[thepossum]

Deleted due to attitude of responders

[smerriman]

GIB bids very poorly when it comes to slam auctions. It doesn't understand the principles of cuebidding (as you can see 4♠ is defined as 26+ total points which is literally impossible) or what high level bids mean.

Here you need a way of telling if partner has the diamond queen. I'll leave it up to somebody else for how expert humans would handle this, since I don't know. But I can tell you it is impossible to find out with GIB.

If you're solely looking to improve your scores with GIB, you're often better off bidding badly / lying about your hand in order to take control of the auction (ie, bid 4NT yourself instead of 4S, hear partner has two aces, then take a wild guess what to play).

This is why bidding with GIB can hurt your game in general - but it's still fun to learn how to outwit the robot.

PS - the probability of making grand slam with your hand and dummy is 52%. This means if you're playing matchpoints, 7♣ (well, I guess 7NT) is the correct contract, even though it goes down on this layout - but you don't want to be in it if playing IMPs, which needs to be at least 55%.

Edit - good point, Cyberyeti. It's 57.9%. Maybe minus a tiny bit for a ruff in another suit. So in fact, grand slam might be the right spot on these hands. Just not the right way to get there.

[Cyberyeti]

It's actually a bit better than 52%, you can pick up Q9xx onside as well as 2-2 and stiff Q.

I wouldn't worry too much about being in this one, 7♣ is a ridiculous bid, you could have something like AKx, Qx, AKx, KQxxx where it's almost no play.

[The_Badger]

Just consider yourself unlucky, possum, with a bot bid that turned out to be a rot(ten) bid. At least you ended up in a slam, albeit one level too high. Those who ended up in just 5♣ with the North hand seriously underbid their (two ace) hand opposite a 2NT opener with a fit, though the irony of all this is that 5♣ is the par contract on best defence: ♦lead, ♣ruff (though who on earth would lead a ♦ without a Lightner double?)

[Cyberyeti]

 The_Badger, on 2018-November-02, 03:49, said:

Just consider yourself unlucky, possum, with a bot bid that turned out to be a rot(ten) bid. At least you ended up in a slam, albeit one level too high. Those who ended up in just 5♣ with the North hand seriously underbid their (two ace) hand opposite a 2NT opener with a fit, though the irony of all this is that 5♣ is the par contract on best defence: ♦lead, ♣ruff (though who on earth would lead a ♦ without a Lightner double?)

If you know how the diamonds break, 6♣(N) or 6♦ are par

[Tramticket]

 thepossum, on 2018-November-02, 02:33, said:

As with most of my posts I'm not looking at critique of my play

Hi Thepossum,

You post regularly here, so you are clearly looking to improve. So my advice, given in the spirit of trying to help is:
- Most newer players focus too much on working on bidding (particularly systems) and too little on declarer play and defence.
- You can improve your results more quickly by focusing on declarer play. (your bidding will even improve, because understanding how to play a hand helps you value a hand properly).
- You won't learn good bidding habits from BBO robots.
- You can and should learn good declarer play techniques from playing on BBO, because BBO gives you the unique ability to replay a hand trick by trick analysing exactly where you went wrong, including comparing and contrasting with other players who have played the same hand.

Looking at the current hand, I would make the following comments:
- The robot's 7♣ bid is nuts, Nothing more needs to be said.
- Much of your play looks fine, but good technique would have limited you to one-off:

You correctly drew trumps and eliminated the clubs, this was good technique. You should have also eliminated the hearts before tackling diamonds. This is good technique for two reasons: (1) you might find out something useful about the hand the hand by playing on a side suit (if an opponent discards, this might give you a better count of the distribution). (2) You take away the opponents' exit cards. So I would have won the opening lead, drawn trumps in two rounds, ruff a heart, cash the two spade winners and play the ace of diamonds before ruffing the last heart and finishing in dummy. On this occasion you will get the bad news when you play the ace of diamonds, but cashing the ace first allows you to drop a singleton queen with west. If both players had followed small you would need to decide whether to take a second-round finesse or play for the drop. The odds are very close, but without better clues, playing for clubs 2-2 is slightly better odds. On the actual cards, you should take a second round finesse in diamonds. You know it will lose, but because you eliminated the other two suits, west will have no choice but to play the diamonds back to you or give you a ruff and discard. you will save one trick.

Good luck!

[The_Badger]

 Cyberyeti, on 2018-November-02, 03:55, said:

If you know how the diamonds break, 6♣(N) or 6♦ are par

Absolutely right, Cyber. Thank you. I was just looking at South's point of view given the bidding to date.

[pescetom]

 smerriman, on 2018-November-02, 02:48, said:

Here you need a way of telling if partner has the diamond queen. I'll leave it up to somebody else for how expert humans would handle this, since I don't know. But I can tell you it is impossible to find out with GIB.

I'm more human than expert, but I would start by bidding 4♦ instead of 4♣
Even GIB understands RKCB and Queen asks.

[Left2Right]

Let's do the math. The risk-reward ratio for a small slam is anything better than 50%. For a grand, it's more than 67%. Thus, the 59ish correctly worked out in the forum doesn't justify the risk of seven.

The pros tell me that since it's too tough to calculate that 68% when you're sitting at the table, we must instead discover that 13th trick, in this case, the diamond queen.

One method for this is "6 Ace Roman Keycard Blackwood."
Try this link to read all about it.

And no, GIB doesn't know it.

[smerriman]

 Left2Right, on 2018-November-02, 23:35, said:

Let's do the math. The risk-reward ratio for a small slam is anything better than 50%. For a grand, it's more than 67%. Thus, the 59ish correctly worked out in the forum doesn't justify the risk of seven.

That's not correct. When small slam is certain, a vulnerable minor suit grand needs to be 55% to make it profitable at IMPs (13 IMP gain vs 16 IMP loss).

It would go up to your 68% if you thought three quarters of the field would be in just game, but surely everyone will be in at least 6 here.

[nekthen]

My 2 cents. I believe this works with humans too. Given the choice I would bid 4♦ because we can then discover if partner holds the Q. We can’t do that if they are not trumps

[Cyberyeti]

 smerriman, on 2018-November-03, 00:42, said:

That's not correct. When small slam is certain, a vulnerable minor suit grand needs to be 55% to make it profitable at IMPs (13 IMP gain vs 16 IMP loss).

It would go up to your 68% if you thought three quarters of the field would be in just game, but surely everyone will be in at least 6 here.

21 opposite 8 in a field of unknown quality by no means everybody will be in 6.

[pescetom]

 Left2Right, on 2018-November-02, 23:35, said:

One method for this is "6 Ace Roman Keycard Blackwood."
Try this link to read all about it.

See also Another link to understand better.
Bridgeguys reads like it was translated from Russian to English by a Spaniard who never played bridge

[smerriman]

 Cyberyeti, on 2018-November-03, 04:28, said:

21 opposite 8 in a field of unknown quality by no means everybody will be in 6.

Everyone will open 2NT. And every North in the field is guaranteed to bid minor suit Stayman. I can't think of any sequence from there that won't end in slam.

[dsLawsd]

The hand has a nasty flaw- xxx in hearts. So I underbid a bit knowing that GIB doesn't use logic insofar as suit quality and often jumps to ridiculous grands. So 4 dia and try to stop in some 6. So much for getting better- for that we need a regular partner!

[Cyberyeti]

 smerriman, on 2018-November-03, 13:43, said:

Everyone will open 2NT. And every North in the field is guaranteed to bid minor suit Stayman. I can't think of any sequence from there that won't end in slam.

About 10% of club players even have minor suit stayman available where I am.

[rmnka447]

Realistically, real bridge players will only bid grand slams when they can count 13 tricks or know the odds of making are at least 67% or better. The robot's 7 ♣ bid is completely from outer space.

There's no problem with your bidding other than possibly bidding 4 ♦ instead of 4 ♣. But 4 ♣ wasn't the problem on this hand.

[smerriman]

 Cyberyeti, on 2018-November-03, 15:52, said:

About 10% of club players even have minor suit stayman available where I am.

Agreed. Am in the context of this being a robot tournament.

[thawp66]

 Tramticket, on 2018-November-02, 04:07, said:

Hi Thepossum,

You post regularly here, so you are clearly looking to improve. So my advice, given in the spirit of trying to help is:
- Most newer players focus too much on working on bidding (particularly systems) and too little on declarer play and defence.
- You can improve your results more quickly by focusing on declarer play. (your bidding will even improve, because understanding how to play a hand helps you value a hand properly).
- You won't learn good bidding habits from BBO robots.
- You can and should learn good declarer play techniques from playing on BBO, because BBO gives you the unique ability to replay a hand trick by trick analysing exactly where you went wrong, including comparing and contrasting with other players who have played the same hand.

Looking at the current hand, I would make the following comments:
- The robot's 7♣ bid is nuts, Nothing more needs to be said.
- Much of your play looks fine, but good technique would have limited you to one-off:

You correctly drew trumps and eliminated the clubs, this was good technique. You should have also eliminated the hearts before tackling diamonds. This is good technique for two reasons: (1) you might find out something useful about the hand the hand by playing on a side suit (if an opponent discards, this might give you a better count of the distribution). (2) You take away the opponents' exit cards. So I would have won the opening lead, drawn trumps in two rounds, ruff a heart, cash the two spade winners and play the ace of diamonds before ruffing the last heart and finishing in dummy. On this occasion you will get the bad news when you play the ace of diamonds, but cashing the ace first allows you to drop a singleton queen with west. If both players had followed small you would need to decide whether to take a second-round finesse or play for the drop. The odds are very close, but without better clues, playing for clubs 2-2 is slightly better odds. On the actual cards, you should take a second round finesse in diamonds. You know it will lose, but because you eliminated the other two suits, west will have no choice but to play the diamonds back to you or give you a ruff and discard. you will save one trick.

Good luck!

 dsLawsd, on 2018-November-03, 14:10, said:

The hand has a nasty flaw- xxx in hearts. So I underbid a bit knowing that GIB doesn't use logic insofar as suit quality and often jumps to ridiculous grands. So 4 dia and try to stop in some 6. So much for getting better- for that we need a regular partner!

Hi, Tramticket. I agree with almost all of your comments and analysis. However, I have a few critiques. I’m almost certain that your comment near the end “...clubs 2-2...” was supposed to be “...diamonds 2-2....”.

As to the last 2 sentences and -1 vs -2, it seems academic. We have 6 trumps (with 2 heart ruffs) and 3 top tricks in the majors. We needed 3 of 4 diamond tricks. The fifth diamond is irrelevant. Though everything you said about technique seems reasonable, in this layout I don’t see an endplay. After playing the DA, simply playing diamonds top down gets the necessary D tricks. Starting the D suit by leading the T lead “Thepossum” to -2

To make 7, we need the first 4 diamonds. Cyberyeti’s final comments that this is about 58%, a reasonable contract, but a horrible auction, all seem reasonable.

As to your comment about not learning good bidding from the robots and “dsLawds’s” comments about GIB and needing a regular partner to get better, I have some thoughts.

Does having 1 regular (human) partner improve your play or your results, and are these mutually exclusive? If you focus on having a strong partnership system, with various conventions and defensive signals, and you play enough to refine these, your results will probably improve significantly. This seems also true if you play a lot with GIB. By learning their system and tendencies, you can improve your results.

However, what these have in common is that may not be transferable. That is to say, what happens when you then play with another partner, or in an an individual? If your results seem to be the same as what they were beforehand, it might indicate your play hasn’t improved.

For example, as many have commented about this hand, it highlights a weakness in GIB slam bidding and leads to a conclusion that bidding 4D might lead to better results. It’s debatable whether this is a better call anyway, but it may lead to an interesting conclusion; sometimes the “normal” call is not the the “best” call when considering your partner’s system and tendencies. However, in order to learn what these are, you need to have a regular partner (human or GIB)!

In order to improve your play independently of partner (or opponents), there are many things to (and not to) do. Playing regularly, a lot, and with players who are stronger than you will all help. Learn to count and observe everything.

[Tramticket]

 thawp66, on 2018-November-05, 05:16, said:

in this layout I don’t see an endplay. After playing the DA, simply playing diamonds top down gets the necessary D tricks. Starting the D suit by leading the T lead “Thepossum” to -2

Oh, I agree. As I advised, you should cash one round of diamonds first and this would show up the 4-0 break.

But good technique, eliminating the other two suits, would allow a partial recovery from leading the 10 initially and he could still have escaped for -1. The trick is to get in the habit of making these routine plays.

As to your other comments, my assumption is that the objective is to play better bridge not, "get better results partnering a robot". If playing better with robots is an end in itself, then fair enough. But, it's mot for me.