Are you kidding me?

[iandayre]

I couldn't believe this one. Especially since my memory tells me I have seen GIB raise in this situation with less.

 

http://tinyurl.com/nuju26c

 

PASS??? How dare you not bid 4H!!!

[iandayre]

I couldn't believe this one. Especially since my memory tells me I have seen GIB raise in this situation with less.

 

http://tinyurl.com/nuju26c

 

PASS??? How dare you not bid 4H!!!

 

The ones who reached game did so one of four ways: Bid 3H and then 4H when 3S came back around, Bid 4H originally, Bid 3H, then 4C when it came back around. GIB now came to life with 5H which made fortunately for those players, and lastly, some doubled 2S. Not sure which is the worst action but all of them are terrible in my opinion. Slightly more than half shared the bottom with me by sensibly passing out 3S.

[spkcp111]

The robot has gotten SIGNIFICANTLY worse since the latest release.

 

I am shocked at how poor the bidding is! The robot continually passes, refuses to bid NT with a stopper in the opps suit. It is shocking! I feel like I'm playing with a NOVICE--not a computer.

 

It is HORRIBLE. I'm tired of spending my $ to play with such an idiot!

 

And finesses are supposed to fail 50% of the time, but you can COUNT on the finesses failing in bridgebase. It's the one thing you can count on. It's just ridiculous.

 

In my opinion? They are training people to play "bad bridge". Rewarding bad bridge cannot be good for the game.

[1eyedjack]

3 comments

 

1) GIB has been upgraded since the hand that initiated this thread. It is possible that your rant is intended to be a polemic against the latest version, in which case I suggest that you should have started a new thread. Or at least one that started on GIB version 31.

 

2) I have seen no evidence that finesses fail (or succeed for that matter) with anything other than a statistically insignificant departure from 50%, which is my experience. I suspect that Bayes has something to do with it. You are in 3N with AQx in declarer opposite xxx in Spades and get a Heart lead. Here I would expect the finesse to be less than 50%. GIB doesn't like leading away from Kings. GIB also likes to put you to the guess early when the finesse is working, at a time when you may prefer not to take it (in favour of an alternate line).

 

Presumably you eschewed the percentage (and working) trump finesse on this hand, on the grounds that with GIB they never work?

 

 

[hv=bbo=y&lin=pn|spkcp111,~~M39023,~~M39021,~~M39022|st%7C%7Cmd%7C4S2H679JQAD34KC24A%2CS34679KH3D69C58TQ%2CSTJQH248TD258AC9J%2C%7Crh%7C%7Cah%7CBoard%206%7Csv%7Ce%7Cmb%7C1D%7Can%7CMinor%20suit%20opening%20--%203%2B%20%21D%3B%2011-21%20HCP%3B%2012-22%20total%20points%20%7Cmb%7C1H%7Can%7COne-level%20overcall%20--%205%2B%20%21H%3B%208-17%20HCP%3B%209%7Cmb%7C1S%7Can%7CFree%20bid%20--%205%2B%20%21S%3B%206%2B%20total%20points%3B%20forcing%20%7Cmb%7C3D%7Can%7CMixed%20raise%20--%204%2B%20%21H%3B%207-9%20total%20points%3B%20forcing%20%7Cmb%7Cp%7Cmb%7C4H%7Can%7C5%2B%20%21H%3B%208-17%20HCP%3B%209-19%20total%20points%7Cmb%7Cp%7Cmb%7Cp%7Cmb%7Cp%7Cpc%7CD9%7Cpc%7CD2%7Cpc%7CD7%7Cpc%7CDK%7Cpc%7CHA%7Cpc%7CH3%7Cpc%7CH2%7Cpc%7CH5%7Cpc%7CH6%7Cpc%7CS9%7Cpc%7CHT%7Cpc%7CHK%7Cpc%7CDQ%7Cpc%7CD3%7Cpc%7CD6%7Cpc%7CDA%7Cpc%7CSQ%7Cpc%7CSA%7Cpc%7CS2%7Cpc%7CS4%7Cpc%7CDJ%7Cpc%7CD4%7Cpc%7CS3%7Cpc%7CD5%7Cpc%7CC6%7Cpc%7CCA%7Cpc%7CCT%7Cpc%7CC9%7Cpc%7CH7%7Cpc%7CS6%7Cpc%7CH8%7Cpc%7CC3%7Cpc%7CD8%7Cpc%7CDT%7Cpc%7CH9%7Cpc%7CC5%7Cpc%7CC2%7Cpc%7CCQ%7Cpc%7CCJ%7Cpc%7CC7%7Cpc%7CSK%7Cpc%7CST%7Cpc%7CS8%7Cpc%7CHJ%7Cmc%7C9%7C]400|300[/hv]

 

 

3) GIB has been tested and shown to play a 57% game (at least the advanced version). And that was before the latest upgrade. Your GIB games appear to be exclusively robot tourneys. May I point out that you are competing not against robots in these tourneys but against human opponents sitting in exactly the same seat as you and being presented with exactly the same problems to solve, with the other three table incumbents behaving identically to those at your table. It may not be bridge as you know it, and success may require some familiarity with GIB predilections (or is that peccadillos?). But with that caveat, skill is definitely rewarded. Some think that that is worthy of a $ entry fee.

 

My highly unscientific observation to date is that the latest upgrade may be a little overhyped in its overall extent of improvement, but it is early days. I hope we don't have to wait for the same length of time until the next upgrade.

[Bbradley62]

Rather than ranting and raving, maybe posters can try to make constructive suggestions as to how to fix specific problems.

 

It seems to me that the problem with the posted hand is that the lower limit of the 3♥ overcall is too low... to overcall at the 3 level, vul vs non-vul opposite a passed partner, should indicate a minimum of 14 or 15 total points rather than 12. With this knowledge, North should be more likely to raise to 4♥.

[Bbradley62]

And finesses are supposed to fail 50% of the time, but you can COUNT on the finesses failing in bridgebase. It's the one thing you can count on. It's just ridiculous.

Since this seems to be a popular subject for infrequent posters who choose to vent their spleens here, I decided to do a little analysis on hands that I have played at tables with 3 robots. I've played 88 such hands (excluding video bridge hands which are not included in MyHands) in the past month (an admittedly small sample size) and found the following in hands that my side declared:

• Of the 23 times I finessed to find a missing A, that A was onside 15 times (65%)
  
• Of the 31 times I finessed for a missing K, that K was onside 20 times (65%)
  
• Of the 9 times I had a one-way finesse for a missing Q, that Q was onside 5 times (56%)
  
• Of the 6 times I was missing KQ and needed at least one onside, I was successful 4 times (67%)

If you can produce some actual statistics about finesses failing too often, maybe someone will take you seriously.

[johnu]

Since this seems to be a popular subject for infrequent posters who choose to vent their spleens here, I decided to do a little analysis on hands that I have played at tables with 3 robots. I've played 88 such hands (excluding video bridge hands which are not included in MyHands) in the past month (an admittedly small sample size) and found the following in hands that my side declared:

• Of the 23 times I finessed to find a missing A, that A was onside 15 times (65%)
  
• Of the 31 times I finessed for a missing K, that K was onside 20 times (65%)
  
• Of the 9 times I had a one-way finesse for a missing Q, that Q was onside 5 times (56%)
  
• Of the 6 times I was missing KQ and needed at least one onside, I was successful 4 times (67%)

If you can produce some actual statistics about finesses failing too often, maybe someone will take you seriously.

 

Obviously the reason spkcp111 loses so many finesses is that people like you are hogging the successful finesses. In the future, please be fair and only win your fair share of finesses so the rest of us don't have to suffer.