[TimG]

Cascade, on Mar 11 2008, 03:15 PM, said:

Yes I have these working.

I think the problem is that bridge.exe might have been upgraded and the upgraded version does not work with dealer. This is some vague recollection of having problems with this several years ago.

I was using a bridge.exe from May of 1999. I got out the original disc and switched to the March 1999 version. Neither work. What is the date on your version?

Thanks.

[hotShot]

try: ftp://present.cirl.uoregon.edu/pub/users/.../bridge/gib.exe

Sorry! Seems not to work from here.

[Cascade]

TimG, on Mar 12 2008, 11:21 AM, said:

Cascade, on Mar 11 2008, 03:15 PM, said:

Yes I have these working.

I think the problem is that bridge.exe might have been upgraded and the upgraded version does not work with dealer.  This is some vague recollection of having problems with this several years ago.

I was using a bridge.exe from May of 1999. I got out the original disc and switched to the March 1999 version. Neither work. What is the date on your version?

Thanks.

April 1999

12/04/1999 9:52pm (DD/MM/YYYY format)

I am using a version of dealer named mdealer.exe which is dated 17/06/1999 9:30am

[TimG]

Cascade, on Mar 11 2008, 07:12 PM, said:

TimG, on Mar 12 2008, 11:21 AM, said:

Cascade, on Mar 11 2008, 03:15 PM, said:

Yes I have these working.

I think the problem is that bridge.exe might have been upgraded and the upgraded version does not work with dealer.  This is some vague recollection of having problems with this several years ago.

I was using a bridge.exe from May of 1999. I got out the original disc and switched to the March 1999 version. Neither work. What is the date on your version?

Thanks.

April 1999

12/04/1999 9:52pm (DD/MM/YYYY format)

I am using a version of dealer named mdealer.exe which is dated 17/06/1999 9:30am

Thanks again. The version of dealer.exe that I am using is dated 19 May 1999. (I modified and re-compiled it to include an alternate evaluation method, so the date isn't likely to match anyone else's version.)

I re-downloaded it today and attempted to compile, but cygwin gave me an error when I issued the "make" command.

What is the significance of the "m" in mdealer.exe?

[whereagles]

Well, I finally got it to work.

I was surprised by the results: on a 1♣ opener that's basically natural or any 21+ hcp, it seems like the 21+ hcp comes up 11% of the time. Gotta check the code... lol.

[TimG]

I still haven't managed to get dealer and GIB to work together. Can anyone point me to a source for mdealer.exe? From what I have read, that version might work, but I can't seem to find the file.

Thanks.

[awm]

Here's what I get. There are some things that are unclear about what Whereagles asked, so I will clarify with my assumptions after listing results. This is over one million random hands.

Balanced 12-14: 97,007
Balanced 12-14 that open 1C Natural: 42,114
Balanced 18-19: 13,245
Balanced 18-19 that open 1C Natural: 5,895
Balanced 20+: 7,751
Unbalanced Natural 1C with 11-20: 52,060
Unbalanced 21+: 3,597
One Suited 18-20 NOT CLUBS: 5,008

Option A: 171,318
Option B: 116,425
Option C: 160,415
Option D: 105,522

I've made the following assumptions:

(1) Balanced hands are hands with no singleton/void and at most one doubleton. This includes all 5332 patterns (even those with a 5cM) and excludes all 4225/6223 patterns.

(2) Balanced hands that open 1C natural are those where there is no five-card major, at least as many clubs as diamonds, and not 4-4 in the minors.

(3) Similarly, 4441 hands except for 4-4-1-4 exactly are assumed to open 1D.

(4) I am always assuming first seat, so not conditioning on opponents not having opened etc.

(5) One-suited hands are defined as those with a 6+ card suit, where every other suit is at least two cards shorter. So I have included 6421 and 7501 for example as "one-suited" hands.

[Cascade]

TimG, on Mar 17 2008, 06:02 AM, said:

I still haven't managed to get dealer and GIB to work together. Can anyone point me to a source for mdealer.exe? From what I have read, that version might work, but I can't seem to find the file.

Thanks.

I have put what I have here


mdealer.exe

It is zipped.

I assume it is ok since the program is in the public domain.

[TimG]

Thank you, thank you, thank you!

[whereagles]

awm: what do your results mean? That in case A 1♣ is strong 3,597 times out of 171,318?

[awm]

Everything is out of a million hands. For agreement A:

(1) You open 1♣ about 17% of hands, which is very high.
(2) The division is roughly:

10% strong artificial; 20+ hcp balanced or 21+ hcp any or 18-20 hcp one-suited not clubs.
30% unbalanced hands that open 1♣ in standard
25% balanced 12-14 that open 1♣ in standard
30% balanced 12-14 that open with 1♦ or 1M in standard
5% balanced 18-19 that open 1♣ in standard

Basically each of these percentages is the number of hands thus described divided by the number of A hands (i.e. divided by 171,000). The percentage of 1♣ openings is the same 171,000 divided by a million.

What Whereagles said is basically right, if you replace "strong" with "21+ hcp and unbalanced." Most of the strong hands will be 20+ balanced hands or 18-20 one-suited hands (either of which is more frequent than 21+ and unbalanced).

[whereagles]

Ah.. allright, I see it. It's more or less the same conclusion I got to.

So basically 1♣ will be strong 10% of the times if there's no 12-14 balanced in the fray, and drops to about 7% if you dump that hand into it.

Still surprising, though... I was expecting something more like 2-3%.