Join Bridge Winners
All comments by Ping Hu
1 2 3 4 ... 8 9 10 11 12 13 14 15 16 ... 25 26 27 28
You are ignoring the author of this comment. Click to temporarily show the comment.
I use 2 as multi with weak two in majors. The traditional precision 2 is very infrequent, not a very good of use of this bid. Now it could be incorporate into 1 opening that shows a limited unbalanced hand.
Dec. 28, 2017
Ping Hu edited this comment Dec. 28, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I would assign a score of AVE- to NS.

A similar mistake happened to me in last session of Red Ribbon final five years ago. After first board we were discussing previous board result (how many tricks each side won). I did not notice the board had been changed and picked up my hand and exposed them. When we realized it director was called. He said he never saw this kind of problem. After consulting with other directors he assigned AVE- to us. Despite this board we still finished session top.
Dec. 18, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Jeff, if you read my post about rating, only pair level could be rated. How do you rating individual player? You could not partner with yourself. If each pair maintain its own rating, what does its matter if a player plays with others. Now matter how many player Eric Rodwell play with, the rating for Meckwell is always the same!

I would not rush to a conclusion on something we have not tried. In Commongame we know some lower rated players like this because there is a way for them to get rewards when they were only compared against his/her peer. While in typical open club game they could not get MP.
Dec. 7, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
There are two distinct issues here. One is reward for player's performance. The other is rating that measures player's ability. These are separate issues.

Player should be awarded for their performance no matter what their rating/ability is. To be fair, the tournament should be set up in such a way that players are competing against their peer (same rating within certain range). Lower rated players could play in higher group/flight, but higher rated players should not be permitted to play in lower group/flight.

As how to award players, we could do it either the size of MP, or combined with pigment. For chess tournament that gives cash award, different group/class get different size of award. See the following link to upcoming North American Open.
http://www.chesstour.com/nao17.htm

Currently ACBL uses pigmented MP for different classification of tournament (club, section and regional). We could redefine it to reflect the strength of competition (measured by rating). If we could add more levels of Life Masters (like Ruby, Sapphire and Emerald), why not more types of pigmented points (Emerald, Diamond, Sapphire and Ruby)?
Dec. 6, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Jeff, the problem is how you define “merit of accomplishment”.

If you define it as winning Bermuda Bowl or an NABC championship, most of us probably should stay home not playing. If you agree winning an opening event in regional, a limited event in NABC are also “merit of accomplishment”, what about for a new player winning a 299er event, or just winning a club game? If they could all be considered “accomplishment”, then the question is whether or not they should be awarded, but how they should be awarded.

You did pointed out a valid issue that the current system encourage players not to seek challenge strong player but play in lesser event they had better chance to win. This is because current MP system allows it, not players fault. A proper rating system would prevent it. In chess tournament players have to play in their own rating group or higher. They could not play in a lower group. There are also anti-sandbag rule that prevents player to purposely play badly so they drop their rating and could play in lower group.

Now going back to the question of winning an A/X pair is an “accomplishment” or not. Let consider two different scenario. First, a pair of upstarting young pair or a flight B player win this event. Second, let's assume Meckwell (they probably will never enter such event) or a pair of other seasoned pro win this event. Which one is an “accomplishment”? I think the first case certainly IS and second is not.
Dec. 5, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I'm glad this post started some interesting discussion. It looks like it is in common agreement that the primary award should be based on field of strength. Current MP also adjust for length of match, table counts and restrictions (senior or mixed). They are OK as secondary factors.

A more important question is how to award players based on their individual strength. In order encourage player to play (I believe this is what MP should do) player should be awarded when they perform above their expectations. This means we need a system that could predict what player's expected performance should be. Then their actual performance is better, they should be awarded.

I have previous presented an Elo based rating system for Bridge. It could make such prediction based player's own strength, their opponent and other players played in the same game. In order to be fair, players award should consider players in the same strength group. In Commongame we have divided players based on their strength up to 12 different groups. This is much more than 3 stratum in ACBLscore. In principle we could create awards based on each group. This is more like chess tournament where play and award are based on player's rating classes.

Need to board my flight a couple minutes. Will write more later.
Dec. 4, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
The Swiss team using Bridgemate was a success last night. There were 30 teams with 2 stationary. Players used Bridgemate to enter player numbers and board results just like pair game.

Director used computer program to pull them into ACBLscore. Players did not have to turn in score cards. The new round pairing was published soon after last score got into Bridgemate. They were published using both projector and rack (traditional way) for first 3 rounds. Last round director decided to use only projector to make it even faster.

The only issue slowed down the game was 31 entries was sold but one team was missing at start. Later it found one team had bought two entries. So pairing had to be rearranged to change a 3-way to a head to head and Bridgemate had to be restarted. There were a couple players had questioned about scores during and after game. Director was able to pull out the board results in question right way and computer was correct.

Bridgemate allows director to monitor game progress on every table. When there were slow tables, he was able to go to that table to pull out a board. So even the game started a little late, but last three rounds were all on time.
Nov. 29, 2017
Ping Hu edited this comment Nov. 29, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
What happened was then entries were more than expected. It was expected 60 teams but 86 entries were sold. When the entry was closed, it was already about 10 minutes passed game time. The computer generated pairing soon after that, probably about 1:15. But the wireless connection to the project went down.

Some later account by other players was there was a lady who could not see screen clearly so she tried to press some buttons on projector but ended up crashing it. This was right at the time we were about publish pairing to projector. We tried to reconnect to projector but was not successful. The pairing is on computer screen but we could not publish it. So TD has to revert to manual process to publish pairing on the wall.

Initially we tried to follow the pairing given by the program and did that for top bracket and half of second bracket, but it was slow to do this exact match. So TD went back to old way of randomly pairing within the brackets. However the last two bracket had some three ways and we started with 72 tables, as entries was more than that number we pulled some tables from 4 other section to make 86. It took TD a while to figure out what table could be used for round robin. When the last pairing was put on the wall and game started, it was 1:40pm. The ACBL Director of Operation decided to issue a refund/discount coupon to all players.

I have posted a message early on this thread that we are going to use Bridgemate to enter player numbers early in this thread and advised to come early. However most players were not prepared, so they have to use another laptop to look for player number. Considerable amount of them also did not fill the bracket form (form with team masterpoint). So the selling area was very crowed and added operation difficulties.

I have previously mentioned Bridgemate will be used in Swiss team on Sunday but it was rescheduled for tonight (Tuesday evening).
Nov. 28, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I don't know if they regular post score on early round. So far the winners (Qualified for tomorrow) are
Nickell,
Mikyska,
Brenner,
McCallum,
Collins,
Robbins,
Kaminski

In head to head matches Bruno and Bender are leading their higher seeded opponents. All other higher seeded teams had upper hands.
Nov. 24, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I have a screenshot but don’t know how to upload it here. I’ll try to type them here.

Head to head:
Nickel vs Hubbert
Levine vs Morris
O’Rourke vs Glickman
Onstott vs Reynolds
Lewis vs VanCleeff
Silverman vs Bruno
Kasle vs Bender

4 way in group of two:
Mikyska vs Vinciguerra
King vs Brenner
McCallum vs Doner
Markowitz vs Collins
Robbins vs McDaniel
Bell vs Kaminski
Nov. 24, 2017
Ping Hu edited this comment Nov. 24, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
ACBL will use Bridgemate to do player entries for Monday 1pm bracket KO. If you are going to enter this event, it is best to arrive early and have all player's number available when you buy the entry. You will be directed to use Bridgemate to enter all player numbers after paying entry fee. The information will be used to create bracket.

In addition, Sunday afternoon Swiss team will also use Bridgemate to enter player numbers and scores.
Nov. 24, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Deleted
Nov. 20, 2017
Ping Hu edited this comment Nov. 20, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
It looks like some links are broken (out of date) now. Try the following one.
http://bridgewebs.com/cgi-bin/bwoj/bw.cgi?club=westwood&pid=display_page5
This has most complete information. It even has movement file if you want to use it with ACBLscore.
Nov. 17, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
There was a prior discussion.
http://bridgewinners.com/article/view/web-movements-help/

You could also find the general description from wiki.
https://en.wikipedia.org/wiki/Duplicate_bridge_movements#Web_movements
Nov. 16, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
You could achieve both goals (maximize pairs playing same boards and simple movement) by using web movement.
Nov. 16, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
No. Nothing was alerted. Another piece of information is this was the least board in Swiss team qualifying round (2nd to last round). Based on the results from previous boards you feel your team falls behind in this round.
Oct. 22, 2017
Ping Hu edited this comment Oct. 22, 2017
Go
You are ignoring the author of this comment. Click to temporarily show the comment.
This(play and defense) is where I think DeepMind's new AI learning ability could help that current Bridge Robot fails. My understanding is it has a policy network and a valuation network. The policy network generate a list of “moves” (in bridge it would be a list of possible cards to play), valuation network will calculate and assign a score for each move the likelihood to win. Comparing with GO, bridge has much less “move” to choose from and much less variation of compute (we only have 52 cards to start with and you could see at least half of them after opening lead).

The current Bridge robot must not have good valuation model. A lot of factors must go into it - what cards have been played, what happened in the bidding etc. I don't believe current program has considered all these factors.

What deep learning could do is it could generate millions a similar hands in training, let's say opponent lead 7 with this pattern (suppose opponents was quiet in bidding) it could use its learning to determine what the best card to play.
Oct. 21, 2017
Go
You are ignoring the author of this comment. Click to temporarily show the comment.
I think it is just a matter of time computer is going to beat human in bridge. Three years ago nobody expected computer could beat human players. The most optimistic expectation was it needs at least a decade. Now even world champion could not compete with AlphaGo.

The argument that Bridge needs experience and intuition is same as Go players described their game before AlphaGo. It is clear current computer Bridge programs are not up to the tasks. It doesn't mean something revolutionary like AlphaGo could not.

One big problem is Bridge is bidding. In order for computer to achieve its potential, we could not limit it to use bidding system just to human one. Computer should be able to invent its own bidding system. This could be helpful to human players as well. It is my firm belief that instead of defining what bids are legal and illegal, we should aim at improving game including bidding but not limit the potentials.
Oct. 21, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Ian,

I would like to add miniBridge could also help players to learn declarer play and defense. Since you know how many HCP each player has, both declarer and defender could use that information during playing.

I used miniBridge mainly as a tool to introduce bridge to new players (most of my students are upper elementary to junior high). This year I decided to stay on miniBridge a little longer. I also had my students who had learnt a couple years to go back to miniBridge and practice declarer play.
Oct. 21, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
If you want to get people to play bridge in 15 minutes, try mini-bridge. It skips the whole bidding part that is most difficult to new players.
http://web2.acbl.org/documentLibrary/teachers/Minibridge.pdf
Oct. 20, 2017
1 2 3 4 ... 8 9 10 11 12 13 14 15 16 ... 25 26 27 28
.

Bottom Home Top