[Gerben42]

Quote

Location determines level? Who has time determines level of tourney?

Naturally. A tourney is as strong as it is. If you organize a tourney in a strong bridge area it will be stronger than if it is in a weak bridge area.

Also 2-day events will draw a different public than 1-day events.

Then higher prize money will attract strong players.

And there is the time issue. A tournament 1 week before a national championship is bound to be stronger than one 1 week after a national championship.

In my experience, the "masterpoint scale" has little to no effect on the level of the tournament. OTOH all the effects mentioned above do.

[tooncestdc]

One possible U.S.-centric rule that has some merit:

What are the largest number of ACBL masterpoints you've won in a single event (lean toward the weaker option if you were the weaker player in the partnership)?

0-1: Beginner
1-3: Beginner/Novice
3-6: Novice/Intermediate
6-10: Intermediate
10-20: Intermediate/Advanced
20-40: Advanced/Expert
40+: Expert/World Class

[Elianna]

If you REALLY want something US-centric you could do it by Platinum points:

Beginner: "Points have colors?"
Novice: "What are platinum points?"
Intermediate: "The Red Ribbons give Platinum points, right?"
Advanced: "I'm so good, I have 4.15 platinum points!"
Expert: "I'm too busy playing in KOs with clients to bother with NABC+ events"
World Class: "Points have colors? "

Anyway, on a more serious note, I like the idea of a series of questions, but it would be better if it were run by the program, rather than make a user follow which number they should go to (considering the level of intelligence of many people...).

[DrTodd13]

1. No online ratings scheme should use anything based on tournament attendance.
2. Any good online ratings scheme should have the ability for the rating to go up and down over time just as a consequence of playing online. Something like "you're an expert if you've won such and such kind of a tournament event" is no good because people get senile and lose ability over time.
3. Initially, rating people on a 1 to 10 scale would be tough. What is a 7? Converging on a number for each player and converging on a community understanding of what a 7 is is an iterative process and in the beginning will be subject to a lot of variance. It is the kind of system where the community will evolve to understand roughly what each value means.

[pigpenz]

if we were to do an oline rating we could eithe ruse avg +/-imps per hand or %rating at matchpoints. Bridge Browser does a real nice job of this already.

Another option would be for those who play in pay games online is avg amt of points won per event....that is why i think bbo should give points to people who play in any event, right now points are only available to those who pay$$$

[Winstonm]

Seeing how I started this thread, I think I'd better chime in about ratings - the last thing I'd want to see is anything other than "self" ratings. OKBridge uses their Lehman system and all it causes IMO is problems because everyone is concerned with their rating. No rating is better than this type.

For me, if given the chance, I'd rather sit and play every night against Soloway and Hamman, knowing I'm going to get my brains beat in and end up with a zero rating than to pick on weaker players just to assuage my ego. So for me, ratings based on results are a poor idea.

All I'd like to see is self rating that doesn't bruise egos yet allows a little better description of ability/interest level. I still don't know what an "expert" player would be - to me, expert is someone who is acknowledged to be of that class, not a self-explanation.

Or maybe this is better: if every day for a month straight you can cut into a game with Mike Passel, Mark Lair, and Grant Baze for a quarter a point and not go broke you are an expert. If you are the underdog, you aren't an expert although you might be darned good, even super-advanced.

That's where I am. I've played each of these guys many, many times and have had my share of wins - but day in and day out they are going to beat me because they are better and they are true experts.

When you look at my profile, you see it marked advanced - because in my mind that is what I am - when the likes of Mark and Mike and Grant start calling me to be an add-on on their teams, I'll change that to expert - and I'm still waiting.

Winston

[DrTodd13]

We all know that self-ratings suck. We also know that if you do ratings based on computer-calculated bridge results then people become nasty and the impetus to cheat increases. Therefore, to me it is pretty clear...if you want any form of a ratings system then it should be one where you rate others. This human factor allows people to factor out results over which the player essentially had no control and to differentiate between who is the better player in a particular partnership. Under this system, if you were nasty then your amiability rating would go down. The amiability ratings provides incentive to be nice. Opps can review hands and if you take a weird line of play that happens to work that the opps know isn't the best line then they won't give you as high an ability rating. This is something a system based purely on results could never do.
So, maybe this approach is the least of all evils...one evil including not having a ratings system at all. Lots of sites provide them and people are always trying to invent better ones. They are useful but you need just to find a way to provide as much benefit as possible without the negatives.

[pbleighton]

"Minority view.

I don't see anything wrong with the number of "BBO experts." Who cares? The goal is to have fun and play some bridge, and in their eyes theyre experts. I don't think a rating system where these people get offended because their peers think theyre intermediate (which they likely are) is good for BBO or for the goal of having fun. If you are an expert you figure out who the real experts are pretty quickly. If you enjoy playing with randoms, just leave if you're playing with a partner who sucks yet has expert written and move on to another table."

Agree.

"I do think the ACBL and WBF need rating systems, but I don't think they're for online bridge."

Disagree. The rating system for the ACBL is a fundraising mechanism. As far as the top players go, people know who won and placed in the big events.

I don't fell strongly about this, however. If people like it, they can have it.

Peter

[DrTodd13]

Jlall, on Jan 20 2006, 03:39 PM, said:

Minority view.

I don't see anything wrong with the number of "BBO experts." Who cares? The goal is to have fun and play some bridge, and in their eyes theyre experts. I don't think a rating system where these people get offended because their peers think theyre intermediate (which they likely are) is good for BBO or for the goal of having fun. If you are an expert you figure out who the real experts are pretty quickly. If you enjoy playing with randoms, just leave if you're playing with a partner who sucks yet has expert written and move on to another table. I do think the ACBL and WBF need rating systems, but I don't think they're for online bridge.

If we just want to have fun then do away with ratings altogether. If we are going to have a rating then presumably it is going to be used for something. It is going to be used to find partners and opponents with which we hope to maximize our fun. I'll tell you that my personal experience has been that if you really want to partner with or play against experts you'll be hopping from table to table constantly. This isn't my idea of fun. It is just my view that the fun of more accurately selecting who to play with and against would be greater than the hurt feelings caused by people who aren't rated the way they would like.

[uday]

Quote

The amiability ratings provides incentive to be nice

It is not possible to distinguish between a new member and an old member from a new location with a new name.

Isnt this going to cause an issue with any ratings schemes? Anyone with a bad rating will drop back to newbie status and start over with a new name. A system that makes it easy to avoid these people might make it hard for true newbies to gain their footing.

Maybe it isnt an issue. You wouldnt be able to trust the neutral/new ratings but you'd be able to trust the bad/good ones.


How do you see people rating each other? Simply right click and assign a number from 1..100 in either category ?

[Dwingo]

uday, on Jan 21 2006, 08:13 AM, said:

How do you see people rating each other? Simply right click and assign a number from 1..100 in either category ?

Over such a big population, you would expert the Novice-Beginner-Intemediate-Advanced-Expert-Worldclass spectrum to follow the Normal Distribution Bell Curve.

With people rating others, it is quite likely that you will see more Novices, as people in general are very harsh in criticising their online partners. One bad play and your partner is likely to take it on you with some bad rating.

With novices and beginners also given this right to vote, are they knowledgeable enough to distinguish between Advanced, Expert and WorldClass?

Godwin

[Clinch]

DrTodd13, on Jan 20 2006, 05:58 PM, said:

We all know that self-ratings suck.  We also know that if you do ratings based on computer-calculated bridge results then people become nasty and the impetus to cheat increases. 

I don't *know* either of these things.

Self-ratings, imho, don't suck, but they are meaningless as measures. The motivation of someone who self-designates expert is endlessly debatable, but one clear possibility is that such a player is passing a message - "I do not wish to play with or against novices and/or intermediates". In general, those who designate themselves expert (but are not) tend to forgive their own errors and be scathing about those of partners and opponents. Self-rating does not, for that reason, make for a friendlier game amonst strangers, as one is constantly stumbling over people whose self-judgments differ from one's own.

I also question that independent ratings make people want to cheat. I agree that the desire to improve ratings when one hits the ceiling tends to make people believe that others are cheating around them, or to engage in unsavoury strategies such as annihilating poor opponents to improve ratings and self-esteem, but I don't think ratings themselves increase cheating, except in very few cases. If used judiciously, I have found them to be good ways of finding compatible games.

Peter (Quite good, although I say so myself).
New York, NY.

[DrTodd13]

Quote

Quote

The amiability ratings provides incentive to be nice

It is not possible to distinguish between a new member and an old member from a new location with a new name.

Isnt this going to cause an issue with any ratings schemes? Anyone with a bad rating will drop back to newbie status and start over with a new name. A system that makes it easy to avoid these people might make it hard for true newbies to gain their footing.

Along with a person's rating, you need to see how many times they've been rated . If they've only been rated once or twice then you shouldn't take the rating too seriously. I think the hassle of telling all your friends your new ID will make few people switch IDs frequently but you're right, if they are determined then you can't stop this behavior but I don't think it will be a big deal.

Quote

How do you see people rating each other? Simply right click and assign a number from 1..100 in either category ?

I'm thinking that when you quite playing with (or potentially against) someone that a box would pop up asking if you'd like to rate them. You could right click on their name and change your mind later on but unless you pop the box up I don't know how much participation you'd get in the ratings system.

Quote

With people rating others, it is quite likely that you will see more Novices, as people in general are very harsh in criticising their online partners. One bad play and your partner is likely to take it on you with some bad rating.

With novices and beginners also given this right to vote, are they knowledgeable enough to distinguish between Advanced, Expert and WorldClass?

Under my scheme, if you play a small number of hands with someone and then get pissed off and give them a 1, your rating won't have much weight because you didn't play that many hands with them. At normal tables, I don't think the novice has to rate a worldclass is a problem because I don't think such pairings are very commonplace. In tournaments, you may get this mix of people (assuming you rate opps as well as pd) but in that case the number of boards against any opp is small (see argument above) and a user could always refuse to rate someone if they can't differentiate between adv and wc.

Quote

I don't *know* either of these things.

Self-ratings, imho, don't suck, but they are meaningless as measures. The motivation of someone who self-designates expert is endlessly debatable, but one clear possibility is that such a player is passing a message - "I do not wish to play with or against novices and/or intermediates". In general, those who designate themselves expert (but are not) tend to forgive their own errors and be scathing about those of partners and opponents. Self-rating does not, for that reason, make for a friendlier game amonst strangers, as one is constantly stumbling over people whose self-judgments differ from one's own.

I also question that independent ratings make people want to cheat. I agree that the desire to improve ratings when one hits the ceiling tends to make people believe that others are cheating around them, or to engage in unsavoury strategies such as annihilating poor opponents to improve ratings and self-esteem, but I don't think ratings themselves increase cheating, except in very few cases. If used judiciously, I have found them to be good ways of finding compatible games.

The prima facie meaning of a rating is as a measure of ability. You admit they are meaningless as a measure so I think that pretty much means they suck. Maybe we should start a poll to see how many thing the self-rating is a self-measure of ability or a statement of desire who they want to play against? I don't see why anyone should choose the latter when that is clearly not its intent. Just because people who self-rate expert tend to share traits does not mean people are self-rating that way because they know they share those traits with other people who self-rate that way.

[tooncestdc]

Just a thought, but I think that a "prisoner's dilemma"-type system could be use to make peer-rating work. Imagine the following system:

Each player is asked to click one of the following:
- My partner is better than me
- My partner is about equal skill level than me
- My partner is worse than me

If both people agree to which is the better player, the better player's rating goes up and the lower player's rating goes down (to a degree...if the better player is rated significantly higher, no change).

If both think that their partner is the better player or equal players, their skill levels become closer together.

If both people think they are better than their opponent, both of their skill levels go down.

The idea is to discourage people from trashing their partner as it can come back to bite him... just a sliver of an idea here.

[Al_U_Card]

Look at the results for the last month. It shows results for enough hands that you can tell a lot about the player. You can even check out some of his pards to see how good they seem to be.

[mikeh]

Good topic, and I am not surprised that there appears to be no consensus.

I would like to see an objective rating system if one could be developed that did not come fully-equipped with major drawbacks, but I see little likelihood of that happening.

The 'objective' systems I have seen or read about carry huge ethical problems. Cheating is one of them,and probably the worst, but it is not the only problem.

As for voting opponents and partners, the problem is that bridge skill is not purely subjective.

And looking at results is misleading. When I attemot to assess the strength of an opponent or am kibitizing, it is not the result per board that I look at but the apparent thought process reflected by the bidding and/or the card play. On most hands I have watched between average players, all players, apart from dummy, usually make mistakes in the play, but most of the mistakes either do not matter or cancel out, because the opps give them back.

Give me someone like Ritong or Justin or Frances to comment on another player's skill level, and I would have a lot of confidence in that assessment, because my interaction with them, and their discussions about how they think, persuade me that their opinion counts. But give me the opinion of perhaps 80% of the bridge world (I include real life as well as on-line), and I would pay little attention. The harsh reality is that almost all bridge players believe themselves to be better than the rest of the world thinks they are, and very few players are capable of accurately judging the ability of anyone significantly better than themselves. I hasten to add that I fully recognize that this almost certainly applies to me as well.

The self rating method on BBO is as good as any, at least until inflationary pressure makes the majority of players WC instead of expert

[hotShot]

Al_U_Card, on Jan 23 2006, 10:25 PM, said:

Look at the results for the last month. It shows results for enough hands that you can tell a lot about the player. You can even check out some of his pards to see how good they seem to be.

I disagree, even at world championships someone has to be last and they will have a bad result shown. The average score of the championship loosers will be worse than that of the winner of the "Totaly incompetent Bridge players turnament".

[hotShot]

Let us for a minute sassume we would have a peer rating system. How will it work for me?

Some time ago i met a lady in the MBC. The first board in 4th seat i held 0 HCP but a nice 5card♠ and a void in a minor. She bid 1♠ in second seat over a 1♥ opening. Since it we were white and they red, i bid 4♠ over LHO's 4♥. During the whole game she told me that whatever i play is not bridge ....., she went down one dbled while almost everybody else lost to 4♥ making. The second board was similar, but the tone of her complaints got a little unplesant. After those 2 boards she left, with +12 imp at our side and -12 with opps. She will never play with me again, but both opps invited me privatly to play with them any time.

[DrTodd13]

hotShot, on Jan 23 2006, 02:59 PM, said:

Let us for a minute sassume we would have a peer rating system. How will it work for me?

Some time ago i met a lady in the MBC. The first board in 4th seat i held 0 HCP but a nice 5card♠ and a void in a minor. She bid 1♠ in second seat over a 1♥ opening. Since it we were white and they red, i bid 4♠ over LHO's 4♥. During the whole game she told me that whatever i play is not bridge ....., she went down one dbled while almost everybody else lost to 4♥ making. The second board was similar, but the tone of her complaints got a little unplesant. After those 2 boards she left, with +12 imp at our side and -12 with opps. She will never play with me again, but both opps invited me privatly to play with them any time.

You say you played 5 hands with this lady? Well, let's say the cutoff for a sufficient number of boards to know how good a person is is 50. So, you played 5 hands with her out of a possible 50. Therefore, her opinion would count only 1/10th as much as people who have played 50 hands with you. Moreover, give the opps the ability to rate you as well and you get 2 good ratings and one mediocre one.

I think somebody was suggesting that I was suggesting that people look at results in order to rate people. I am suggesting the exact opposite. Results are largely irrelevant. The quality of the field and how lucky you are can both cause a purely "look at results" approach to fail, not to mention what a pure results-based approach does to human psychology. If we just wanted to look at results then we could automate it and not need human input. If the purpose of a rating system is to find people you might like to play with then how other people view you as a partner is all that's important. So, to people that are very passive and like to sit back and not take sacrifices then your style will be unpleasant to them. I admire skepticism and I'm not 100% sure my approach will work but I very much suspect that it is better than what we have now and puts in place several motivations to get people to do the right thing. If people want to rating system to be useful then they'll have to bite the bullet and say, "ok..we played 5 hands and on most of them he was ok and on the other two it wasn't my style so I'll rate him a 4 or 5." If everyone starts handing out 1's and this becomes a problem in that everyone's ratings are tending towards 1 or 2 then a gentle reminder that the system might actually be useful if people tried to do it unemotionally rather than emotionally might be in order. I tend to think that your somewhat regular partners who must like you otherwise they wouldn't keep playing with you will keep your rating somewhere around where it should be because they know you better and their opinions carry more weight due to number of boards played.