Join Bridge Winners
All comments by Richard Willey
You are ignoring the author of this comment. Click to temporarily show the comment.
> Obviously the health of bridge is a complex multi-faceted
> issue, but If any ONE thing were killing the game, it would
> be this attitude.

Let me play devil's advocate:

During its hey day, bridge was an extremely popular form of mass market entertainment. I don't believe that it can survive as such.

I believe that the key to bridge's survival is to manage its transition to a niche game with a smaller target audience who appreciate its complexity and want a challenge.

I very much worry that if you try to satisfy both audiences we'll lose both (and I think that you're going to lose the social players regardless of what you do)
March 28, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> People leave the game because the game is too
> complicated. They leave because they have to leave
> their comfort zone deal with opponents playing
> unfamiliar systems. They leave because it's too hard
> and they have no chance at winning.

Perhaps these people should not be competing in a game of skill…
March 28, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
this is great! Exactly the type of think that I was hoping to see people discussing…
March 27, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Silly question:

Has anyone compare the outcome from the Fantasy Vanderbilt to the results of the actual event? It would be interesting to see how well a “wisdom of crowds” type ensemble was doing…

If the ensemble is actually doing a good job, perhaps it would make sense to provide a cash prize and use this as the seeding system…
March 27, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Because there is a strong tradition that natural methods should not be restricted and canape bidding is considered natural.

If you do plan to restrict canape openings, I do hope that you will also ban Walsh responses…
March 25, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Few quick comments

1. I think that it is possible to make a strong case for licensing a single system for novice events. I think that once can also make a strong case for opening the floodgates to SAYC, 2/1, Acol, Polish Club, Precision, EHAA, what have you.

I am quite uncomfortable with intermediate solutions. Once the camel has its nose in the tent, I think that you need to let the whole create climb on in.

2. If other convention charts will be based on a principle that everything that is not banned is permitted, I think that you should try to structure this chart in the same manner.
March 25, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> Do you know of any mathematical problems that have been
> solved by majority vote of a committee?

https://en.wikipedia.org/wiki/Indiana_Pi_Bill
March 23, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I admit to having a naive view wrt seeding, however, from my perspective the best seeding system is the one that produces the smallest number of upsets…
March 23, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> I agree there is currently no objective way to measure bridge
> skill. In my opinion, people should stop focusing on a system
> of seeding based on skill but rather a seeding system based
> on accomplishments.

To my knowledge, no one has yet implemented a good objective bridge rating system.
This does not mean that this is technically challenging.

From my own perspective, the three things that make this difficult is

1. Telling 50% of the players that they are below average
2. Convincing 80% of the players that they should trust some black box algorithm
3. Getting the powers that be to collect the data necessary to implement such a system
March 23, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> What is your advice for this hypothetical pair?

My advise for the hypothetical pair is that they need to convince folks that it is bonkers to use a system that restricts itself to considering binary match win / loss records for seeding purposes.
March 23, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Very much agree with the comments that Jurgen is making.

One point that bears repeating:

If you want good ratings, you need good data.
If you want good data, you need to improve the record keeping.
This means recording the score for each an every board during a match, not just the results of the match.

This should be MUCH easier than implementing a comprehensive rating system.

Moreover, this could be done by the USBF before the ACBL comes on board…

Once I see these sorts of changes put in place I'll believe that folks are serious about wanting a solution rather than just venting…
March 23, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
@michael

Let me give you an example of the type of analysis that I would want to conduct in order to answer your question.

Please note, I did this analysis for the Vanderbilt Quarter Final Match between Rosenthal and Schemer. I deliberately picked this match because it was exceptionally close - it was only decided by a single IMP… This match also featured some really large quarter by quarter swings. Lots of IMPs ended up changing hands.

In order to determine my confidence in the outcome of the match, I used a statistical technique called a “bootstrap”. I created 10,000 virtual matches, each of which had the same statistical properties as the original match. I then looked at the mean and standard deviation for the 10,000 virtual matches.

I created each virtual match as follows

Q1 of the real match consisted of 15 boards.

I sampled with replacement 15 times from the set of 15 results. This creates a new dataset with the same moments as the original, but included a slightly different set of results (some board might be sampled two or even three times. Some boards not at all).

I did the same thing for Q2, Q3, and Q4 and then summed across the results to create a virtual round. I repeated this process 10K times and ended up with an array of 10K virtual matches.

On average, the result was almost identical. Rosenthal won by a smidge. However, the standard deviation was enormous (12.5). Based on the variance that we saw in this match, I would expect that Schermer would have won the match 50% of the time and they would have won by 12+ IMPs roughly 16% of of the time….

[Please note: the standard deviation for the all the boards in the match was “only” 6.18. However, I think that the bootstrap methodology which treats the match as the sum of 4 quarters with different pairs facing each other is a more accurate treatment….
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> If the seeding was perfect, how many upsets would you expect
> between the 17-32 and 33-48 teams and how many upsets would
> you expect between the top 1/4 and bottom 1/4 teams?
> Do the KC results as a data point surprise you, or is that
> what you expected?

really difficult to make a meaningful statement about this when all you are giving us is information about the mean and nothing about the variance…
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> This is not about missing data: this is about improving how
> the seeding is done, if possible. Can you add to that?

Almost certainly, but I doubt that it is worth my investing the time/effort to do so…

Conceptually, I believe that it is possible to develop an accurate rating system. However, in order to do so we would require

1. Some kind of non-parametric modeling technique. At the end of the whole process, we'd have some kind of black box model that outputs a list of seeds. However, it might not be possible to explain precisely why the model had decided to rate team X rated higher than team Y. (Please note: this is no worse than the situation that you have today, however, I think that people will insist on higher standards when a computer is involved than they require from the existing system)

2. Much better record keeping. The types of models that I would want to use work best with lots of data. Let's suppose that Nickell beats Lavazza in a 128 board match. It would be MUCH more useful to have the scores from each and every board that was played rather than a single data point showing the final results. Until we make a sustained effort to invest in comprehensive record keeping we don't get to have good seeding methods.

Convince me that people care more about accuracy than simplicity. Then invest in processes to record useful information. At which point I am more than happy to sacrifice a chicken to Baron Samedi, spin up some machine learning algorithms, and address this properly.
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> My view is that the event should be seeded based on past
> performance of the players, not on a subjective guess on
> how the teams should be ordered?

If you already know the answer you're looking for, why did you bother to ask your question?

Or, more specifically:

Great! New players have zero history, there fore they have zero skill, there fore they get no seeding points.

Problem solved…
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Looks like there are two separate discussions going on

1. How should we run seeding (in general)
2. How should the seeding system account for missing data

#1 Would appear to be a special case of “How do we develop an accurate rating system for bridge”…

(Does anyone disagree that if we had an accurate rating system, we could use this for seeding purposes?)
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
From my perspective, it makes sense to treat this as a “missing data” problem.

The field of Statistics has a bunch of different methods to handle situations where specific observations are missing from a data set. “Mean replacement” or “median replacement” are one (well known) technique.

Different techniques get used depending on why the data is “missing”. For example, there is one set of methods that one might employ if the data is “MCAR” (missing completely at random). In this example, my suspicion is that the data is not MCAR. The set of players where we are missing seeding information share certain characteristics (for example, they are new players or they are players who have not yet played in a top level event in a US national)

From my own perspective, it might make sense to use the strength of one's team mates as a predictor of the strength of the new pair. I would expect that a team that is looking to augment will invest more time / effort in investigating the (potential) skill of a new pair than anyone else. As such, there should be a reasonable good match between the average skill level of the existing team members and the folks that are being added.
March 22, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
The purpose of the letter that Jacobs is citing is to warn perspective Uber drivers that their insurance won't necessarily cover them, not to warn the passengers…
March 18, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
> Oh, convention/agreement disruption exists. It is a
> violation of the CoC .. we are required to know and
> be able to disclose our methods.

I don't think that anyone disagrees that players should be able know and disclose their methods.

What distinguishes convention disruption is the notion that some sets of methods should be treated differently than others. If your conventions are on the approved list and you screw up, its “just bridge”. If you make the mistake of playing methods that Bobby Wolff disapproves of, he will try to invent a new form of punishment…
March 18, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Given how secure the ACBL's hand generator was (perhaps is), you should be thankful that you were using shuffle and play…

Speaking of which, does anyone know whether the migration to Dib Deal actually happened?
March 17, 2017
.

Bottom Home Top