Join Bridge Winners
All comments by Ping Hu
1 2 3 4 ... 7 8 9 10 11 12 13 14 15 ... 25 26 27 28
You are ignoring the author of this comment. Click to temporarily show the comment.
I think it only needs a small change in the procedure to fix this problem. Instead of asking EW to move boards after they finish playing, move them after their home table finishes playing.
Jan. 16, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
@Richard, thanks. I'll relay your message to Jay Whipple who runs Commongame.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
At this time this data is private and we don't have plan to publish it except using them to study.

In the future if ACBL develops a rating system, I could see games could potentially be separated into two types: those who want to be rated, and those who do not. This is similar to chess game where some sections could be rated, and others that are not. This might be a natural way to separate out those who play seriously from those who just want to play social bridge.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Richard,

Yes, the predictability is a big thing. In commongame database we kept an expected (predicted) score and actual score for each pair every game. These data could be used for study.

Thanks for alerting us about legal issues with player data. For rating purpose we only kept player ID and partnership ID. The commongame handles all other data. I believe it generate its own player ID for players that don't have ACBL player numbers. In principle the rating system don't even need to know player's name, just their ID (ACBL number or Commongame player ID). In practice it kept a map between the ID and player name so it could generate report with player names. No other player information is saved.
Jan. 1, 2018
Ping Hu edited this comment Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
That would be a topic that needs separate study. There might be a few different ways to do it. I have some studies in my original document two years ago. Now we are getting more data and I'm sure we'll do some more studies this year.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Jim,

There is not public access to the database directly. If you want to get information, send me private message.

However the information about the handicap could be found at
http://thecommongame.com/PingHu/PingHuRating.html
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
The results were kept in the commongame database. It has rating data for over 66K players and 220K different partnerships.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Marty,

Most of the questions you raised about Elo rating has been solved. Look at my post below.
https://bridgewinners.com/article/view/an-elo-rating-system-for-bridge/
This system has been running for the Commongame for over two years. If you do to thecommongame.com, you could see a handicapped score based on this system and related information.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Jeff, I agree with you.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I think the original OP probably should ask in this way: if it is the same 30 tables of players, play a team game, a pair game or an individual game, should the masterpoint awarded to be the same?

If it were asked in this way, my answer would be YES. The first place award should be same, the total masterpoint awarded should be about the same.

There are a number of ways you could achieve this. The 2nd and lower place MP award is calculated as a fraction of 1st place MP on a sliding scale. All it needs is to have a different sliding scale for team, pair and individual game so you could get the total MP awarded to be same. Currently ACBL uses the same sliding scale for team, pair and individual game. Then it tried to twist other parameter to make them compatible but it did not work.
Jan. 1, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
The bracket size is usually determined my tournament organizer and directors. For bracketed Swiss, it is usually targeted for 8 per bracket so it is easy to pair teams head to head.

However some tournaments have chosen 9 teams per bracket. The reason is masterpoint award. For 8 teams per bracket, 3 of them get overall awards. With 9 teams 4 would get overall MP. So for the same 72 teams, 8 team per bracket produces 27 overall MP winners, 9 team per bracket would have 32.

As you found out 5-6 per bracket only awards top 2.
Dec. 31, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Kevin,

You are correct that it does get a little complicated in term of tournament organization. However we could try to model it to reach some rational decisions.

Let say I reorganize the event in this way: all teams entering CKO will pay for 2 session fees. If they are out of 1st session, they automatically enters 2nd session for single session Swiss. This CKO + one session Swiss is equivalent to two session bracketed Swiss. Furthermore let's assume no additional teams enter the one session Swiss. I would say the MP awards total (all places) for two session Swiss should be about the same as those from CKO + one session Swiss.

In practice you may want to have the one session Swiss to pay a little more to attract more teams.

There are a number of ways you could do to achieve this. CKO overall award is limited to top 25%. You could pay two session Swiss up to 30 or 35% so that they could be compatible to CKO + one session Swiss.
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Moved
Dec. 30, 2017
Ping Hu edited this comment Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I abstained for this question because I think it is not well formed. Are we trying to compare a 60 entry pair game against a 30 entry team game? If so, there are actually two questions? First, should pair game pay the same MP (first place and aggregate) as team game? Second, how should field size be factored in MP awards.

For the first question, philosophically I think they should be same, both for first place and aggregate. For the second question, it certainly will affect on the depth of overall awards. However I would favor total table count as secondary factor in MP awards, for example it could change the MP determined by strength by +-20%. See my previous post from link below.
https://bridgewinners.com/article/view/fixing-acbl-mp-system/
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Kevin,

This should not be a problem. If we just normalize the first place MP to be the same and the award depth the same (25%), the MP awards will be compatible.

For teams out of KO, in principle they could enter another event (so called loser's Swiss). This would not be much different from two session Swiss where at lease half of the teams are usually out of competition for overall after one session.
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Max,

I don't think we are in much disagreement. The OP simply asks a philosophical question: should Swiss and KO pay same amount of MP award for same teams.

To put them in correct context we have to create the similar condition of contest. Assuming we have a good strength based rating system that we could rate each team correctly, for 40 teams we could either divide them into 3 brackets or 3 strata. More teams should be an issue in the future. We should be able to divide them into as many strata as the brackets we do today. With the same “condition of contest” there is not a reason to favor one over the other.

The pair/team disparity is slightly different. Pair game is scored by matchpoint, team game is usually IMP. So the first question is if MP and IMP should be awarded differently? Philosophically I don't think they should. Next is a more practical issue. Players tends to like to play in pairs games over team game in tournament, because most of them don't have a fixed teammates. So I think there is a reason that team game should play a slightly higher MPs (in aggregate term) than pair game in order to draw these players into team game.
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
The problem with reward different masterpoint for different event format is it will drive players to one format over the other. Overtime the less favorable format will have less and less players and eventually it will die. The 2015 change to reward pair event higher MP over team has shown this. Players don't want to play in Swiss team or KO when there is a pair event available in a tournament.
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Multi weak two in majors as I said in my earlier response.
Dec. 29, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
It is absolutely true what best use of 2 depends on the entire system. My design of limited bids in a strong club is like the following:
1) 1/: 5+ card;
2) 2: original Precision
for the rest of hands,
3) 1: unbalanced;
4) 1NT: balanced or semi-balanced;

The 1 seems ambiguous. However if you think it in the context of you open any 5+4M with 2 and semi-balanced with 1NT, you'll find it must either has 4+ or 0-1 . The later case is just Precision 2 with 4-4 in majors. This make 1 open much better than “nebulous” 1. It promises something!

What could gain with extra 3 bids (2//NT)? I could use them to open 5-5 hands without full opening value.

I don't view weak two as “weak” hand. It has a lot of offensive values. Consider hands with 6 card suit and 10 HCP vs a balanced 11, which one could make more tricks? In term of loser count, a typical weak two is 7-8 losers. A hand with 5-5 two suits and similar HCP strength is 6-7 losers. It would be a big advantage to be able to open these 5-5 hands. I consider these bids all constructive, not “weak”.
Dec. 29, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
In my system, this would be easily handled. Partner open 1 to show unbalanced hand. You respond with 1. Partner rebid 2NT to show 4 with s/v in and minimum. Now you could judge to pass or play 3.
Dec. 28, 2017
Ping Hu edited this comment Dec. 28, 2017
1 2 3 4 ... 7 8 9 10 11 12 13 14 15 ... 25 26 27 28
.

Bottom Home Top