Join Bridge Winners
All comments by Ping Hu
You are ignoring the author of this comment. Click to temporarily show the comment.
We could have a slow migration process. ACBLscore is good at managing club pair games. As long as we don't add new movement and change scoring method (the last was 4th undertrick doubled from 200 to 300), it could still be used to run club pair game and produce game files. If the function to calculate award and club financial is put into a new application that only runs on ACBL master server (as I suggested early and Nic commented it was ACBLscore+ design), we just need to upload game file from old ACBLscore and there is no need to upgrade ACBLscore any more. It could still run without internet connection.

We do need an ACBLscore replacement to run tournament and team games where ACBLscore is not very good now.
Jan. 12, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
TD could use existing ACBLscore to print out the estimated award (from the copy that has the bug fixed). However the final award is always calculated and controlled by the ACBL master server.
Jan. 7, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
You don't need internet connection while running the game. Just need to upload the game file after the game or tournament. Most TDs do it now for other reasons like upload the game to website and publish the results.
Jan. 7, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
In addition to simple software testing issue, this also illustrates an software architecture problem very typical to old programs. ACBLscore tried to do everything in one place. With this master point change, it needs to upgrade hundreds of copies of installations.
A better architecture is to calculate final master point award on a ACBL server. All club and tournament TD just upload the game files to the ACBL server. In case a software problem discovered, all it needs to do is to fix the program on ACBL server and recalculate the award. In this way, TD does not have to calculate it by hands.
Jan. 7, 2015
You are ignoring the author of this comment. Click to temporarily show the comment.
Nic,

What I have in mind is some kind of common format to capture per board result. What is the card distribution, what are the bidding records and playing records. Of course it needs to keep information about who played it. These are the common data no matter where it was played or in that event it was played.

ACBLscore packaged a lot of data in its file. Each country might have different data for their tournament. Those data do not need to be in common game file.

PBN/LIN is probably best close to this. XML has too much overhead. It is best to use XML to keep tournament level data.
Dec. 21, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Kevin,

I completely agree with you that we could not expect ACBL taking a lead on this. It should be a bridge-community wide effort. In practice it should be driven by bridge software developers.

I think the first priority is to have a standard game file format. In chess, they have used PGN as game file format early on and all chess applications use it. Bridge does not have many moves as chess. It is limited to 52 cards and 13 rounds of card play. It should not be difficult to come up with a common standard. Once we have a standard, developers who want to do report and analysis software could all use it, and it is possible to build large database.

This should not limited to US only.
Dec. 21, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Nic, this requirement to support different club is not unique to ACBL. As we discussed in Providence, all it needs is ACBL has an central computer to process all game files, calculate master point award and club sanction fees. I say this based on my experience as a chess TD. I could work for different clubs. All I need to do after each game is to upload the game file to USCF website (for TD) with proper information (club ID etc) and USCF computer will calculate rating and how much the sanction fee is.
Dec. 19, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Greg, since you are on the technical committee, I urge you to seriously consider separating out club financial, membership and other features from ACBLscore. I have discussed this with you in private message. If ACBL just use another program to manage its own financial instead of bundle it into a scoring program (like USCF) it would make everyone's life much easier.
While talking about ‘big data’, I think bridge needs to have a game data base like chess while you could find major tournament data. All chess players know that chess game database is a big business. All serious players use it to study how their opponents' play. More frequently they use it to prepare their openings. Bridge needs something to capture all bidding and card play of a game.
Dec. 18, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
As a player who is in early 50s I could discuss some obstacles that players in age 30-50 would face since I just passed this time period.
1. The availability of game. Players in this age group often have regular work and family obligations. They don't have a lot of time to spent on the game. A 3-4 hour bridge game during weekday is not always possible. In the city I lived, on the club that offered evening game Tuesday evening closed out a couple years ago. The best option now is online bridge like Speedball on BBO. The question is how much ACBL would like to promote online game over club.
2. Difficult to find partners. The players in this age group who wants to play are players with some experiences and generally bridge enthusiasts. However since they could not go to a lot of club games and tournament, it is difficult for them to find good partners. From my own experience I quitted playing in ACBL a couple times during past 20 years because of my partner could not play any more or my job change resulted in moving to a different city.

Bridge as a family activities might work for some but I doubt it works for most families. I have posted my experiences trying to teach bridge to my son on another thread. It was not successful. Kids are more likely to play with kids in this modern age.
Dec. 10, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
All these could be automatically handled by an online partnership desk. Once potential partners are found, player will get notified. Once they agreed partnership an online entry for that event will be generated and the name will be automatically removed from partnership desk. ACBL sent me a survey and ask what features should be added to ACBLlive. I made my comments to add those features.
Dec. 8, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
It snowed yesterday evening and overnight, but it was not heavy. It should get better over the weekend. If you rent a car, make sure you get a snow brush. When I picked up my rental car yesterday, it did not have one and I had to ask for it.
Nov. 28, 2014
Ping Hu edited this comment Nov. 28, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
I'd like to suggest we go one step further. We could create an electronic bidding box and connecting it to BridgeMate. The system will work like online bidding on BBO. It would eliminate director calls from bid out of turn, insufficient bid etc. It could also record the time each player spent on bidding so it won't be what you said vs what he said. If a mandatory delay of 10 second is desired it could be designed so no bid could be entered within 10 seconds. It could work with or without screen.
Nov. 24, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
I would bid 4 directly. The card you need are honors, A and K for slam. So even it is minimum, you could still have slam if they are the right cards. The problem is where winners come. It has to be our suit. If is 3-3 it is OK you just need 2 ruff. If it is 4-2 (more likely) or worse you have losers. It is hard to investigate without any special agreement. So I just have to make a call. It is very lucky in this hand you have 2 out of 3 key cards and 6 should make, but the probability for this to happen is <30% I guess.
Nov. 20, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
The best thing to do is to replace STOP card with a 10 second timer.
Nov. 19, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Gary, “good” 12 HCP means at least 14 points. “Good” 9 HCP means 10+ points. So the combined hand has at least 24 points for a two level contract. Meckwell bid and made a lot of 3N with 23 HCP.
Nov. 17, 2014
Ping Hu edited this comment Nov. 17, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
As I mentioned there are downside of wide NT and you need some tools to manage it. One of my tool is opener double of opponent's overcall shows maximum. The other is negative double. Because 1NT takes a lot of space, sometimes you NEED to use opponent's overcall to describe your hand better. This is not unsafe because your partner knows you have a balanced hand. So whatever he bids, he knows you have at least two card support. If opponent happens to bid his suit, he knows you have at least two trumps in defense. So there are some risks associated with bidding at higher level but it is manageable. In these days you could hardly buy your contract at one level, sometimes not even at two level.

The wide range NT does make game try a little bit difficult. This is why I exclude 16 with 5 card suit because it has too much offensive potential. The hand you showed is only 15 but it is probably better than some 16 because of its structure. This leads to hand evaluation. An opening hand has to have certain offensive potential. In no trump you win trick by two ways, high card or long suit. I use the following assumption in my card evaluation: every card that could win a trick worth 3 points. It could be either a high card or a long suit. So you need 21 points for 1N and 24 for 2N etc. High card points are easy to count. Length count are estimated as following: 4 card suit worth 1 point (you have 1/3 chance to develop a trick), 5 card suit worth 4 points. These estimate needs to be adjusted later depending on if partner has support or not.

With above assumptions 1NT open would have a minimum of 14 points and maximum of 18 points with my treatment. In this way responder with good 9 or 10 points could make a game try and still be able to stop at 2NT. The hand you showed had at least 6.5 tricks of offensive potential. It would be outside of my 1NT range. I would upgrade and open 1. If you don't have 10, I would open 1NT.
Nov. 17, 2014
Ping Hu edited this comment Nov. 17, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
I've gone one step further. I play 1NT as 12-16 and 1D as unbalanced hand. 1D open will have a singleton/void somewhere. This is very useful to partner in subsequent bidding. In practice I tried avoid open 1NT with bad 12 and 4333. If it is 16, it would not have a 5 card minor.

I have played this system from club to nationals in the last a couple years with a lot of success. It is a double edge sword. On the positive side, 30% of our opening bid is 1NT and opponent has to compete at 2 level. They sometimes could not find the good defense due to the wide range. On the negative side we need to work hard to find 4-4 major fit. For that purpose we play negative double and modified Stayman response.
Nov. 17, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
This is a great service. Thanks to all team members at bridgewinners.com.
A couple years back there was a reception at summer NABC for BBO players. Is there any possibilities to organize a bridgewinners member reception at Providence? There are a lot of good discussions about bridge in this forum. It would be good to meet people in person.
Nov. 8, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
No. I have two formulas. One for MP rating and the other for IMP rating. A game could be rated for MP only, IMP only or both. It is similar to chess where it has regular rating and quick rating. The difference here is that MP and IMP game needs difference strategies so they may need to be rated separately. One could not simply covert MP game result into IMP game result or vice verse.

The rating calculation only depends the game results and the pair's rating. It does not depend on population. For example in a team game, one team has a pair rating 2500 and the other 2300, the second team has a pair 2000 and 1900. The 2500 is playing against 2000 and 2300 is against 1900. Let's assume for a rating difference of 400 the stronger pair suppose to win an average of 1 IMP per board. So I could calculate and get 2500 against 2000 expected to win 1.2 IMP per board, 2300 against 1900 should win 1 IMP per board. So in this particular match-up, the strong team is expected to win 2.2 IMP per board. If you switch the match-up the expected score could be different. The game results are compared against this expected score and if the strong team win more than their expected score, their rating increase, if it is less, they lose rating. The same is true for the weak team so they just need to not to lose more than 2.2 IMP per board to gain rating. Please note the numbers are hypothetical. The real number needs come from tournament data to determine what score difference is appropriate for a rating difference.

The pair game is similar except one board could be played by a lot of pairs. So this one calculation is done for every pair on that board. Here the population is a factor because the more a board is played it get calculated more. So these board will have more weight comparing with those boards that played only by 2 or 3 pairs. In my calculation the weight increase is not linear so that one bad/lucky board will not skew the rating.
Nov. 7, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Peg, there was a separate thread about bridge rating recently and I have discussed my idea.

You are absolutely correct that bridge is a partnership game so the rating need to be on partnership not on individual. This is one of the key difference of my system from others (OKB or Power Rating). The other difference is that I used per board result and compared with every pairs who played that board rather than a single session result of percentage in MP game.

My test on a local club for two month's games generated 156 players and 128 different partnership. Each of partnership get its rating. For entire ACBL universe, it is going to be huge. However the modern computer technology is capable to handle it.
Nov. 7, 2014
.

Bottom Home Top