Sigi_BC84

This is how it basically works on both declaring and defense. AFAIK MG has improved on this idea to make decisions during later stages of play when GIB has collected enough information -- the computer then tries to select the best line of play rather than just solving the problem locally. The paper can be found here: http://www.jair.org/abstracts/ginsberg01a.html --Sigi

Very nice idea, but incompatible with "real" bridge. Ideally the computer accepts explanations as they are given in f2f bridge across the table (or typed into an alert window on BBO). To correctly interpret such information, some natural language processing would be needed as well as an expert system that would, for example, know what "capp" means. To implement your "disclosure by example" scheme, the pair opposing the computer would have to carry around a sufficiently large example database (or have enough records in BridgeBrowser or BBO or whatever you'd be using). This is often not the case. Furthermore you'll be ending up without enough information too often. Take denial cuebids after a relay sequence as an example. The denial cues are very easily formalized in the context of the preceding relays. However, finding a sample for a given denial cue sequence will be hard (since it is rather unlikely that a sufficient number of identical sequences has occured so far). --Sigi

:-) Having a strong-three suiter in the Multi is a very common, if not standard, treatment (at least in my club, if not in most of Germany, if not in most of Europe). --Sigi

Matt's basic approach is to let GIB "solve" bridge by search and simulation, without domain-specific knowledge (or at least minimising this). That precludes the use of "learning" techniques. [...] There are other approaches to computer bridge, such as data mining and rule-based expert systems, but Positronic and GIB are probably at polar opposites of the spectrum. I was simply wondering why Matt's paper from 2000 didn't mention this learning approach. He makes it very clear that he thinks brute force and simulation is superior to applying expert knowledge during play. However he also admits that GIB is much weaker during the auction. This still holds true with the current incarnation of the program (the paper is from several years ago). I'm only saying that it looks promising to me at first sight to try to improve GIB's bidding by implementing some kind of learning/mining/expert system. Probably different approaches are needed concerning bidding and play. I also see major problems in communicating the meaning of opposing bids to the computer. To play "real" bridge, the computer has to be able to understand simple explanations, as there is no time to formulate complicated expressions describing a call during play. --Sigi

Hey, it's getting interesting. Blasphemy. What do you use (please don't say GNU style)? Yes. --Sigi

I would call 2♥, especially since this shows ♠ and ♣ specifically in our system. Partner is passed and I have only one useful defensive card and with luck might get one ♦ ruff, but not more. So why not put my offensive weight into the auction now? Next round might be too late (I do not see how to show my spades on a reasonable level after a Bergen raise or J2NT or whatever they might do). --Sigi

Wow, this is cool. Combine that with GIB's playing abilities and you'll get a nice computer contestant. Unfortunately probably not many people will go so far as to practice 1000 hands with the computer (of which the first few hundred will probably be quite frustrating). Maybe one could efficiently seed the training database with information about the system and then improve on that. I'm wondering why Matt Ginsberg doesn't mention this approach in his paper about GIB (maybe I've missed it, however). --Sigi

Why not include the strong 3-suiters into the multi as well? Did I miss something here? --Sigi

that would be agood task for the bridgebrowser wouldn`t it? i wont mind playing against other scores than the original stuff if they are prooven by a sufficient number of deals. as far as i know the original list is based on the data of 2.500.000 boards. After calculating the pars for 2.5 million boards you are probably really close to the perfect average for most HCP counts. Maybe I'll repeat the calculations myself one day using GIB to see how quickly the figures converge. Bridge Browser has 100 Mio hands but my prediction is that you wouldn't get different numbers when analyzing all these hands and calculating the means (I'd certainly be very interested in actual results, but I don't own the hand records). Zar Petkov has analyzed 10 Mio hands to validate his point count and is offering the records on his web site. DD analysis has already been done on those, so you'd only have to average the scores and compare to Russian Scoring. --Sigi

DF (or GIBs DD analyzer) doesn't help you at all here. You need perfect knowledge of all four hands in order to to double dummy analysis. Now for computer assistance: Matt Ginsberg (the author of GIB) has claimed several years ago already that his program plays on "expert level" (he also conducted experiments and pitched the computer against experts, Rosenberg among others, to prove that). Since then he has improved the program considerably and there is a lot more computing power available now (steadily increasing). Considering this, I think it is quite feasible to obtain very valuable automatic or semi-automatic computer assistance by running GIB (or Jack) client side to help with the card play. I would probably bid myself (GIB sucks at bidding) and then let the computer play the hand for me (maybe after giving it hints on the meaning of the bids and so on, which is easily possible with GIB). You can check GIB's strength yourself by pre-dealing hands, setting the final contract and playing against it. --Sigi

The Russian Scoring table is "valid" in the way that it represents the expected score for a given combined holding of HCPs. This has been determined by double dummy analysis and averaging of the par scores. So when using this table to adjust the scores, you are playing against the average. The averaging has been done for both vulnerable and nonvulnerable scores, so basically there are two separate tables. This still allows for lucky swings of course, since you can have a 20 HCP game or a 24 HCP slam or whatever. Therefore it doesn't eliminate luck but diminishes the effect of having "good cards" since better (higher) cards mean you need to score higher. The Russian Scoring table was made for f2f games, in an online world you could even calculate the par score for every hand and compare against the perfect score. This is as fair as you can get on a single table I would say. The question is if people want that (I understand that part of the thrill when playing rubber comes from the possibility of getting good hands). --Sigi

There are coded decks for minibridge (24 lesson deals each) which get you around this problem quite nicely (they include a booklet with all the deals and a few notes on each deal). The main problem with the system is that it's very different from SAYC and eventually you want to teach your students SAYC and then you'll have to tell them, "well, forget the complicated system you've learned, here's the 'real' thing". I don't want to do that. --Sigi

Do you drink it with coke, banana juice? To those of you who might wonder: It is very common in Germany to mix all kinds of beer with coke. Wheat beer actually is mixed with banana juice even. Very tasty. --Sigi

Indeed. (It also made me pee more often than "regular" beer.) --Sigi

LOL. Also be prepared at all times for GIB bypassing 3NT for no reason (happened to me in an offline game, uncontested auction!). I think it's playing the cards really well, but bidding is no fun. --Sigi

I think the default value (setting #3) is the best option for most people -- most will probably feel no need to change it. It should have been the default behaviour of FD to begin with, so basically introducing it can be seen as a bug fix. If people change that option one should assume that they knew what they were doing (probably they are FD users themselves and are annoyed by bad FD files or whatever). If somebody gets damaged because he or she randomly changed options in the BBO client, well, bad luck for them I'd say. In my eyes the worst case is assumed too often at the moment when FD features are being discussed. Somebody is confused by a certain aspect of FD and downright assumes that this will lead to major cases of misinformation and appeals. Let's be a bit more optimistic and actually assume some basic common sense and intelligence with BBO players. The "possible outcomes" part of the explanation box should get a label, however. You can deduce what it must mean, but here one should in fact assume that many players will be confused -- especially since it's completely unusual to point out the potential strains at the current point of an auction. People probably don't expect this kind of information so it will be hard for them to understand that it's being presented to them at all. An explanatory label will make it clearer. --Sigi

I'd rather have George Clooney or Brad Pitt take up Bridge than the German chancellor (who might be called a little old lady :-). --Sigi

Pearson point count is a guideline for opening in 4th seat. In short, when deciding whether to pass out or open, open if(HCP + # of ♠ >= 15) Atul Oh, I know that rule, only not by the name of Pearson Count. I didn't mention it because I thought it's obvious anyway, LOL :-). --Sigi

Unfortunately he also displayed kind of an arrogant attitude and left with $500. Not the best ambassador of Bridge... He should have won the Million, with some more opportunities to promote the game. Oh well, at least now we've had Bridge mentioned on German prime-time TV once. --Sigi

It would already help a lot if only one or two class A celebrities would play, and promote, Bridge. Omar Sharif promoted Bridge and he hasn't been an A-list star for a long time now. Actually I knew long before I learned Bridge that it's a game that Sharif played, while I had not even seen a single movie starring Sharif at that time! That's how well celebrity "endorsement" works. A Chinese player from my club has told me that bridge has become popular in China while their leader was playing and promoting it. Now you might say, this must have been propaganda, or that people in China blindly replicate their leaders to please the ones in charge or whatever, but the point is that it really helps if the game gets into the public eye in such a way. The main problem of the game is the image problem; there might be other reasons (learning curve and all), but the main issue is the picture of the typical bridge player (little old lady) that people are carrying around in their head. This stopped me from considering Bridge for a long time, and I was convinced in about a second that it's the perfect game for me when finally I took the opportunity to take lessons. If George Clooney started to play bridge and talk about it in the countless interviews he gives, believe me that image problem would be gone at that moment (only slightly exaggerating). --Sigi

What I find really compelling is an example Charles Stross gives in one of his books: Assuming that AI (artificial intelligence) is possible at all (this being an open question), you will have a god-like AI in no time, for the AI could use the exponential growth in computing power to improve itself, which will in turn add to the potential growth (the "base" of the exponential function). Actually in his stories, such AIs already exist, providing some really hilarious settings :-). I love the AI gedanken experiment because it makes it so clear what exponential technological growth really means. --Sigi

Charles Stross is writing novels that deal with the technology singularity phenomenon. Very high-paced, hard science-fiction, some really mind-boggling ideas. In the foreword of one of his books he says that it has already become close to impossible for a science fiction author to make any credible projections that stretch into the future for more than a few years for the reason that the singularity might be quite close and even with todays rate of technological improvement most sci-fi predictions are rendered invalid within a really short time. Wikipedia entry for Charles Stross --Sigi

I don't know the Pearson count, but you have 25 Zar Points plus spade suit, so according to Zar this is a proper opener... (At the table I would pass.) --Sigi

You should not be ashamed for liking those two movies. Neither has been made to inspire the intellects of the audience, and that's fine, as there are plenty of other valid reasons to make a movie. I've watched Titatic three times when it was running in the cinemas. I think it is flawless craftsmanship and some of the acting was very, very good (Kate Winslet should have received that Academy Award for her role, unfortunately there was a strong contender that year [Helen Hunt in As Good as it Gets...). The first two times of watching I did not realize at all how long the movie actually is (my backside hurt after the 3.5 hours, but during the film I completely forgot time). It's a big accomplishment to begin with if a movie manages to make you forget that you are sitting in a theatre, let alone for over 180 minutes. American Pie is a very funny movie with likable actors; well, the raunchy subject matter might put off (or even offend) people, but then some people are even offended by Forrest Gump, so what the heck. Four thumbs up for these two films. --Sigi

AHHHHHH. I thought I was reading worst movies when I saw this :) Fargo is brilliant. What did you dislike that much about it (I'm seriously interested)? --Sigi