[dwar0123]

Ok, maybe I should have looked at the point range I was showing before accepting the quantitative invite, even though the displayed point range was flatly forbidden by the auction it would have at least caused me to not accept.

However, I believe the robots would be less frustrating to play with if they didn't make such invites with such hands.
Perhaps either have the robot actually have the 5-6 points it shows instead of the 2 points it had or lower the range of the 3nt call to something that is actually physically possible.

---

This was an advanced bot.

[mike777]

again people complain about the bots bid but they grossly misbid.

We see these examples all the time.

I can understand why wc players prefer to play with bots againt everyone else.

[fuburules3]

At least E decided his hand wasn't good enough for a double.

[dwar0123]

mike777, on 2012-April-09, 19:55, said:

again people complain about the bots bid but they grossly misbid.

We see these examples all the time.

I can understand why wc players prefer to play with bots againt everyone else.

I would argue the 3nt bid was reasonable, maybe you wouldn't do it, but its hardly a gross misbid. I know where all the points are and I have a fair shot of endplaying him a couple of times to score 9 tricks. I in fact made 3nt opposite my 2 point partner.

The 6nt bid is irrelevant as it had no influence on the bots 4nt bid. For the sake of argument pretend I passed and went down 1.

If it really bothers you what I did after the robot made a gross misbid, chalk it up to frustration, regardless it is irrelevant to the computers bid.

[Stephen Tu]

3nt is a pretty gross bid IMO, "making 3nt" isn't much of a justification of the bid given the bots gross misdefense; supposedly they were supposed to hold you to 1nt except for east's mysterious diamond exit (the "3nt" and "6nt" probably made east think west can't have QJs).

Anyway when humans make weird bids that shouldn't come up often (or at all, in some cases), it's more likely to hit a portion of the database that hasn't been thoroughly debugged.

That said, I do think there are some adjustments to be made here in the DB here:
- 2h shouldn't show "rebiddable hearts", maybe just 5 hearts, and the points should be limited to something like 0-4 hcp, maybe 6+ total points
- 3nt probably shouldn't show 25+ hcp. It's most likely a running suit + stoppers, don't know if that can be defined in GIB's bidding language

- probably in general there should be checks for "too many hcp" auctions, some way for GIB to think in terms of "7 or 8 cd suits", massive distribution, instead of large # of HCP if it seems there are too many in the deck from the bidding. Any place in the DB where things are defaulting to be defined as "25+ hcp" are probably suspect.

[dwar0123]

Stephen Tu, on 2012-April-09, 21:37, said:

3nt is a pretty gross bid IMO, "making 3nt" isn't much of a justification of the bid given the bots gross misdefense; supposedly they were supposed to hold you to 1nt except for east's mysterious diamond exit (the "3nt" and "6nt" probably made east think west can't have QJs).

Anyway when humans make weird bids that shouldn't come up often (or at all, in some cases), it's more likely to hit a portion of the database that hasn't been thoroughly debugged.

That said, I do think there are some adjustments to be made here in the DB here:
- 2h shouldn't show "rebiddable hearts", maybe just 5 hearts, and the points should be limited to something like 0-4 hcp, maybe 6+ total points
- 3nt probably shouldn't show 25+ hcp. It's most likely a running suit + stoppers, don't know if that can be defined in GIB's bidding language

You hit upon what the real problem may be. The computer lied about the potential strength of his hand. I know that for most people, 2 anything in this auction is a bailout showing a crap hand, but I do not presume to know what gib is doing(i've been punished a few times for that as well). I read 11- and twice rebiddable hearts and suddenly my 3nt seems a lot more reasonable?

You may say my 3nt bid is foolish but it is being induced by inaccurate descriptions.

Quote

- probably in general there should be checks for "too many hcp" auctions, some way for GIB to think in terms of "7 or 8 cd suits", massive distribution, instead of large # of HCP if it seems there are too many in the deck from the bidding. Any place in the DB where things are defaulting to be defined as "25+ hcp" are probably suspect.

Thanks, I am not going to presume to suggest how this gets solved, I am just annoyed that some people here would excuse poor gib behavior just because a less then expert player is reporting it. No matter how bad I am, that is no excuse for bidding 4nt with that hand.

[Stephen Tu]

Even with the "11- hcp" description somewhat misleading you, I think 3nt still isn't reasonable. Your RHO opened 1nt, and it's a robot that doesn't psych. So partner's range is 0-5, and on average will have less than half of that.

4nt is pretty weird. It's probably because of the rather loose definition of 3nt as "5- hearts", probably 3nt should be 2- hearts and it doesn't think some miracle is happening and getting 5 heart tricks when you have AKQx.

So tighten up the 3nt definition and probably it passes and you go down two.

[dwar0123]

Stephen Tu, on 2012-April-09, 22:21, said:

Even with the "11- hcp" description somewhat misleading you, I think 3nt still isn't reasonable. Your RHO opened 1nt, and it's a robot that doesn't psych. So partner's range is 0-5, and on average will have less than half of that.

I really feel like you can't appreciate this problem from my perspective because you know on some fundamental level that 2♥ was a busted hand. Change his Q♦ into a K and I am cold with a marked finesse. 3 points. I really don't think you are viewing what 11- hcp and twice rebiddable hearts looks like to someone who isn't as sure as you are that 2♥ means i have nothing.

I think we have to also tighten up that description of the 2♥ bid.

[Bbradley62]

---

As you can see from the auction (and descriptions) at another table, there is a good second bid for your hand which is not to unilaterally jump to 3N opposite a potential yarborough.

I'd be willing to bet that the "11- HCP" and "12- total points" are remnants of the fact that North is a passed hand. I wonder what strength info would be included if he were not.

Of course, this does not change the fact that the "rebiddable hearts" part of the description is nonsense, and the 4NT bid was horrendous.

[pigpenz]

unless were playing with a pinochle deck this auction is impossible.
Having started in the midwest with Alan Stout, David King, and others when partner doubles 1NT if you have anything at all you sit for 1NT doubled...so by bidding 2♥ your pretty broke. I think whole process started with the 3NT bid most likely there can only by 3-5 hcp between west and north, if partner had those 5hcp he would have sat for 1NTX....2NT= may be more than you get for 1NTX, takes small diamond out for max tricks on this hand.

[dwar0123]

Well, thanks all for the critique of my bidding, even though I posted this in the GIB robot discussion forum, it is nice to know that everyone feels the need to repeat each other on my bidding problems. Message duly received, thanks.

Now if we could get back on topic, cover up the south hand and ask yourself these 3 questions.

1. Is the description of 2♥ bid as accurate as it should be? Pretend that people play with robots that know less then you about bidding.
2. Is the description of the 3[nt] bid as having 25-30 points and less then 5♥ accurate? Keep in mind that there are plenty of legitimate hands south could have to make that bid, 7 running clubs and two outside aces for instance.
3. Is bidding a quantitative nt here wise?

[Stephen Tu]

On your 3 questions I basically already agreed with you in my previous posts. I don't think anybody here disagrees with those issues.

You are the one dragging it off topic by repeatedly trying to justify your 3nt bid. When you keep on submitting defenses to your bid, that aren't particularly convincing, you basically invite repeated critiques of it. You should either stop trying to claim 3nt as reasonable, admit that upon further reflection it was unwise, or come up with better justifications, or just drop the subject and not mention it again. If you can't take the heat and defend your positions then don't get involved in internet debates.

BTW even giving partner DK doesn't make it "cold" as you claimed. You still need CJ to come down which although a favorite isn't 100%. And LHO can still have DQ if RHO has SQJ, so the hook wouldn't necessarily be "marked".

[dwar0123]

Stephen Tu, on 2012-April-10, 14:29, said:

On your 3 questions I basically already agreed with you in my previous posts. I don't think anybody here disagrees with those issues.

Yes, I appreciate your constructive contribution to the thread. I fully understand that 3nt wasn't a reasonable bid for you. I fully understand that you are a better player and that you are correct. I also fully understand that you are incapable of appreciating how to a lesser player the 2♥ bid is a frustrating example of gross inaccuracies that lead to confusion and unnecessary misunderstandings. I am sure you also do not think its reasonable that a player who has never seen a deck of cards before could miss a guard squeeze. I get it, we are not on the same page here. Given the inaccurate description of 2♥ it is not reasonable for anyone to make that mistake because the only reasonable thing is that everyone is good enough to overlook that description. IE. I shouldn't be paying to play with gibs because clearly I'm not good enough.

Quote

You are the one dragging it off topic by repeatedly trying to justify your 3nt bid. When you keep on submitting defenses to your bid, that aren't particularly convincing, you basically invite repeated critiques of it. You should either stop trying to claim 3nt as reasonable, admit that upon further reflection it was unwise, or come up with better justifications, or just drop the subject and not mention it again. If you can't take the heat and defend your positions then don't get involved in internet debates.

Well thanks for dragging it off topic again, despite the fact that in my previous post, contrary to your claim, I offered no defense of my bidding. It is almost like I dropped the subject and choose not to mention it again.

And really, I kept going back to my 'reasonable' 3nt bid in an attempt to show how the 2♥ bid is confusing. I don't care how reasonable it is to you, you are at a different point in the evolution of your game. I didn't come here for critique of my bidding, I came here to help try and make GIB better, yet everyone seems intent on pointing out my flaws, as if they matter.

Quote

BTW even giving partner DK doesn't make it "cold" as you claimed. You still need CJ to come down which although a favorite isn't 100%. And LHO can still have DQ if RHO has SQJ, so the hook wouldn't necessarily be "marked".

Yea, the hand was laid out double dummy, clubs were splitting and wtf are we even talking about this for?

You really need to move past your biases of what is reasonable, it is self evidently reasonable for me to bid 3nt here because I did and felt it reasonable. I don't need to justify it to you, because that isn't the point, the point is that 2♥ bid is described inaccurately, please fix it. Also, please fix bidding quantitative 4nt too on 2 points, really annoying.

[cloa513]

dwar0123, on 2012-April-09, 19:30, said:

Ok, maybe I should have looked at the point range I was showing before accepting the quantitative invite, even though the displayed point range was flatly forbidden by the auction it would have at least caused me to not accept.

However, I believe the robots would be less frustrating to play with if they didn't make such invites with such hands.
Perhaps either have the robot actually have the 5-6 points it shows instead of the 2 points it had or lower the range of the 3nt call to something that is actually physically possible.

---

This was an advanced bot.

GIB is apparently capable of psyching.
I saw this somewhere about GIB's poor play.
GIBs playing against humans
1H (GIB North) ,-,1S,1NT,-,2C,-,3NT.
and the 1H was a psyche
GIB misplayed because the 1NT opening was actual 18-19 rather than its expected 15-17.

[pigpenz]

say you were playing in a tourney and your partner bid 2♥
and you didnt have a mouse to hover over your partner to find
out what his bid meant, what would you expect him to have when you kow
your right hand opp had 15-17HPC and you have 20HCP?

so we have 35-37HCP already accounted for, partner is either 3433 or has
5-6 ♥'s on a good day, thats about the best we can hope for.

I have had to learn the hard way to check before I bid and check what GIBs bids
mean, even though I dont think that is appropriate....because it doesnt happen
at a real bridge table.

were lucky that we can do this, cause the results would be even more bizarre
if we couldnt look before we leap.

[dwar0123]

This is the kind of post I am talking about, it adds nothing to the topic of gib bidding and is just further reinforcing the notion that I suck. I am getting tired of it, how many times do I have to agree that I suck before people get tired of beating a dead horse that was NEVER relevant.

No wonder there is so much tension in this forum, seriously guys, why is everyone so intent on putting others down?

pigpenz, on 2012-April-10, 16:27, said:

say you were playing in a tourney and your partner bid 2♥
and you didnt have a mouse to hover over your partner to find
out what his bid meant, what would you expect him to have when you kow
your right hand opp had 15-17HPC and you have 20HCP?

so we have 35-37HCP already accounted for, partner is either 3433 or has
5-6 ♥'s on a good day, thats about the best we can hope for.

If I had 15 points, partner is still limited to 4. If I had 3 points partner is still limited to 4. My bid is immaterial, my logic is immaterial. There is no way that 11- hcp and twice rebiddable ♥ is a useful or accurate description of that bid. It doesn't matter what kind of other evidence I have to be ignoring to proceed to 3nt, clearly I ignored it, that does not excuse that description. I am not suggesting that bid description should be informed by my 20 points, clearly that would be a horrible idea, but if we are playing with bid descriptions then they should be accurate and they should be telling newbies like me what experts like you already know. If it is limited to 4 points then it should say that.

Quote

I have had to learn the hard way to check before I bid and check what GIBs bids
mean, even though I dont think that is appropriate....because it doesnt happen
at a real bridge table.

were lucky that we can do this, cause the results would be even more bizarre
if we couldnt look before we leap.

You have issues with full disclosure bidding, that's great, I can appreciate the arguments on both sides, however this isn't the thread to have this discussion. Clearly these bids were made under the auspice of full disclosure bidding and so long as that is what we are doing, we should strive to do it well.

If your point is to instead defend known poor/vague bid descriptions as a way to articulate differences between good players and bad players, that would be relevant. However I hope that is not what you are suggesting.

[daveharty]

I am curious what rules GIB applies when the auction takes an "impossible" turn. Can GIB allow for a psyche by opener? Does it "trust" partner's bidding over that of the opponents if a conflict arises? If so, then GIB's bidding here is barely credible, if it thinks the partnership could have 32 HCP between them (leaving aside the issue of the bid descriptions; I have no idea how closely they are actually aligned with the decisions GIB makes). Bad, sure, but if the relevant algorithm goes something like "Partner has 25-30 points and I have a card for him" then I guess it makes a kind of warped sense. I'm also curious what "twice rebiddable" means to GIB; I guess I can imagine a scenario where I would bid a suit like 98742 twice (although that's not trivial), but not one where I would bid it three times.

[Bbradley62]

daveharty, on 2012-April-11, 11:31, said:

I'm also curious what "twice rebiddable" means to GIB...

GIB's definitions of biddable, rebiddable, etc are explained in post #10 of this thread: http://www.bridgebas...-played-on-bbo/

[barmar]

GIB's requirements for the 2♥ bid are either twice rebiddable ♥ (any strength) OR 5+ ♥ and <4 HCP. But we can't express OR in the specification language. So the author of the rules chose to display the first possibility rather than the second. But maybe we should change it to just say 5+ ♥ -- the second option is probably more likely. We don't want to put 0-3 HCP in the specification, though -- stronger hands are still possible because of the first option.

BBradley is correct that in this auction, the 11- HCP is due to North being a passed hand; if East were dealer, the description would just have said "twice rebiddable H".

Regarding psyching, I think GIB always believes partner. It does have the notion of partner making a bid that contradicts something it said earlier, but I don't think it does anything with it in practice (when debugging, it displays a special character to indicate the conflict). And bidding rules can make use of what the opponents have shown (e.g. that's how it distinguishes a cue bid from a natural bid). But it doesn't have any general mechanism that tries to reconcile all the information from both sides -- there's nothing that automatically kicks in and concludes "Partner can't have that many points, because it would require a 50-point deck."

[Bbradley62]

barmar, on 2012-April-11, 20:34, said:

GIB's requirements for the 2♥ bid are either twice rebiddable ♥ (any strength) OR 5+ ♥ and <4 HCP. But we can't express OR in the specification language. So the author of the rules chose to display the first possibility rather than the second. But maybe we should change it to just say 5+ ♥ -- the second option is probably more likely. We don't want to put 0-3 HCP in the specification, though -- stronger hands are still possible because of the first option.

I thought it was GIB's usual policy to provide the least common denominator for bids, which (as you suggest) would be simply 5+H. I'm surprised this one is different.