Join Bridge Winners
All comments by Ping Hu
1 2 3 4 ... 16 17 18 19 20 21 22 23 24 ... 27 28 29 30
You are ignoring the author of this comment. Click to temporarily show the comment.
ACBL BoD is not like board of directors of a typical company. It plays part of role of CEO like making major policy and rule changes. The CEO position used to called executive secretary. I think it is more appropriate to be called COO. There is a good writing about how ACBL BoD and BoG operates from the following link.
http://kenmonzingo.com/reports/Blakley_Report.htm

Is this the best governing structure going forward?
May 12, 2016
Ping Hu edited this comment May 12, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
Matthew, you started a good discussion. In your statement you said district was once created with balancing of population. However with ACBL's success in marketing game in retiree in the past two decades, the membership is not proportional to population any more.

A few people had raised the question of functionality of district. Maybe we should ask some more fundamental questions. What should be the best organization structure of ACBL? Is district necessary? Could ACBL members elect its governing body(BoD etc.) directly? Most other organizations have direct elections.

We have seen letters from Bob Hammon and Chris Compton expressing their dissatisfaction (and suggestions) over past few months. I'm sure a lot of other players have similar thought. Is it time to consider a major reform in ACBL to make it more responsive to its members?
May 11, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
It would be best to consider metropolitan area in redistricting to cut down travel. Let district boundary to be in rural area instead of state boundary that would cut a metropolitan area in half.
May 10, 2016
Ping Hu edited this comment May 10, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
AlphaGo has 3 major components: a policy neural network, a valuation neural network and a tree search algorithm that connect them together. It could be trained by self playing the game.

We could use it to solve this opening lead problem. A policy network could identify 2, 2 and 2 as most likely candidates to lead. The valuation network would evaluate the hand based on other cards it sees and assign a score to each lead (presumably it needs some Monte Carlo). It could also take into account of what scoring method used in calculation. So a scenario where a hand has 9 or 10 could have a slight different evaluation that could change the order of results.

Comparing with Go, the card distribution in bridge is limited. When you could see two hands the possibility from the other two hands are about 10 million. When you add constraint from bidding information, it reduces this number a lot. Opening lead is most difficult. If you don't have any information, the possibility would be multiply by a factor of 1 billion from above but it is still much less than game of Go. With bidding information (like declarer is at least 5-4 in major here) this number could be greatly reduced. So a computer simulation is workable, but we may need a drastic different computer program - just like AlphaGo comparing to other computer Go program before it.
April 28, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
AlphaGo has defeated Lee Sedol. Would bridge be next for AI to conquer?
April 27, 2016
Ping Hu edited this comment April 27, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
If X is not alerted as not for penalty, what does it suppose to mean? Are we in an upside down world that every bid does not have a natural meaning?
April 11, 2016
Ping Hu edited this comment April 11, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
If this was “a great experiment”, was it ever concluded as successful?
March 27, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
When I first read this appeal yesterday my first thought was an appeal committee made up by directors (that will start next NABC) is likely to make a different ruling than current player committee.

The issue here is whether there was an misinformation. Based on the fact stated, NS agreement is garbage Stayman and it was explained correctly. So there was no misinformation.

North bid 2 without the typical hand for garbage Stayman. That was a bridge judgement. The only possible issue I could see is whether it falls into the category of psyching a conventional response. It is not allowed under GCC. I'm not clear it is allowed under Mid-Chart. Someone with better knowledge could explain.

I think 2 bid here is a good bridge judgement call with no downside risk, unless NS had pre-agreement that they would bid 2 without heart suit in this sequence, there is no misinformation.
March 18, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
I hope everyone following this thread also read the following post.
http://bridgewinners.com/article/view/an-unsigned-letter-to-the-acbl/
March 16, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
I had a similar experience from past NABC. The case was misinformation. My partner opened 1N and RHO bid 2 (alerted as suction). I had 5 diamonds and 3-3 in majors and doubled 2 to show values. It went all pass. The result was 2 X made 5. It turned out their agreement for 2 is either minor or both majors. We called director and he ruled table result stand. We went to appeal and “screening director” thought we don't have a case. It seems the screening director tries to do is just keep the cases out of appeal committee. We went to appeal committee and got an adjusted score.

My concern with current change is that it avoided a group of players to judge other players. It solves one kind of conflict of interest. However it let a group of TDs to judge other TD's ruling and created another kind of conflict interest. How often is it likely a TD would overturn one of his/her colleague's ruling considering most of the cases went to appeal are not exactly back and white, but could go either way?
March 15, 2016
Ping Hu edited this comment March 15, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
You are ignoring the author of this comment. Click to temporarily show the comment.
The assumption of this statistical analysis is flawed.
March 8, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
The 100% way to make sure opponents “see” alert in BBO is to add an Alert acknowledge response box (like the pop-up box for claim) so bidding could only go on once both opponent acknowledge the alert. However it would slow down the game considerably.
Feb. 20, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
Eugene,

You forum has featured article that could stay on top for a long time. Could you implement something like that for partnership desk and allow some requests for long team partnership to remain their longer while other requests (for specific tournament) would expire in a couple weeks (for example)?
Jan. 25, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
This rule only applies to NABC, not regional and sectional tournament. The original version is in Denver meeting note that you could find from the following link.
http://www.acbl.org/acbl-content/wp-content/uploads/2014/02/Denver.pdf
Jan. 23, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
“We are also looking for ways to automatically record all bidding and play into a database.” The solution will have to be an electronic bidding system. Such a solution may bring additional benefit to game, like eliminate directors from insufficient bid and bid out of turn. It could also record the exact time spent on each bid so there would be no dispute about BIP.

One possibility is to add a feature in BridgeMate to allow each player enter their bid. Each player may have one BridgeMate. I asked BridgeMate representative about if they could keep additional game data a couple years ago in a NABC. They said they could but they were waiting for ACBL to define the new game file format. This was when ACBLScore+ was still in place. Perhaps ACBL technology committee should take a lead to define a standard for bridge game file format.
Jan. 20, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
Kevin, you don't need to know A and B exactly. You only need to have a good estimate for A-B (based on distribution and other information you have). That is enough to get a count of total tricks.
Jan. 16, 2016
Ping Hu edited this comment Jan. 16, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
I use an alternative way to estimate Total Tricks. It is based on the difference of offensive and defensive tricks. Let's say your side could take A tricks in offense but only B tricks in defense. The opponent could take 13-B tricks if they declare. So the total tricks are
A + (13-B) = (A-B) + 13.

If you could estimate A-B, you could get total tricks by adding 13. For this hand, West knows partner has 5 hearts. So his side could get 3 more tricks from hearts in offense (assuming opponent 2-2). He could also get a ruff from Diamond. In this layout opponents could get a club ruff if they play the hand. So E-W could also get one more trick from club. So A-B = 5 and total tricks is 18.

It is not easy to calculate this precisely during bidding. Heart could be 3-1. In that case it could easily be 18 total tricks because B is reduced by one. Holding West hand, I would estimate the total trick to be at least 17, most probably 18.
Jan. 13, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
Robin, if Q could win a trick, it could win in both offense or defense. It only depends on position of the K. Yes, you might come up with some cases where opponents could take tricks from and and Q does not help. So you should consider that probability in theory. In practice, I think Q is like to be loser/winner both in offense and defense. If you have extra length in , Q might be an winner in defense but not offense.
Jan. 3, 2016
You are ignoring the author of this comment. Click to temporarily show the comment.
The calculation would be done offline and you don't have to calculate every time you decide if you are going to bid 5 over 4.

The general question is how many total tricks you need to bid 5 over 4. Based on your calculation (it could be done offline) it is 19 (at this vulnerability).

The next question how to determine the number of total tricks this hand has. LOTT is one method that said you need to count number of total trumps from both side. This works sometimes but is known to have issues with distributional hands.

I estimate number of total tricks by estimating the difference of tricks between offense and defense. The logic is like this. Assume you could win X tricks from offense (you bid) and Y tricks from defense (let them play). So opponent would be able to make 13-Y tricks if they play. Then the total trick for this hand is

X + (13 - Y) = 13 + (X-Y)

So if I could estimate X-Y (dif. between offense and defense), adding 13 would give me total tricks for this hand.

If you look at your holding. A is going to be one trick in both offense and defense. Assuming partner has 7 diamonds, your side probably could get 6 more tricks from playing offense than defense (if you assume opponent is 3-1). It is not likely you could get any tricks in or . There might be a remote chance you could ruff a . Overall the best estimate for X-Y is 6. That would give 19 total tricks. So the answer is to BID.

On the other hand if you think partner bid 3 with only 6 cards. There could only be 18 total tricks at most, more likely 17(trump 3-2). Then it is a clear PASS.
Jan. 3, 2016
Ping Hu edited this comment Jan. 4, 2016
1 2 3 4 ... 16 17 18 19 20 21 22 23 24 ... 27 28 29 30
.

Bottom Home Top