Join Bridge Winners
All comments by Nicolas Hammond
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.
ACBL Bulletin. April 2019. Page 40. Matt Smith answers your question.

You may not ask your opponents to stop alerting in ACBL events.
Aug. 1
You are ignoring the author of this comment. Click to temporarily show the comment.
Represented your country is a little tough as there are countries with very few members. I would suggest reached the KO stage of Bermuda/Venice etc. not qualified for BB, Venice etc.
July 31
You are ignoring the author of this comment. Click to temporarily show the comment.
No idea; I saw the date in an eBay listing. But then again I recently saw eBay listing for the 1968 Laws. They meant 1963.
July 30
You are ignoring the author of this comment. Click to temporarily show the comment.
0.5% of ACBL members have > 10K points. See https://web3.acbl.org/mpholdings

For ACBL resident members, who rarely play outside ACBL land, I have no problem with a limit of 10K points. I would not care if you won NABCs or World championships. You have < 10K points.

For non-ACBL resident members, who regularly play in other jurisdictions, it is hard to convert your skill level/foreign “masterpoints” to ACBL points. Typically a foreign “pro” will be assigned 5K when they first come to play in the USA.

I would suggest two MP ratings.

One for qualification to limited events. Keep the initial 5K assignment for foreign players. But have a committee/person have the ability to assign more. If you win a World Championship, you are assigned 10,001 MPs. If you win an NABC and are a foreign player who regularly play outside ACBL, you are assigned 10,001 qualification points. You can no longer play in ACBL limited events.

Your other MP value would be the assigned value you report for KO events. It does not need to be the same as your qualification MPs.

If you have made the finals of four World Championships, you should not be allowed to play in a 0-10K ACBL National event. The limited events are not designed for those that have won a World event. Just my opinion.

There are multiple players that should not have been allowed to play in the 0-10K mixed. Sorry that does not work for the professional players.
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
European players are known to be better in certain areas of the game than ACBL players, for example, opening leads. Over a long match this will have an impact.
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
Agreed. I did see something on eBay which mentioned 1971, but there are many errors on eBay.
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
How is a player who is twice World Champion, twice runner-up World Champion, six time winner of NABC events allowed to play in 0-10K events?

Foreign top players are supposed to inform the TD before the event starts.

The winners of 0-10K Mixed Swiss are Nikolaos Delimpaltadakis - Ioannis Papakyriakopoul - Vasileios Vroustis, Athens Greece; Benedicte Cronier, Paris France; May Sakr, Ardmore PA
July 29
You are ignoring the author of this comment. Click to temporarily show the comment.
I've published some details in my book, see http://www.detectingcheatinginbridge.com. I've got three chapters on opening leads.

There is a snippet on opening leads at http://www.detectingcheatinginbridge.com/selections.html but this shows how good the top players are on opening leads; not the effectiveness of trump leads. I do have the data, it's just not in the book. The book shows the different styles of opening leads of the top pairs.

The pair that leads trumps the most (Sementa/Duboin) leads them over 11% of the time; the pair that leads them the least is Hamman/Wolff at under 4%. I was lucky enough to chat with Bobby yesterday and asked him why he doesn't like trump leads. He gave a long explanation with examples.

The top pairs get opening leads wrong about 1 in 5 hands.

As others point out; when you first learn you are given “rules” to follow. As you play more, these “rules” should become “guidelines”.

I have over 200K records, all from top tournaments. However, trying to dissect the data to find the right circumstances is something I haven't done.

One general rule is to always lead a trump against a low level doubled contract to cut down on ruffs.
July 27
You are ignoring the author of this comment. Click to temporarily show the comment.
If only we knew someone with a lot of time on their hands, say, someone who has just retired, and someone who knew the BBO format relatively well and was an excellent Bridge player who knew what stats to create…

Oh well, wishful thinking.

On a serious note… my book includes data on the elite players by name - Meckwell, HHs, Lauria/Versace. According to the lawyer these are “public figures” by the fact that they have established themselves at the top level. As they are “public figures”, I did not ask, nor need, their permission. Frankly, a lot of the graphs also show the number of boards played so it is trivial to pick these three pairs (+ Fantoni/Nunes). Strictly I don't need anyone's permission to list their names - this is statistical data therefore publishable. However, the conclusions are obvious for some data so the list of pairs in some tables has some “names withheld” because the result is too obvious.

I did ask some other top pairs below the top three if I could include their data. If either player said no I did not include them in most statistics.

I specifically asked Kit (+ Fred) if I could include them. They are an excellent reference pair.

I also include Boye with his two main partners (with permission).

There were some other top pairs that gave permission, but I cut the data for space reasons.

Without permission (but perfectly legal), I included data on same well known pairs in some charts. This is so you can see how some pairs rank compared to other pairs, including the known cheating pairs. For example, Sabine/Roy; top women players, top Seniors, top USA, top European. These are all on lists that cannot be used for cheating detection.

Back to Fred's suggestion.

There is the law of unintended consequences.

The goal of Bridge is to win.

If there are published statistics of how often you do x, y or z, then you may stop playing the game to win and work on how to improve your statistics. Think of the Lehman rating on OKB.

Lists are both good and bad.
July 25
You are ignoring the author of this comment. Click to temporarily show the comment.
bridgescoreplus.com downloads all the data.

There is an option to output handrecords in various formats, including BW.

Same for EBL/WBF/BBO events.

You have my email and cell #, so email, or text. For others, PM me.
July 25
You are ignoring the author of this comment. Click to temporarily show the comment.
@Barry: I'll turn the question around. If a Bridge Organization had a list of suspected cheaters, what would you want them to do with it?

The answer is probably legal based on the organization. Almost certainly there are political issues, but that is outside my domain.

I have provided the information in your last sentence in a proposal given about a year ago. I provided recommended guidelines on usage.

I would expect the likes of WBF to disinvite certain pairs.

For ACBL, there are different legal issues. IANAL.

I don't know if EBL has the ability to disinvite.

Almost certainly you would stuck cameras on highly suspected pairs and also be more active in assigning limited resources to those suspected. This is probably the biggest benefit.
July 24
You are ignoring the author of this comment. Click to temporarily show the comment.
@Stefan: I have multiple algorithms.
ACBL only records the table result, so there is less data to go on; which means I need more data from the results to be statistically valid.

WBF/EBL record the opening lead. But the data can't be trusted - see the book for details.

Vugraph record bidding and play. More information can be processed. Less data is needed for the results to become statistically valid.

@all: The software detects humans cheating in Bridge. If you use Bots, you will have to train Bots to cheat in Bridge. Now you are into an area beyond most people's ability and knowledge base.

@Michal: I am happy for someone to create an independent test. I've written what I think is a fair way of doing things. I've made some predictions in the book, when (not if!) they are proved to be true, I can reveal the names.
July 24
You are ignoring the author of this comment. Click to temporarily show the comment.
Here's the test:

Take all data that I'm using - raw data from ACBL, EBL, WBF web sites.

Take this same data, but change all the names. So I can't have code to check on names.

Give me this new data which has random names for the players.

Put me in a monitored setting/environment. I start with a clean database. I run the software against this data set. I show you the list of cheating players.

==

Here's another test:

Give me a database from a country where I do not know the players.

I will ask the software to detect who might be cheating.

See if it matches any convicted pairs.

==

Here's another:

I provide a set of MD5 hashes on players that are cheating. I put it in a book. See how accurate the predictions are. Wait… I already did that.

==

The software is written to detect humans cheating. There are multiple algorithms, not one.

Writing an algorithm to detect bot cheating would be different. I think.

Remember… a cheating pair has unauthorized information. A smart cheating pair knows when to use that unauthorized information or not. Cheating pairs make mistakes. Just fewer mistakes in certain parts of the game in certain circumstances. Knowing what they are, and how to measure them, is what is new.
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
Yes/no. (If I understand your question).

What is very interesting is that I can tell when a pair suddenly “improved” or got worse.

For example, I believe I know when F/N started to cheat; the same with the Doktors being monitored.

What is very interesting is the number of top pairs that suddenly got a lot worse after the summer of 2015. Many pairs have a dip in their performance starting in Chennai.

The book shows the big change in defense ability in European and World events post 2015.

Before the non-Americans get upset… there is little data kept in American events to run data on.
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
I have a lot of stuff. Data on all top players. Data on all ACBL, EBL, WBF players.

Data is statistically significant given a large number of boards.

Using this I can work out weaknesses in someone's game.

Annoyingly I played against someone (pro pair) this tournament that I had given advice to (their opening leads were poor). Against us they promptly found all the right opening leads. They had actively worked on improving their opening leads for the last 2-3 months.

Obviously a couple of boards is meaningless, but I'll be curiously to check this pair in about a year when I have more data on them to see if their opening leads have improved.
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
@Stefan: “What happened in 1983/84?”

There is suddenly less public data. If someone wants to transcribe from World Championship books, that would be great.

Would love to get the data so that the software can process it.

See https://www.sarantakos.com/bridge/vugraph.html

If not for Nikos Sarantakos there would be no data before about 2003.

Edit: clarify statement in third paragraph.
July 23
Nicolas Hammond edited this comment July 25
You are ignoring the author of this comment. Click to temporarily show the comment.
I do know how often world class players get this, or similar situations, right. I also know the same but for cheating pairs. There is a difference. This is how probable cheating pairs are identified. Details in book….
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
@Barry: “I'd love to see the results from a large data set of club games. I need a good laugh now and then.”

I don't do (much) club data. There are various reasons. It is possibly to accurately process some club data, but I'm not going to give the details.

The problem with club data is the wide variance of skill level. The same problem exists at tournaments, but to a lesser extent.

In the book, I show the difference between top players, then top players when they only play against top players. It is the latter where it is easier to detect cheating.
July 23
You are ignoring the author of this comment. Click to temporarily show the comment.
There is much more than one algorithm. There are different algorithms for different parts of the game.

If the algorithms are fully described, it becomes easy for a player to know how to manipulate some plays to avoid being detected.

It then becomes a cat and mouse game of trying to stay ahead of the cheating pairs with each successive generation of cheating detection tools requiring more sophistication.

Also… being blunt, the software is a commercial product. There is no benefit to releasing the source code or algorithms to non-licensed customers.
July 23
.

Bottom Home Top