[virgosrock]

---

my fault for not bidding 4NT over 3 NT. But 14????? instead of 17-21 hcp?

vrock

[Stephen Tu]

It's absurd to require 17-21 in this spot. It's a negative double, opener with a spade stop and no heart fit is more or less forced to bid 3nt unless he simulates that passing for penalties will work out better.

So they should fix the definition of 3nt, not so much North's choice to bid it.

[virgosrock]

Stephen Tu, on 2018-May-17, 14:35, said:

It's absurd to require 17-21 in this spot. It's a negative double, opener with a spade stop and no heart fit is more or less forced to bid 3nt unless he simulates that passing for penalties will work out better.

So they should fix the definition of 3nt, not so much North's choice to bid it.

yep. no issues with 3NT. issue with blurb,

vrock

[jdonn]

Will be fixed some point I believe.

[cherdano]

I don't want to disrespect the programmers. But it's just genuinely surprising that a bridge programme that is actively being worked on in 2018 AD wouldn't have an automated check that explanations are roughly compatible with both internal rules and actual bids.

[helene_t]

Stephen Tu, on 2018-May-17, 14:35, said:

opener with a spade stop and no heart fit is more or less forced to bid 3nt unless he simulates that passing for penalties will work out better.

Even with a hearts fit, if the spade stopper is good. Sometimes South will double with less than four hearts.

[virgosrock]

helene_t, on 2018-May-24, 15:29, said:

Even with a hearts fit, if the spade stopper is good. Sometimes South will double with less than four hearts.

seems to be such a simple thing Helene. make the bid, compose the blurb based on the bid, push both out simultaneously.
But Sir Donn says undo-able.

vrock

[Stephen Tu]

The meanings of bids have to worked out ahead of time. Composing the description in real time based on the hand held would be equivalent to cheating. "Partner, I hold exactly x=x=x=x shape and N HCP and stoppers in suits: a,b,c". Bids mean whatever you want them to mean so that you always have a convenient bid that shows your hand available. Bidding would be so easy and accurate!!!
So I don't think we really want that. Descriptions have to be in the database beforehand. Some meanings are just going to be off because it's hard to construct the rules database to cover all possible auctions. But they could do a better job in getting all first round competitive actions completely covered with better definitions than now, and the first rebid after that. And the bot has to be allowed to deviate from the description because sometimes no bid really fits; humans improvise and do this all the time, just we tend to be a lot better at sussing out what really is least bad mis-description than bots tend to be.

As for Helene's comment, it depends on your agreements, whether double at this level is played as pure neg (and absolutely promising 4cd hearts in theory), meaning some good hands with < 4H will be stuck for a good call and will try something least misleading and sometimes screw your side, or if you are playing more "thrump" doubles ala Bergen where the priority is 3nt and alternately you will play some hands in 3nt when hearts was a better spot.

[virgosrock]

Stephen Tu, on 2018-May-24, 18:48, said:

The meanings of bids have to worked out ahead of time. Composing the description in real time based on the hand held would be equivalent to cheating. "Partner, I hold exactly x=x=x=x shape and N HCP and stoppers in suits: a,b,c". Bids mean whatever you want them to mean so that you always have a convenient bid that shows your hand available. Bidding would be so easy and accurate!!!
So I don't think we really want that. Descriptions have to be in the database beforehand. Some meanings are just going to be off because it's hard to construct the rules database to cover all possible auctions. But they could do a better job in getting all first round competitive actions completely covered with better definitions than now, and the first rebid after that. And the bot has to be allowed to deviate from the description because sometimes no bid really fits; humans improvise and do this all the time, just we tend to be a lot better at sussing out what really is least bad mis-description than bots tend to be.

As for Helene's comment, it depends on your agreements, whether double at this level is played as pure neg (and absolutely promising 4cd hearts in theory), meaning some good hands with < 4H will be stuck for a good call and will try something least misleading and sometimes screw your side, or if you are playing more "thrump" doubles ala Bergen where the priority is 3nt and alternately you will play some hands in 3nt when hearts was a better spot.

Right Stephen. would be cheating. Actually I meant what I have said often. If blurb is a wacky distance away from bid then re-simulate or re-whatever.

vrock

[cherdano]

Stephen, you misunderstand my point. I am not suggesting that GIB should adjust the situation on the fly. I am suggesting that the GIB team shouldn't have to wait for user reports to identify nonsensical explanations such as this one. GIB probably plays hundreds of thousands of hands every day on BBO - if the programmers had a tool to automatically hand for non-sense such as this one, they would have known about this particular wrong explanation years ago (it's not a complicated auction). And they'd hopefully found a way to fix all such blatantly wrong and common explanations systematically.

[Stephen Tu]

I was commenting on virgosrock's post not yours, cherdano. I agree they could implement some sort of checking mechanism to flag problematic auctions automatically. I kind of feel like they just don't have enough programmers to make progress at more than snail's pace, as jdonn has said they have years of backlog of bugs already. Implementing this might just open flood of too many duplicate reports to classify and sort through to be useful.
I think they have to try and systemically list out all plausible classes of first round competitive sequences and try to get everyone's first bid and opener's and intervenor's 2nd bid, basically cover 1.5 rounds of competitive bidding and make sure everything is sane. It's bad that they still have problems just getting the first response to a takeout double consistently reasonable.

[virgosrock]

Stephen Tu, on 2018-May-25, 04:22, said:

I was commenting on virgosrock's post not yours, cherdano. I agree they could implement some sort of checking mechanism to flag problematic auctions automatically. I kind of feel like they just don't have enough programmers to make progress at more than snail's pace, as jdonn has said they have years of backlog of bugs already. Implementing this might just open flood of too many duplicate reports to classify and sort through to be useful.
I think they have to try and systemically list out all plausible classes of first round competitive sequences and try to get everyone's first bid and opener's and intervenor's 2nd bid, basically cover 1.5 rounds of competitive bidding and make sure everything is sane. It's bad that they still have problems just getting the first response to a takeout double consistently reasonable.

Perhaps, ONLY for money bridge,they could have humans edit GUBBO's bids before passing them on. prevent the ridiculous things GUBBO does. Money bridge tables are not that many and to some slight extent will justify the table fees. LOL

vrock