MULTI DATA PROJECT

[Free]

Is it? I thought 2♥ gains a bit on 2♦ because opps are under more pressure to act, but it loses slightly because opener can pull 2♥ to 2♠ (2♦-2♥; 2♠) and if responder has a weak hand with long diamonds. If we played a long team match and the dealer had always only hands that are one card away from 4415, would the 2♦ team beat the 2♥ team?

I don't have any numbers to support my statement, but 2♦ sure looks better to me. Responder can pass 2♦ and opener can pull the 2♥ response with a 4-3-1-5 (which is a huge one imo - although not so frequent ofcourse). But another reason you didn't mention is that defenders usually don't know anything about declarer's hand (even if responder is strong).

 

After a 2♥ opening, you may be playing a 3-3 Major fit (since you can't pass with 6♦ and you probably don't want to bid 3♦ either, you'll bid 2♠ on a 3 card suit more often), you may play any ♥ contract and let the known hand play (almost DD defense), and all you get is a little pressure. Imo it's just not worth it.

[Mbodell]

Interesting. When I play this I pass with 4315 when partner responds 2♥ over 2♦. Is this not standard?

[Free]

Interesting. When I play this I pass with 4315 when partner responds 2♥ over 2♦. Is this not standard?

There's been a thread a few months ago about this approach.

[TimG]

1321281338[/url]' post='588786']

.

 

From my perspective, if folks are seriously interested in this topic, the best course of action would be to try to get Jack's developer's interested. (or any other serious software group)

 

There are a lot of advantages to being able to run this all using computer players. Once this has been coded up,

 

1. You can run large numbers of simulations at a relatively small cost

2. You have identically skilled players competing

3. You can play the same hands multiple times without worrying that folks with remember them

 

Won't you be stuck with the biases of the computer program used? With a sample from many players the problems of system context and individual strengths and weaknesses would be evened out.

[32519]

Has any new data on the Multi been gathered yet from those who agreed to assist Han? Its almost two weeks on since this project was started.

[Free]

I'm not sure it actually started, because we're still looking at the best way to gather and interpret any data.

[Cthulhu D]

Won't you be stuck with the biases of the computer program used? With a sample from many players the problems of system context and individual strengths and weaknesses would be evened out.

 

The bias's should cancel each other out because the same 'player' is in all 8 seats in the simulated teams match. The advantage is you can easily play a very large number of boards (say, 200+) with the same weak 2.

 

The way I'd do it is:

 

• Choose your 2 level opening, and determine which one of the three standard weak 2s it replaces.
  
• Caculate the relative frequence on the traditional weak 2 opening vs whatever you are opening.
  
• Create a reference set of 200+ boards split between traditional weak 2 openings and whatever your new pre-empt is in the ratio of the relative frequences of the two bids.
  

 

So if the tradiational weak 2D comes up (disclaimer: Number is made up), 4% of the time, and your Ekrens 2D comes up 6% of the time (disclaimer: number equally made up), you'd generate 80 boards with a weak 2 diamonds opener, and 120 boards with a Ekrens 2D opener. Obviously if your different structure incorporates multiple openings (say, a 2D mini-multi, Ekrens 2H, and some intermediate spades hand), you're going to need to do the reference boards split between all three weak 2s, the mini multi, ekrens 2H and some intermediate spades stuff.

 

The main issue would be the strength of the defence the computer was given against the 2<x> opening, and the difficulty the computer would have in responding to the pass/correct weak 2s in the 2h: Hearts or spades vein. I would suggest having a well recognised expert design the defence, and not using the very strange 2H: Hearts or Spades style weak 2s in testing.

 

While the computer bridge programs are weaker than top humans, this controls for field strength and bidding style very effectively.

[mycroft]

Meckwell could easily have changed off Multi also because they wanted to play together in the Reisinger and the Blue Ribbon Pairs, etc., and at least one of them knew about the change coming down the pike that meant they would have to have a different system to do so.

 

Only mildly sarcastic, and probably not really their concern, though.

[bluecalm]

If it could be any help I can dig up every hand ever played on vugraph by given pair opened in given way (or satisfying given distribution/points).

I didn't play with my db project for some months but if such data could help with that I will resurrect it a bit.

 

Can you conclude from this that they don't think multi is worth playing? No!

 

If we are to believe what they say in interviews they in fact think multi is worse than weak twos. They need one 2 level bid to plug precision leaks though so it's not comparable to dilemma most people have.

 

Won't you be stuck with the biases of the computer program used?

 

If some sensible defense could be programmed and executed well by the program (like dbl = t/o to spades, rest natural) then such data could be very valuable because of sample sized we could have. I think it would be more valuable than almost anything we can gather from human play.

 

Is it? I thought 2♥ gains a bit on 2♦ because opps are under more pressure to act, but it loses slightly because opener can pull 2♥ to 2♠

 

The biggest advantage of 2♦ over 2♥ is that well defined hand is in dummy after 2D - 4H.

[mwalimu02]

when i become a pro at this *smh* i think i'll be an addict :)