[Gerben42]

The defensive methods from the Magic Diamond team can be found by downloading Magic Defence.

Anti-nonsens, the defense against the garbage opening in a forcing pass system, can be found in paragraph 4.14.

What I am currently using as defence to DrTodd's Dejeuner can be found here:
Dejeuner defence

[DrTodd13]

Dejeuner has several bids that are un-anchored. Gergen's defense to
dejeuner has even more bids that are un-anchored. This level of
uncertainty is fine with me but it certainly does add an element of
randomness to the game. If every initial bid and initial overcall could
show multiple hands types, I wonder if even the most liberal among us
would enjoy such a scenario. In principle, I wonder what the optimum
amount of randomness is for sifting out the best players and for the
enjoyment of the players themselves.

[Gerben42]

Quote

Gerben's defense to
dejeuner has even more bids that are un-anchored.

But only in situations where we are weak and the opponents are strong: Either they have shown 14+ and we are too weak for a constructive 1-level bid, or they have started the GF relay.

In all other situations there is always a known suit: shape first and try to get to known ground. The only thing I'm worried about is the defence to the garbage 1♥ but this is the toughest thing to do.

[Winstonm]

Jlall, on Sep 2 2005, 10:33 AM, said:

The less randomness the better in terms of enjoyment (for me).

For me as well. Also, seems that randomness would not ferret out the better players as much as level the field.

Though no expert on the subject, my understanding of chess is that a lesser player simply cannot win against a better player - that chess is the ultimate game of skill and randomness is not a factor at all.

However, if you've ever played Texas Hold 'Em, you would realize that randomness only succeeds in diminishing the width of the gap between good players and weaker players when a J9 offsuit takes down all the cash against AA.



Winston

[EricK]

Winstonm, on Sep 3 2005, 07:42 AM, said:

Jlall, on Sep 2 2005, 10:33 AM, said:

The less randomness the better in terms of enjoyment (for me).

For me as well. Also, seems that randomness would not ferret out the better players as much as level the field.

Though no expert on the subject, my understanding of chess is that a lesser player simply cannot win against a better player - that chess is the ultimate game of skill and randomness is not a factor at all.

However, if you've ever played Texas Hold 'Em, you would realize that randomness only succeeds in diminishing the width of the gap between good players and weaker players when a J9 offsuit takes down all the cash against AA.



Winston

Suppose you are up against a better player at chess. What can you do?

If your opponent is a better overall player but handles certain sorts of positions worse than you, you can try to steer the postion into one of those.

But if he is better than you in all aspects then what? Your best chance is to complicate the position. You will still probably lose, but there is a chance that he might go wrong in the complications.

A similar thing applies to bridge. If you are up against better players you will probably lose (although you might win luckily if eg the opponents bid to a lot of good games which happen to go down on bad distribution). So what can you do?

If they are better than you at all aspects of the game you have to try to complicate as much as possible. Pre-empting more randomly, obstructing their auctions and using complicated methods are good ways of doing this.

It is no surprise then that good players want there to be as little randomness as possible. The ironic thing is that the people in charge of bridge are generally good players, and they claim that they outlaw certain methods to help the run-of-the-mill player - but the people who really gain from everybody playing less complicated methods are the best players, not the worst!

Eric

[Free]

I prefer to play strong pass against a strong pass opening Especially auctions starting with pass!-pass!-? are pure fun imo.

[Winstonm]

Suppose you are up against a better player at chess. What can you do?

Improve or lose?

[EricK]

Winstonm, on Sep 3 2005, 04:04 PM, said:

Suppose you are up against a better player at chess. What can you do?

Improve or lose?

Improving is a good long term strategy, but not necessarily practical when in the middle of a game!

I was once a reasonable chess player but I did manage to beat a few better players (and draw with a much better player) by sufficiently complicating the game. I have also lost to worse players when the situation got too complicated for me.

As I say, it is the better players who have more to fear from complicatedness - both in chess and bridge.

Eric

[DrTodd13]

If you as a player feel that you are better able to cope with complicated auctions than your opponents then I don't see why you wouldn't want to try to create complicated auctions. It is not necessarily a sign of weakness in some other area that you try to do this. It may only be that you're using everything that you can to your advantage. I think the situation is exaggerated a bit too for you have to retain some constructive to your agreements or you'll never reach the right contract if the opps don't fall all over themselves getting in your way. So, hand type will be resolved somehow and this will provide a mechanism for competition.

[the hog]

The Swedes use AntiNonsens - some here use what is called the Antifert:

Over the fert:
X = bal 12+, hoping to catch them. Herbert responses
Bids = normal openings
1NT = any 0-10!!

Over transfer openings:
X = opening hand with the suit opened
1suit shown = takeout of suit shown

[Impact]

As the developer of the anti-fert 20+ years ago, I consider that you have given it poor press by stipulating that it must be played with your accompanying material (and we tend to play it slightly weaker - 0-9 with certain handtypes suitable for other actions removed)....it does risk poker-like confrontations but that too is part of the charm of the game.

There are a host of continuations against the varying types of "sub-minimum" openings. I have typically split methods between countering one level openings which routinely are based on 7/8 HCP and those which approximate as their general base light standard eg 10+HCP (without commenting on the accuracy of HCP evaluation but as a known base for comment).

When you open lower than the suit shown you gain in many instances in auctions without competition (both by virtue of the levels gained and often by the trasfer effect particularly in the case of interrogatory type auctions initiated by relayer as opposed to dialogue).

However like almost all matters there is a loss as well: it allows additional ease of competition by opponents (as the level of pre-emption is less) - and of course the lighter you open the more you disclose to opponents should you "lose" the auction and they become declarer.

Notwithstanding a plethora of material on maximising the usage of different opening bids for "efficiency", which is appropriate for unwinding a relay system, there is little consideration of the reverse view which pertains to competition. I think the codification of the relevant parameters (and disagreements likely to pertain) has militated against such analysis.

IMPACT operates as a forcing pass (but potentially 0-4HCP) 5-9 FERT or singlesuiter 10-15 6+C and 2C opening as frequently 0-4 any when n/v; but strong club vul.

It is fairly easy to demonstrate that different systems are more likely to be effective at different vulnerabilities (to the extent that at unfavourable vul there is a fair argument for a ROmex or Roman or Power NT while at favourable a mini-NT of 9/10-12 is superior for dealing with pre-emption/pre-emption). We also used different ferts at nil vul (1H) and favourable (1S) as the one step is often crucial in disrupting relay auctions of opponents who might seek to relay over the fert at this level.

Most good partners have enough trouble absorbing the intricacies of 1 system let alone many (even as the designer in the heat of battle the forces of alternative logic structures may be too strong).

Further, the administrators in Oz, in their wisdom, have determined that "Two systems" is the maximum that may be played - putting paid to true optimisation of methods based on vulnerability - albeit there appears a general acceptance that methods which do not include forcing pass entail sensible variations for 4th hand opener (eg if multi twos it would be obvious foolishness in 4th seat to open to show a truly weak hand of say 5-8 HCP without a Major etc).

Game theory has suggested for a long time that optimal theoretical solutions to multi openings of all sorts is multi defences; unfortunately the practical ramifications of sorting such out have proved harder than you might have thought and the effort has often been wasted by partner's non-computer like brain.

Try to remember when designing systems that the ideal paper solution in each conceivable auction ends up imposing an all but intolerable memory load as the ramifications of application of each layer of logic to each bid is immense (and time-consuming). Hence the operation of another of S J Simon's rules about the the best possible result as opposed to best result possible...

Accordingly, both for simplification and speed there is a tendency to create "rules" with general application, and only specifying the most crucial exceptions (of course the devil is in the details in determining where the dividing lines fall). To give just one example of my current natural responses over intereference to limited opening bids :-

1S (3D) ?
3H is initially a DAB based on USP - but need not be if one was content to reverse its meaning with the obvious sounding 3S raise.

You may be familiar with this type of paradox in the more familiar position of responding to 1NT/2NT opening where Fibonacci analysis specifies the number of bids available for relays below 3NT and accordingly allocation of various bids including pivots. If not used for natural/semi-natural purposes (eg S or S shortage or minors), the 3S bid could easily be used asa pivot to 3NT allowing a range of clarifications by responder. If that is done, the efficient corollary is the absence of a need for a direct 3NT as natural - so you may use it as a pivot to 4C with all sorts of clarifications available.

Unfortunately in an all to human world the latter such bids (3NT in all equivalent auctions as a puppet/pivot to 4C) which I favoured for system optimisation 25-30 years ago are just too prone to disaster in the real world - the theoretical merits being defeated by the human operators.

Accordingly many of the "finer" system ideas (including a complete encrypted bidding system) have been relegated to the basement for occasional intellectual tinkering but no practical purpose.

This ramble has already gone too long and I should post under a separate topic...

[the hog]

This has to be Bob Sebesfi unless I am sadly mistaken.

[akhare]

The_Hog, on Sep 3 2005, 04:56 PM, said:

The Swedes use AntiNonsens - some here use what is called the Antifert:

Over the fert:
X = bal 12+, hoping to catch them. Herbert responses
Bids = normal openings
1NT = any 0-10!!

Over transfer openings:
X = opening hand with the suit opened
1suit shown = takeout of suit shown

Hi Ron,

Why employ 1N as any 0-10 over the fert?

Atul

[the hog]

Bob, who I suspect strongly is "Impact", as that is the system Sebesfi designed, liked to create action. You have to admit this is a bit chaotic : 0-8, 0-10, your bid at the 2 level?

[Impact]

Not Bob - rather his partner in crime and system designer.

[the hog]

Ah, now I have it. Fred Curtis.

[hrothgar]

Impact, on Sep 4 2005, 09:17 AM, said:

Game theory has suggested for a long time that optimal theoretical solutions to multi openings of all sorts is multi defences; unfortunately the practical ramifications of sorting such out have proved harder than you might have thought and the effort has often been wasted by partner's non-computer like brain.

Hi there

Thanks for an interesting post. I was curious whether you could provide a pointer to any results supporting the supplied quote. I've done a bit of work trying to apply formal game theory type analysis to bridge. For example, I've argued for a long time that the traditional notion of "psyches" do not exist. Rather, the disclosure rules force players to use an imperfect establish vocabulary to describe complex mixed strategies.

I very much agree that most sub-game perfect equilibria involving a mixed strategy require both pairs to randomize according to some optimal probability density function. However, its not immediately intuitive that a "pure" multi-bid necessarily requires a mixed response.

I was curious where the result came from...

[Echognome]

hrothgar, on Sep 5 2005, 12:17 PM, said:

Thanks for an interesting post.  I was curious whether you could provide a pointer to any results supporting the supplied quote.  I've done a bit of work trying to apply formal game theory type analysis to bridge.  For example, I've argued for a long time that the traditional notion of "psyches" do not exist.  Rather, the disclosure rules force players to use an imperfect establish vocabulary to describe complex mixed strategies.

I very much agree that most sub-game perfect equilibria involving a mixed strategy require both pairs to randomize according to some optimal probability density function.  However, its not immediately intuitive that a "pure" multi-bid necessarily requires a mixed response.

An interesting post and also idea. I certainly like the idea of using some formal game theory for analysis. Seeing this is part of my area of research I feel qualified enough to post on it.

As far 'psyches' being part of a mixed strategy, I don't feel that definition is quite right. I would agree with that definition if the rules allowed partner to account for me having psyched. As it is now, partner is restricted not to field the psyche, so I wouldn't call it a mixed strategy. However, I would also accept an expanded definition of a mixed strategy where your choices of actions led to some more limited paths and others to less limited paths.

In response to the discussion on SPNE's, I'm not sure what you are trying to say. It seems you are mixing the concepts of best responses and subgame perfect equilibria. Let's assume you are conditioning on hand types as your 'type of player'. Then for a given hand type, the best response to a set of calls may be a pure strategy (always double) or a mixed strategy (sometimes bid 3S, sometimes double, sometimes bid 3NT). I'd be interested to see how you formulate an equilibrium as the utility functions seem hard to quantify. However, I'm assuming the best way around this is simulation.

If you have any work that you'd like to share, I'd be interested in seeing it. At least to me, it sounds like some interesting work.

Matt