv+ or exp only tables would drop, people would rather win higher ratings than enjoy competative games

So for the above reasons and probably a few i have forgotten i am against any form of ratings

[Rado]

Hi all friends,
First of all thank you Sceptic for your nice words:-)
I may assure you that neither Stanned, nether me think that we are "world class". Maybe we are good players and we love bridge, but "world class" are those who win or near win at the world stage:-)))))))))))))) and we are happy to have in BBO some of the real winners of the top world live bridge events.
Back to the ratings issue:
I was lucky to discover BBO in its early months after creation. I was twice lucky to discover Mr. Fred Gitelman - a great man and top player, devoting his time and efforts to create and maintain a place for playing bridge with a smile. At those early days of BBO with let's say 100 people logged in we have some conversations with Fred and he pointed that BBO is a place for pleasure, not for competition.

I've tried, I'm trying and I will try always to help the idea for rating-free, pay-free but SMILE piece of the bridge universe

Kind regards
Rado

[mink]

Hi DrTodd,

yes, the problem that you are refering to does exist. However, I really doubt that people would not go for higher ratings because nobody else can see them. They would still go for them! They would do it even without any reason, but the way you propose it, there is a reason: with a low rating you have fewer tables where you could possibly join.

Currently, if my pickup partner makes stupid mistakes in 3 boards in a row, I just think he might be having a bad day. With a rating system I might think I should really leave now because he ruins my rating. I would not like that people think this way.

The only solution to your example problem I can see is you and your partner leave and open another table.

Karl

[paulhar]

I can see I've stirred up a hornet's nest here! I'd like to respond to some of the recent postings. First, let me pose a question: The following wouldn't affect me because I couldn't be that rude, but is it better to have a table with a minimum rating, or to not have any ratings and willy-nilly boot any players not deemed by the host to be competent to play at that table? (I've seen tables listed in a way that suggests that this might have happened!)

In response to the statement, 'can one not trust his friends'? the answer is clearly NO, since I have a friend from real life who plays a wretched game but I would play with because I like him/her. Someone who marked me as a friend is going to be sorely disappointed to play with this person.

In response to the statements, 'nobody would want to play against strong players', and 'and experienced partnership could wait for an hour for a game', I noticed that 'mycult' stated that he likes a game where he's the weakest player. So do I. I'm sure we're not the only two. I don't really care about my own personal rating as much as I care that the people that join my table aren't going to be such that their partners are going to want to leave after one hand. Alternatively, if people are scared, they could start an unrated table with a minimum rating, so that their opponents wouldn't lose points. Besides, if they're that strong, they have a high rating and lower-rated people shouln't lose rating points to them if they play as expected by their own rating! If people don't understand that, they could be educated. An article simple enough to be understand by the non-math minded could be posted about ratings in the articles under 'explore bridge'. Do people read those things? I don't know, but I have. The news window could point them in the right direction the first few days of ratings implementation.

In response to the 'kibitzer bots' to check people's ratings, can't you get an idea of someone's rating by simply playing a few hands with/against them? And really, why should anyone care if someone wants to go to all that work to do that? I could care less if anybody knows my rating. I suggested not being able to see other's ratings to elminate people looking down on lower-rated players. Besides, I would rather have someone spend their time on kibitzer bots than on writing a virus to bring down Bridge Base (seems like the same kind of mentality)

One common complaint is that nobody wants to play with an overrated partner or against underrated opponents. If you can't see the ratings, how can you tell? You assume someone's rating is commensurate with the way they play.

Bunny bashing? I hope that the initial ratings tutorial will explain why that won't work. Firstly, the 'bunnies' will have low ratings if they're truly bad players, so your expected results are pretty good. You have to exceed these excellent results to gain rating points. OK, lets say you CAN exceed these great results. You are now overrated. Being overrated is transitory. As soon as you play in the general pool of players your extra rating points will flow back into the system and your rating will revert to the mean (i.e. YOUR expected rating based on your skill.) If bunny-hopping works, you will have to ALWAYS do it just to MAINTAIN (note, I did not say INCREASE, beacause you can't) your current over-rating! Does an experienced partnership really want to do this all the time just to be overrated when nobody can see the ratings?? Of course, this assumes that you can increase your rating by bunny-hopping, and I'm not convinced that you can. Of course, there is one way to increase your rating - and that is to learn the game! Play more seriously. Read books. Discuss bidding with your favorite partners. The Bridge Base store offers many fine products to help you toward that goal. Pay more attention when you're playing. Not only will it help you, but your future online partners will be happier too.

Speaking of serious games - don't you think that ratings will make for more play where people aren't constantly leaving to stir dinner, check their stock quotes, and do whatever else they do which needlessly slows down the game? If people care at all about their rating, wouldn't they play at a rated table when they wanted to focus on bridge? I have a friend who rarely plays online bridge, calling it 'not real bridge' because people frequently don't play an obvious card for several seconds, and it's because they're not at the table. He claims that marking it a 'fast table' dosen't get the desired result because he does not mind if someone takes a long time with a real problem, just like in a real game. Its those unexplainable slowdowns that irk him. "And frankly", he says, "in self defense, I do something else while playing online bridge. I wouldn't do this if everybody else didn't." Don't you think these avoidable delays would be cut down in a rated game?

In response to "similar to ATP rankings in tennis" - of course not! The ATP ranking is almost as flawed as a ranking system as the ACBL masterpoint system! I'd be happy to point the reasons if anybody asks but that's off-topic. If Lehman ratings work like chess ratings, and I think they do, they would be so far superior to anything that ATP or the ACBL does. Don't get me started on ladders.

I think somebody suggested self-rating your partners and opponents after playing with them. NNNNOOOO!!!! (1) A good American player with little online experience is going to give a fine European player thumbs down for (a) leading low from a doubleton, which must be common practice in some coutnries, ( passing a free bid in a new suit, which is commonly non-forcing, etc. (2) If you don't like someone, you can log on with many different user names and give them a black mark from each one. (3) If you really want to play with somebody, you give them a black marks so that others WON'T want to play with them. If I thought awhile, I could come up with many more reasons, but do I really need more?

Once again, the people who think that strong players won't partner weaker partners need to be educated about the rating system. The weak partner has a low rating because he is weak. The expectation, therefore, of the partnership of those two players, is not anywhere near the expectation for a partnership of two strong players. Of course, the results won't be that good. But they're expected to be not that good! If you don't do even more poorly than the bad expectation, then your rating won't go down. Sure, it won't go up either. If you expect it to go up when you play, you need to be educated again. Your rating should only go up when you exceed your own expectation, which, unless you're constantly improving, will only happen about half the time. The other half, it will go down.

Yes, there may be more rudeness by the players that don't understand ratings. In response to the player who said 'my partner just booted three boards, I'd better leave.', if this is your partner's expectation, then his rating is just gawdawful and you're not losing points by staying. On the other hand, if your parnter isn't such a bad player but has just made the worst play of the century, you can be sure he feels a lot worse about it than you do and he's going to dig in and make sure it doesn't happen again. (Provided he's not spending energy deflecting your caustic comments.) Ah, yes! Maybe ratings will DECREASE rudeness! Yes, you're rude, your partner plays worse for it, and YOUR RATING DECREASES and you can't play at tables you should be entitled to play at. Or, my goodness, no - I have to be NICE to my partner to maintain my rating?? No, that can't be possible...., no not just possible but obvious! Certainly, that will be in the tutorial on ratings... Oh, the flip side, somebody asks? I can be rude to my opponents and get a better rating for it? Won't work. Unless you enjoy people leaving your table and waiting ten minutes for someone else to fill in. Perhaps eventually the BBO administrators will catch up to you and boot you from the site. Maybe when you come in with your new username, you'll realize it doesn't work and play nice. However, I've found that most of the venom is directed at partners. Hardly anybody has ever said anything rude to me as an opponent (except once when I forgot to say BRB and I deserved it!) and usually I only say things to an opponent which might be considered bad by some when he maligns his partner (usually incorrectly) and then usually in private as I don't want to make him look bad to the whole table.

Those of you that oppose ratings can perhaps help me out. I don't play often enough on BBO to establish set games, but would like to come on and play at a competent table. (I think this might be common among the membership of BBO.) I have clicked many people as friends, but when I log on, those that are on are usually already playing. When I try to click onto a table that looks like it might be competent, I am rejected since according to the BBO guidelines, I am Advanced, not having won any national titles and only a few regional ones, and they are looking only for Experts and World Class players. I presume that at a lot of these tables, I might be the best player at the table (judging from what I see from most people that click themselves as Expert) but cannot bring myself to call myself an Expert. At the tables where I am not rejected, the play is often deplorable (mostly by opponents), which people clicking in and out after almost every hand because they can't stand their partners. If I start a table, the people that click into the table haven't bothered to read the comments. I might say "2/1 partner" and get someone playing the Polish Club. Yeah, I know the solution - I need to learn the Polish Club! There are probably a lot of other people that feel the same way. I would be nice to be able to log on, play a dozen hands or so with the same congenial partner and opponents, all of which can make 3NT with nine top tricks and can trump a loser in dummy if need be. Given the lack of ability to put a minimum rating on a table, I don't see a good solution. Is online bridge not right for me?
Paul Harrington (paulhar@juno.com)

[hotShot]

Hi Paul!

Online-Bridge has a very high potential for abuse.

If one could only see his/her own rating, players could be asked to send a screenshot to prove their rating ....

There are lots of ways to cheat. I don't what to make this a cheaters resource, so i won't give examples here.

I ‘m sure you know, that 2 strong players don't necessarily make a good pair. So how do you calculate the strength of a pair.

Do you know how a tournament result is calculated? Have you ever thought about “fair movements” ?

If you play a BBO tournament with 100 pairs and 5 rounds of 2 boards, you and your partner will have played about 5% of the other pairs. You could have the 5 best pairs as opponents, another pair could have the 5 worst pairs playing. You will have 10 results that are average or worse, the others might get a lot of gifts scoring average or better.
Should their ranking be better than yours ?

Creating an offline tournament the TD tries to create a movement that makes you play at least half of the other pairs.

To create a serious ranking one would need to play tournaments with lots of rounds with very few people. Handling divisions where people can move up or down.

Until that can be done, any numerical ranking system will not be accurate.

Have a nice day
hotShot

[badderzboy]

As an 'average' player on BBO, I think ratings would be an unmitigated disaster for two reasons,
(i) People would leave tables with even greater frequency than now to avoid a 'bad' score.
(ii) Snobbish behaviour towards people with low ratings and new players to BBO!

Ratings prove only to boost ego's and not good manners / ethics!

To cheat the telephone gets round all electronic countermeasures but never done it as I play to enjoy not to win by cheating !!!

Steve

[paulhar]

Responding to HotShot's thoughtful post:

"Online bridge has a very high potential for abuse".

Normal bridge doesn't? It's the nature of the game. I've heard anecdotes about people sitting at table 8 N-S in a tournament so that they could see table 7's from two different sections! And having a minimum rating requirement on a table negates the need for someone to send a screenshot of their rating. There are lots of ways to cheat at any sport or game if one is so inclined. Frankly, I don't see any good reason why one would want to cheat - what's the thrill of winning if cheating was required? I'll admit that this is not a universal opinion and there are those who cheat. Why would they be any more likely to cheat to achieve a rating that nobody else can see? The only reason that comes to mind is that they want to be allowed to play at a table that they don't belong at due to their bridge skills. So, they cheat, they become overrated, join a good table, and it becomes obvious to all that they don't belong there. They get ostracized, which is their due for cheating.

While I have thought about the problem of two strong players not necessarily making a strong pair, it would seem pretty onerous to keep track of details necessary to rate pairs, since everybody could potentially have thousands of partnerships from regular to 'click in and play a hand'. There could be a higher expectation given to a pair who has played more deals together. I'll be the first to admit that it would be quite difficult for an online service to determine that my wife and I have played thousands of deals as partners when we've played only a handful online. This is a tough problem - there is proably some partial solution but I can see that this might be the bane of accurate bridge ratings.

About the strongest five pairs and the weakest five pairs: Let's look at a chess tournament. My friend and I enter a tournament (on the Swiss System) and I lose the first two games and play patsies and win the next three. My friend wins the first three, with the competition getting tougher each time, but then he is far overmatched in the last two games playing against masters and loses the last two. We have identical 3-2 scores and thus place similarly but when the USCF rating is computed, my friend gets a much higher rating that I do, having scored that 3-2 record against much tougher competition. Each game is taken into account - his expected chance of winning, and the rating of his opponent.

Bridge players seem to keep making the mistake of equating 'rating' with 'masterpoints'. They aren't even close. Even forgetting that masterpoints are an attendance award, you could win a club game in East Podunck where the strongest player learned bridge last year and is teaching the rest of the town what she knows, or you could play at a same sized game in a strong field. Winning either will give you the same number of masterpoints. And of course, there's events restricted by things having nothing to do with bridge ability. Others mention ATP rankings, also moderately an attendance award. However, somebody does as well by beating some schlepp who beat Andy Roddick as by beating Andy Roddick. With a chesslike rating structure, that wouldn't be true. Your rating is based on your actual opponents and whether you beat the expected results based on your rating and your opponent's rating.

The same should be true for tournaments and rated tables. It would be conceivable to win the tournament and lose rating points - if you and partner were highly rated and happened to draw five novice pairs as opponents and didn't trounce them as badly as expected by the difference in rating. The other poor pair who played the five strongest pairs in the field would have their ratings adjusted based on how they did compared to how they would be expected to do versus those five strong pairs. Forget the masterpoint mentality! Ratings should be entirely based on expected results against your current opponents. (A really awesome rating system would also take into account the ratings of the N/S pairs you competed against as opposed to the E/W pairs that they competed against - but that would probably only significantly matter in a team game with one other team. When the board is played 16 times, this effect is negligable.)

The current tournament format may be unfair as to determining who wins the tournament, but shouldn't be unfair in calculating ratings. A cool way of fixing the unfair tournament format would to pair each round (trivial for a computer) using the Swiss system - pairing the leading pair against #2, #3 vs #4, etc. Also good because eventually you play pairs at your own level. The downside to Swiss parings: With the current pairings, anybody can win. Everybody knows that and the top players who care can go find 8 of themselves and play a team game and eschew the tournament. But it might be good for online bridge that anybody can win. So, while it's not FAIR, it might be best. Does anybody really care that they can be beaten by inferior pairs? Judging by the number of entries in the Total Points tourneys, not many do care. These tournaments are a lark - they're pure entertainment - and there's nothing wrong with that. But they can be rated a lot more fairly than you give credit. I would assume that most of the rated deals would occur in the Main Bridge Club rather than in the tournaments.

With all the great minds we have playing on/contributing to BBO, I have no doubt that the logistics of a fair rating system could be worked out. If there is a strong consensus that it would increase cheating and boorish behavior, I'd be quite willing to jump to the other side and figure out some other way to avoid the problems I've mentioned on earlier posts. I just can't see how cheating would be increased to attain a rating that only the user can see. Are there really masochists out there who want to play at a table where they're not wanted when they can just as easliy find a table where they would be welcome? Maybe it's just me - but I'd like to play at a table where the other three participants are happy that I'm there.

By the way, if anyone wants to devise fair ratings and/or fair movements, I've done a fair amount of work on both and I'd be happy to share ideas. paulhar@juno.com

[hrothgar]

Hi Paul

I think that I agree with your basic premise: A rating system that was completely private probably would not encourage cheating.

With this said and done: I spent a fair amount of time trying to come up with different types of rating systems. I eventually concluded that it is not possible to develop a rating scheme that

(a) Can accurately rank pairs (or players)
( Can be understood by average players

I have come to believe that the hassle of administering the a system and explaining to end users why their ratings are, indeed, accurate far outweight any perceived benefits of implementing a system.

[sceptic]

Just a quick idea for a ratings system, based on two ratings.

Learner, has not quite mastered the triple flip flop, restrictive choice multiple squeeze

Arroagnt so and so who knows everything

or just make friends and judge people as you find them, mark them as friend and insert comments, i.e. potential partners or decent opponents, good fun player, etc, anyone you dont like you can mark as an enemy.

Anyone that comes up with a site that good probably does not need a rating system.

Alternatively design a bridgeplaying site yourself, make it free and see if you can create and administer a rating system

Good luck with the later, I kinda like the rating system, but then I dont think I suffer from delusions of grandeur, nor do I have an over inflated opinion of my own abilities, nor am I so elietist that I want to pronounce to all how good or bad I am, play bridge with me cos you enjoy bridge or my company, either is good for me

[paulhar]

Reply to brothgar:

(1) Do people understand the USCF rating system? They still happily play and 'believe' the ratings which appear to go up when they play well and go down when they play badly.

(2) Is it that hard really? Four people come to a table with ratings that approximately reflect how well they play. The N/S pair's two ratings are added together, as are the E/W pair's ratings, and the difference taken. This difference can predict the expected number of IMP's or matchpoints that should be won per board by the stronger pair. (In a 2-table team game, you must total the whole team.) On each board, the pair that exceeds their expected result gains rating points derived by some tiny percentage of the excess, while the other pair loses the same. (Examples can be given.) Is this really so hard to understand?

Of course, some adjustments might have to be made to counteract the deflation of ratings due to players improving, and perhaps for regular partnerships, but these are minor indeed and need not be explained to make things work.


Reply to sceptic:

Count me as a 'learner'! Everytime I play is a learning experience. Do I want to prove to others how 'good' I am? Not at all - I know how bad I am! Besides, what I'm proposing, with hidden ratings, the showoffs can't even haughtily show their ratings. (I would even support ratings that not even the user can see - but is allowed to start a table defined by his unseen rating, such as 10 below me to 10 above me.) The reason I'm supporting ratings (an unpopular view, it seems) is that when I play, what it is I'd like to learn to play better bridge against competent players, rather than (1) how to take advantage of players that probably would be happier playing with players at their own level, or (2) how to avoid the unpleasantness that goes with a various string of partners coming in and playing with a weak opponent for a hand and leaving, sometimes with a caustic comment about the player's choice of level.

Reply to badderzboy:

As pointed out in an earlier post, someone shouldn't leave a table with a bad partner simply to avoid having their rating demolished, because the partner probably also comes with a bad rating, which means bad results are expected, and anything acheived higher than expectation will increase your rating. It might teach people to be nice to their bad partners - as they will play better that way. Can you imagine - a way to boost your rating is to go play with bad players, and make them feel good about themselves so they play above their normal game?

As for the snobbish behavior toward people with bad ratings or people new to BBO - how in heck would they know? Your rating is known only to you (or perhaps not even that.)

Actually, I'd like to propose a poll. Please do not answer if you don't have experience playing on rated sites. If a predominate number of replies state that badderzboy is right about ratings promoting bad manners, then I'll admit that all my writing about ratings was a big mistake. Of course, the poll will have a huge bias against my position since I think I would get a lot more support if I polled the entire online bridge playing community rather than BBO members, some of which seemed to have left other (perhaps rated) sites in a huff because they found people unfriendly. Frankly, my online experience is: BBO, the MSN Gaming Zone, and Stepbridge (the only rated one of the three), and found the Stepbridge players to be the friendliest of the bunch. I have encountered a higher percentage of rude players here and on the Zone. Most of the rudeness was not directed at me (happily admitting my errors usually deflects rudeness), but it is unpleasant just the same.

Anyway - the poll is as follows:
In your experience, do the players on rated sites have worse manners than those on non-rated sites?

I look forward to seeing the results and will happily report the findings no matter what they are.

[DrTodd13]

Hmm...lots of comments here. People are cheating now in tournaments and at regular tables. Why are they doing this? There doesn't seem to be a reason to me but people are doing it. So, the question is how much worse would this get if there were a ratings system that only a player could see.

I dismiss statements like people will be asked to send screenshots to prove their ratings. How many people would even know how to do this? Of those that do, how many would actually bother? Just don't play with people who would make such a silly request.

With respect to a kibitzer bot(s) that go around and create ratings and store them external to BBO, nothing is stopping this from happening now. I'm not sure that knowing one person's rating makes this process any easier or better.

Somebody also said that all we have to do is to educate people that the ratings system works so that they wouldn't be afraid to play with players of much lower ability. Well, OKB has a ratings system and they have tried to educate people that their system works and accounts for these sorts of issues. Does anybody believe it? I would say very few. It seems like once a rating exists it gets propelled into the public eye. I think some were suggesting that you could limit tables to those above a certain ratings even though you wouldn't be able to see their rating. This sort of creep to public ratings I wouldn't like. The self-rating would still be useful for that purpose.

[hotShot]

Hi Paul!

In a chess match one of the playes is the winner and one is the looser.

In a bridge tournament you do not compeat with the opponent you play, but with the other pairs/players playing the board from the same side.

Your opponents result is compared to all pairs that played the board his direction.

So when you leave, there is no winner and looser, that is why the chess rating can not simply be used for bridge.

Forthemore some hands are simply not selective.

If you have some sort of ranking you might predict if a pair will perform better than another in a tourney, but to expect some sort of IMP-difference per board you can predict is not the way bridge works.

I'm thinking and experimenting with a rating algorithm for some time now and it's not that simple.

Have a nice day
Robert

[paulhar]

Hi Robert!
I didn't see an E-mail for you - it appears you've worked on the problem so I'd like to discuss it with you. I'll start with this post. The discussion of bridge ratings may be moot as far as BBO is concerned, based on the early results of the poll. However, I think it's an interesting exercise in and of itself. So, here goes...

The date: June 2160. Bridge is a professional sport. The finals are between the heavily favored New York BridgeYanks (boo!) and the underdog upstart Florida Marlintrumps (yay!) Both are 4-man teams. The oddsmakers in Vegas are gleeful as they have set the odds well to get equal money bet on both sides. They have made the Yanks a prohibitive 180 IMP favorite in the 120 board final. The happy Marlin bettors are rooting for a flattish set of boards so that their teams lose by less than 180. But in any event, the smart money doesn't see any advantage to either side, figuring that 1.5 IMPs per board is a reasonable expectation.

How was that figure derived? Simply by comapring the ratings! Yeah, it's not that simple. Or is it?

The Yanks' pair A, based on results earned in the past based on common opponents, is, let's say 0.6 IMPs better than Marlin pair C, and 1.1 IMPs better than Marlin's pair D. The Yank's pair B, based on results in the past based on common opponents, is, let's say 0.4 IMPs better than Marlin pair C, and 0.9 IMPs better than Marlin pair D. Another way of saying this, is that if Pair A were to play neutral pair E in a cross-IMP event, they would expect to do 0.6 IMPs per board better due to ability than pair C would. (Logically, in the above example, they would also do 0.2 IMPs per board better than their teammates, but that's irrelavent.) OK, so if A (N/S) plays against C and D (N/S) plays against B, then you COULD say that A has a 1.1 IMP advantage per board against their counterparts playing the same direction (D), and B has an 0.4 IMP advantage per board vs their E/W counterparts ©. So, the team has a 1.5 IMP advantage per board. Surely this won't be true on every board, a flat board will probably be 0 IMPS, whereas a touch-and-go game might be a larger advantage to the stronger Yanks. But the EXPECTED advantage is 1.5 IMPs per board. Or, alternatively, you could talk about A's 0.9 IMP advantage at THEIR TABLE and B's 0.6 IMP advantage AT THEIR TABLE and come up with the same result. When pairs C and D switch places, the advantages are the same, but the numbers in the above explanation would just switch around.

If these pairs were rated, and the Marlins beat the spread (lost by less than 180), their rating should improve a little, while the Yanks' should decline. If the Yanks win by more than 180 for the match, then their ratings would go up and they should be even more favored next time.

Just as the oddsmakers can handicap a match between the BridgeYanks and the Marlintrumps, certainly there must be some number that you can put on an expectation in a team match consisting of any two pairs against any other two pairs.

Going one step further, isn't crossIMPS just an average of 15 team matches - all 15 being your pair against your table opponents, each of them with one E/W pair as teammates (if you're N/S) and them having a N/S pair as teammates. Each of those 15 mini-matches has an expectation of IMPs advantage to you (possibly negative.) With any kind of decent ratings, this expected advantage would come from taking your advantage over each of your N/S counterparts and taking each of your opponent's E/W counterparts' advantage over your opponents to come up with an expected IMP advantage for that mini-team. (Yeah, I know your teammates are changing every hand! That's immaterial, they're recalculated each time. Computers are fast!) So, you have an expected IMP gain/loss on this board based largely on the difference between your rating and your direct opponents' rating, which will count 15 times, but also on the other competitors you're being teamed up with and against. Beat that expected gain/loss, and your rating goes up some tiny amount. Underperform expectation, and it goes down.

Does it work on every single hand? Of course not! Sometimes you get lucky and are a little bit overrated for a while. But being overrated is transitory, it makes your expectation higher than it should be, and thus less likely to meet, so your rating should revert back to its correct level. Underrated players should correct for the same reason; their lowered expectations should be easy to beat. On average, nobody should be underrated or overrated for long.