[MrAce]

http://tinyurl.com/y745s8lm

[wbartley]

MrAce, on 2017-July-17, 01:47, said:

http://tinyurl.com/y745s8lm

Not sure this is LOL-worthy but certainly an unscientific jump to slam with what appear to be minimum values.

[MrAce]

wbartley, on 2017-July-17, 10:09, said:

Not sure this is LOL-worthy but certainly an unscientific jump to slam with what appear to be minimum values.

It is actually LMFAO worthy. My hand makes slam, if not grandslam vs any hand that makes 5 ♣ cue when coming from pass. All it needs is the similar hand with 1435 or 1426 shape. Is it too much to ask for some shape for this cue? Of course I will be the one bidding slam because I am the one with 3 aces + trump Q.

Another thing that does not make sense in your comment is, you are talking as if you are happy with being at 5 level with this collection. I am not!
Only thing you wrote that make sense could be, it's not LOL worthy but LMFAO worthy!

x Kxxx xxx AKxxx GIB does not open these hands, which is fine, and this is a decent slam and probably grand slam on a finesse.

[cherdano]

x AQxx AKQxxx xx.
I agree 5C is bad. But LOL?

I think 6H is wrong - you do have minimum values. Yes, 4 keycards. But also not a great fit if partner has spade shortness as you hope.

[MrAce]

cherdano, on 2017-July-17, 13:33, said:

x AQxx AKQxxx xx.
I agree 5C is bad. But LOL?

I think 6H is wrong - you do have minimum values. Yes, 4 keycards. But also not a great fit if partner has spade shortness as you hope.

I do not even know what are you talking about with the hand you constructed. Nobody sane bids 4♥ with the hand you construct but I really don't wanna waste my time with this kind of BS, Arend! You do not make cues in case your pd has a 6 card side solid suit + singleton + enough keycards and decided to make a funny 4♥ bid!


FYI, I do not need ♠ shortness and I did not hope for it.
Since when did you lose your ability to count to 12? xx Kxxx xx AKxxx is all it needs for a decent slam. If clubs turn out to be 4-1 we can still make on ♦ finesse. I never said it is cold.
Now your comment makes me understand why I almost always get 80%+ score for slams that I think are biddable even for my grandma!
And is it too much to expect this from a hand coming from pass and cueing specificially 5♣ over your 4♥?
Of course it is LOL. What are we doing at 5 level vs his flat 8 hcp? Assume I bid only 5♥ and went down, would it make my LOL more accurate? No. 5 ♣ is a LOL.

[wbartley]

MrAce, on 2017-July-17, 11:46, said:

It is actually LMFAO worthy. My hand makes slam, if not grandslam vs any hand that makes 5 ♣ cue when coming from pass. All it needs is the similar hand with 1435 or 1426 shape. Is it too much to ask for some shape for this cue? Of course I will be the one bidding slam because I am the one with 3 aces + trump Q.

Another thing that does not make sense in your comment is, you are talking as if you are happy with being at 5 level with this collection. I am not!
Only thing you wrote that make sense could be, it's not LOL worthy but LMFAO worthy!

x Kxxx xxx AKxxx GIB does not open these hands, which is fine, and this is a decent slam and probably grand slam on a finesse.

I would expect a 2♣ bid with the hands you describe rather than 1♥ but I get your point. Also, in the image you posted, you're list as North. I guess you were South when the bidding was going on. It's not "immediately" obvious. Pretty hard to get that big MR ACE label on the North hand out of my mind when I'm trying to figure out what's going on.

My point about this example not being LOL worthy is that this overvaluation of a hand opposite a jump to 4H doesn't compare in absurdity to taking a preference to spades with a singleton holding 1543 when the auction has gone 1S=1N=3D.

When you started this series looked like you were attempting to point out some blatant issues with GIB but now I see it's just a forum for you to whine about how GIB isn't an expert.

[MrAce]

wbartley, on 2017-July-17, 15:58, said:

I would expect a 2♣ bid with the hands you describe rather than 1♥ but I get your point. Also, in the image you posted, you're list as North. I guess you were South when the bidding was going on. It's not "immediately" obvious. Pretty hard to get that big MR ACE label on the North hand out of my mind when I'm trying to figure out what's going on.

My point about this example not being LOL worthy is that this overvaluation of a hand opposite a jump to 4H doesn't compare in absurdity to taking a preference to spades with a singleton holding 1543 when the auction has gone 1S=1N=3D.

When you started this series looked like you were attempting to point out some blatant issues with GIB but now I see it's just a forum for you to whine about how GIB isn't an expert.

Yes, I was actually south.


No I am not really whining about GIB not being expert. It is just the inconsistency that frustrates me the most. You open 2 club and rebid 2 NT and GIb bids 3 NT with 11 hcp. And the cue bids with this flat 8 vs my possible flat 19 hcp! And then passes to your 1 NT overcall with 8 hcp and 5 card minor...And then the guy opens 2 NT with 17 hcp and GIB bids 3 NT with 11 hcp but you open the same hand 1 NT and GIb drives you to slam. This is not really whining about GIB not being expert. It is a pretty bad engine to be honest.

All i hope is in my lifetime to be able to play and compete against good players with a decent engine. Because I love the idea of being able to play vs same opponents with same pd.
This one has fundamental issues and even worse, i feel like (i am not sure 100 %) each time they patch something, they break something more important. That is all.

I called my series (LOL) because to me they are laughable mistakes. and GIB used to not do these things.

[MrAce]

wbartley, on 2017-July-17, 15:58, said:

When you started this series looked like you were attempting to point out some blatant issues with GIB but now I see it's just a forum for you to whine about how GIB isn't an expert.

Let me give you an example.

http://www.bridgebas...C2%7Cmc%7C10%7C


Had I been a whiner, I would whine on this one. I never published this BS. Because I found guilt on my part despite all the nonsense in the alert box. Had I seen my 4♠ being a cuebid, I would have left GIB alone.

[cherdano]

MrAce, on 2017-July-17, 11:46, said:

It is actually LMFAO worthy. My hand makes slam, if not grandslam vs any hand that makes 5 ♣ cue when coming from pass. All it needs is the similar hand with 1435 or 1426 shape.

Timo you are really offtrack here.
X kxxx kxx Axxxx is an obvious slam try, yet slam is horrible. I would argue xx Kxxx kxx Axxx is still a slam try...

For all your aces, you still have a balanced 18 count.

[Phil]

Thats funny, I had the exact hand today and I bid it the same as Timo but 6♥ is too much. 450 would have been a good score.

5♣ is a pretty awful call.

[MrAce]

cherdano, on 2017-July-17, 18:37, said:

Timo you are really offtrack here.
X kxxx kxx Axxxx is an obvious slam try, yet slam is horrible. I would argue xx Kxxx kxx Axxx is still a slam try...

For all your aces, you still have a balanced 18 count.

First hand bids 4♠ if wants to cue unless you are playing first round control only. Bidding 5♣ is a great way to lose slam when pd has xx(x) ♠
Second hand is not even remotely a slam try hand imo.
I do not think there is any hand that comes from pass and cues without concentrated values and some shape.
Does that mean we can make no slam with the second hand type you gave? Of course we can. And of course there will be much more hands that you are in danger at 5 level than the hands you may make slam.
But i should have stopped debating once the first hand you constructed was x AQxx AKQJxx xx for 4♥ bid. This shows you and I are living in totally different planets so I take your "you are offtrack" comment as a compliment. Because that hand alone makes me never want to be in same track with you

[cherdano]

Timo?

[MrAce]

Phil, on 2017-July-17, 19:16, said:

Thats funny, I had the exact hand today and I bid it the same as Timo but 6♥ is too much. 450 would have been a good score.

5♣ is a pretty awful call.

Agreeing with 6♥ or not debate hijacks the topic actually.
These topic are about the LOL actions of GIB. Not ATB topics. I am surprised people miss that.
I am also surprised people miss that the outcome of slam is more likely to make 12 tricks vs a real 5♣ cue than the outcome of 5♥ which is more likely to go down vs the current hand GIB cued. This is why I find 5 ♣ bid a LOL bid.

[Stephen Tu]

MrAce, on 2017-July-17, 16:18, said:

This one has fundamental issues and even worse, i feel like (i am not sure 100 %) each time they patch something, they break something more important. That is all.

I called my series (LOL) because to me they are laughable mistakes. and GIB used to not do these things.

You say this a lot, but I don't know that this is really supported by evidence. Certainly regression bugs (patch breaks something that worked previously) will happen occasionally, but I don't think they are as common as you think. The number of possible auctions, especially competitive ones, is immense. How can you be so sure it worked previously unless you actually get access to a previous version and verify? How can you clearly remember exact same auction and hand type that GIB got it right before? To me most stuff reported probably never really worked.

GIB has been doing LOL bids forever. I don't remember it being better before.

I think the main frustrations are certain classes bugs we complain about here that have gone on forever that we feel should be given more focus but have been unaddressed for years. Plus various periods in the past when BBO staff don't post on here for months without acknowledging looking at any of our reports or giving news on the next update.

Random LOLs that come up, unless they form a clear pattern, it's hard to squash them all.

[MrAce]

Stephen Tu, on 2017-July-17, 23:54, said:

You say this a lot, but I don't know that this is really supported by evidence. Certainly regression bugs (patch breaks something that worked previously) will happen occasionally, but I don't think they are as common as you think. The number of possible auctions, especially competitive ones, is immense. How can you be so sure it worked previously unless you actually get access to a previous version and verify? How can you clearly remember exact same auction and hand type that GIB got it right before? To me most stuff reported probably never really worked.

GIB has been doing LOL bids forever. I don't remember it being better before.

I think the main frustrations are certain classes bugs we complain about here that have gone on forever that we feel should be given more focus but have been unaddressed for years. Plus various periods in the past when BBO staff don't post on here for months without acknowledging looking at any of our reports or giving news on the next update.

Random LOLs that come up, unless they form a clear pattern, it's hard to squash them all.

How many GIB boards do you play average each week?

[Stephen Tu]

Not much recently, but in the few forum challenge events I played in recently, while I certainly saw a decent # of LOL boards, I don't feel like it was greater frequency than say 5-7 years ago when I played more often against the bots. They did tons of crazy stuff back then too. It's not like GIB was awesome before and went downhill. As Josh stated before they test versions vs. older version before release, the newer ones win easily. The number of auctions is just so enormous, you watch it do something nuts, think it didn't do that before, but really it's just that you never got that exact same auction before and the earlier version would screw up also. At least for most of these bugs. Certainly a few regressions are bound to slip through.

The more boards you play per week, of course the more bugs you will run into. Have you been keeping stats, for example, x number of LOLs per 100 boards played, for current version vs. say 3 years ago? If you are just going by gut feeling, it's easy to be wrong about whether the bots are getting better or worse.

Another thing that matters is advanced vs. the basic bots. The advanced bots in a lot of situations do some simulating that prevent it from choosing stupidities the basic bots get themselves into. The basic bots are good at revealing deficiencies in the book bids though, rules that shold be improved. So if you keep stats probably should separate the two.

[MrAce]

cherdano, on 2017-July-17, 21:21, said:

Timo?

I apologize, did not want to snap on you but was having a bad day really. Not even sure if having a bad day is a good excuse but it is what it is, I can't lie to you.

[johnu]

Stephen Tu, on 2017-July-18, 00:46, said:

The advanced bots in a lot of situations do some simulating that prevent it from choosing stupidities the basic bots get themselves into.

Simulations yes, enough simulations or simulations with the best parameters NO. One of the reasons I hated having the robots declare before there was human declare was I would get to a contract and there would be a 100% line to make the contract. GIB would take a line that worked maybe 50% to 90%. Apparently the exceptions to the inferior lines of play didn't come up in the simulations.

[cherdano]

MrAce, on 2017-July-18, 14:48, said:

I apologize, did not want to snap on you but was having a bad day really. Not even sure if having a bad day is a good excuse but it is what it is, I can't lie to you.

Thanks. Good enough for me!