[eamongall]

Good evening folks

I first started using GIB while working in Saudi Arabia in 1996. Some guy called Matt Ginsberg sent me a copy of GIB.
In those days it was a DOS program with loads of switches.
This was before the commercial release of GIB. I was a beta tester.

Even in 1996 GIB played damn well and I still tinker with the DOS version of GIB.

A guy called Stephen Pickett made a front end for GIB called Gibbrowser ...
He also had a program called Bridgebrowser.

GIBbrowser was wonderful as it allowed one replay hands from the other major online bridge outfit run my a chap called Matthew.

Nowadays we have Bridge Composer which allows all sorts of handling of bridge hands.

Gib works by dealing lots of hands to match the bidding so far and then chooses the bid which works the most often.

Lots of times while running the DOS version one sees "hand generation failure set 1 " or "hand generation failure set to 2 "
This means GIB is not finding hands to suit the bidding ..

However Jack the dutch program could always beat GIB badly in any tests I ever ran.

GIB actually has TWO play engines ... GIB and GIBSON ..
if the program can get thru the first 2 tricks without dropping a trick then the more POWERFUL GIBSON kicks in

When using the DOS version one can see it typing ..Switching to GIBSON around trick 3.
On the old GIB client one might recall GIB pausing about trick 3 ... this was GIBSON kicking in.

For the bidding ... GIB relies on a bidding database and some simulation ..
There was a version of GIB that used to think on opponents time but I don't think it lasted or else it confused itself.
It has some programming that allows it to kick in the bidding simulation ...
I think it kicks in when it gets no stand out bid.

The robot on BBO is quite crippled and does not have all the functions of the DOS version ..
The advanced robot probably has more time allowed for GIB and thus more strength can kick in.

However as far as I know the executable program has had little change since about 2001.
BBO I think works on the bidding database and adds/adjusts some rules.

I do recall when we started using WEB based BBO the robots played like lightning ...far too fast to be doing any serious calculation.
The bloody robot was faster than the display could handle.

The robot was finished the hand before the display could catch up

However all in all ... I have had a great kick out of my GIB robot.
but Jack program is much better.
However GIB was the first program to bring bridge programming into the NEW AGE ... ala Monte Carlo and Hill=climbing

This is just a history exercise though ... I am not touting up GIB.

Thanks for reading
Eamon Galligan
eamongall on BBO.

[pescetom]

Thanks for sharing that with us.

[virgosrock]

eamongall, on 2018-May-17, 15:30, said:

Good evening folks

I first started using GIB while working in Saudi Arabia in 1996. Some guy called Matt Ginsberg sent me a copy of GIB.
In those days it was a DOS program with loads of switches.
This was before the commercial release of GIB. I was a beta tester.

Even in 1996 GIB played damn well and I still tinker with the DOS version of GIB.

A guy called Stephen Pickett made a front end for GIB called Gibbrowser ...
He also had a program called Bridgebrowser.

GIBbrowser was wonderful as it allowed one replay hands from the other major online bridge outfit run my a chap called Matthew.

Nowadays we have Bridge Composer which allows all sorts of handling of bridge hands.

Gib works by dealing lots of hands to match the bidding so far and then chooses the bid which works the most often.

Lots of times while running the DOS version one sees "hand generation failure set 1 " or "hand generation failure set to 2 "
This means GIB is not finding hands to suit the bidding ..

However Jack the dutch program could always beat GIB badly in any tests I ever ran.

GIB actually has TWO play engines ... GIB and GIBSON ..
if the program can get thru the first 2 tricks without dropping a trick then the more POWERFUL GIBSON kicks in

When using the DOS version one can see it typing ..Switching to GIBSON around trick 3.
On the old GIB client one might recall GIB pausing about trick 3 ... this was GIBSON kicking in.

For the bidding ... GIB relies on a bidding database and some simulation ..
There was a version of GIB that used to think on opponents time but I don't think it lasted or else it confused itself.
It has some programming that allows it to kick in the bidding simulation ...
I think it kicks in when it gets no stand out bid.

The robot on BBO is quite crippled and does not have all the functions of the DOS version ..
The advanced robot probably has more time allowed for GIB and thus more strength can kick in.

However as far as I know the executable program has had little change since about 2001.
BBO I think works on the bidding database and adds/adjusts some rules.

I do recall when we started using WEB based BBO the robots played like lightning ...far too fast to be doing any serious calculation.
The bloody robot was faster than the display could handle.

The robot was finished the hand before the display could catch up

However all in all ... I have had a great kick out of my GIB robot.
but Jack program is much better.
However GIB was the first program to bring bridge programming into the NEW AGE ... ala Monte Carlo and Hill=climbing

This is just a history exercise though ... I am not touting up GIB.

Thanks for reading
Eamon Galligan
eamongall on BBO.

I am shocked a country like Saudi Arabia allowed an intoxicant like GIBBy.
Seriously, thanks for this info. Did you ever question Matt G about GIBBO's bidding especially in competition, AND defensive signalling and carding?

thanks
vrock