[msheald]

Hello! Poor Robot bid descriptions have been discussed before. I'll add some additional examples - 3 from a recent game of 12 hands.

https://tinyurl.com/2bohtrob

3 Clubs means 4+ clubs, 4+ hearts and 11+ total points. I suppose that some folks might get 11+ total points on distribution with this hand, but after partner bids NT, that becomes less likely. Additionally, one of my early lessons was that distributional points are counted most prominently in dummy - both of which are problematic for this hand. The most important point is that bids should never mislead partner, which is the case here.

https://tinyurl.com/2xjcves7

Double means 5+ spades and 6+ total points. A double as a bid description at the 4 level is problematic, in my opinion. I don't know of many folks who play it other than for penalty. Additionally, I'm not sure I know folks who count a singleton club as 6+ distributional points.

https://tinyurl.com/29mdhdx5

Definition states opener reverse - 5+ diamonds, 4+ spades, 2-hearts, 21- HCP, 18-22 total points. Sometimes, there is no "right" bid, but, as in this case, to me, it appears as a "wrong" bid and misleads partner.

I realize that bid programming is complex and can be expensive and introduce a number of problems. I support BBO's decision not to spend resources in fixing the robots, but I would think that bid descriptions would be easier, and less costly, to fix and significantly increase the enjoyment of robot games. Best regards.

Mike

[smerriman]

msheald, on 2025-June-28, 05:43, said:

I support BBO's decision not to spend resources in fixing the robots, but I would think that bid descriptions would be easier, and less costly, to fix and significantly increase the enjoyment of robot games.

You would think wrong. This is the hardest part of modifying GIB.

[pescetom]

msheald, on 2025-June-28, 05:43, said:

I support BBO's decision not to spend resources in fixing the robots

They are giving you the decades old and buggy "Advanced" robot free (rather than continually increasing the fee) or you just feel they deserve easy profit?

[johnu]

pescetom, on 2025-June-28, 14:54, said:

They are giving you the decades old and buggy "Advanced" robot free (rather than continually increasing the fee) or you just feel they deserve easy profit?

Actually BBO gives the deliberately crippled "Basic" robot for free, and charges for the decades old and buggy "Advanced" robot.

[benellis58]

Given that the GIB robots are a key part of BBO - probably THE key part - it is not at all impressive that BBO appears to be quite comfortable with having them be so consistently and utterly inept.

[mycroft]

Free play, with any 3 players anywhere in the world, any time of the day or night, for 20+years.

I'm a massive introvert, and that's *still* incredibly more important than playing with a C player that won't learn and won't change, but plays the cards well robot.

I think anybody who's nose is out of joint about "bad robot" (except for the people who say that playing with them hurts your real game, who have a point) has never been the designated spare/non-playing director for a club game. The number of worse-than-robots (because not only will they not learn (don't attempt to try, it'll only make it worse!) and won't change (they're having enough trouble remembering what they do play!), but they also play the cards slightly worse than Mrs. Guggenheim) I've played with, and will play with again, makes playing with robots a pleasant experience (at least they aren't telling you what (they think) *you* did wrong!) for whiling away 20-30 minutes on hold with the doctor, or a half hour on the bus, or that hour in the airport, or...

I mean, we should be *glad* that we're not in the chess world where even your laptop can beat all but the absolute top players. We especially should be glad that bridge is not a perfect-information game, and bidding and play methods are so variable, and so is somewhat resistant to T-800.

Should they be better? Sure. Should work be put into it? Yeah, probably. Should BBO do it over ensuring that my team game with my locals still goes, even if half the "locals" are in another country, and one is permanently in another city, 6 times a week? Or even ensuring that (paid) tournaments can be played, whether "for fun" or under the auspices/by one of 50 NBOs, with real players (some of whom the robots would trounce, 8 games out of 10)?

Well, if you're the kind of masochist who will pay to play events knowing how much they hate their partners, okay, sure. But if you're the person who will continue to reward them with your money, time, and energy even if they don't improve it? What's in it for them then?

Is robot important? Sure. Is it THE key part of BBO? Well, once all the 70+bridge players die and no 50+ new retirees come in, then sure, I guess...

[pescetom]

johnu, on 2025-June-28, 16:01, said:

Actually BBO gives the deliberately crippled "Basic" robot for free, and charges for the decades old and buggy "Advanced" robot.

That was precisely my point. Either msheald is the only customer not being charged for some reason, or he just thinks BBO deserve money for old rope.

[msheald]

Another example of a hand that I refuse to play.

https://tinyurl.com/2c84ctlh

2NT description states Lebensohl - transfer to clubs, then passes when I do . . . . Right!

Mike

[smerriman]

That one's entirely on you.. that's not even remotely close to a double.

[msheald]

Sorry, but it is irrelevant whether you agree with a double of not, especially without running 5000 simulations in order to determine the best outcome with the bids.

The point is that the robot's bid description was grossly in error. How can a grossly incorrect robot description error be "entirely on me"? I'm happy to learn if you explain why the robot's bid description was correct. Best regards.

Mike

[msheald]

Thank you for your note. I appreciate it.

We seem to be be discussing two different points. For main point, that the robot bid description was significantly in error, it does not matter what my bid was or whether other people might agree or disagree with it. I realize that my understanding of the bid description might have been in error, and., if so, then I would appreciate the chance to learn from my mistake for future games. If the bid description was in error, then I would appreciate folks's assistance in trying to discern such in the future so the my bids with a robot partner take that into account. In either case, my enjoyment of the game would be increased.

With regards to the correctness of my double, I don't have a bridge simulator in order to run 5000 hands in order to tell the average outcomes for both MP and IMPs. If you have such, I would appreciate the simulation feedback. My experience, at my intermediate level at our club, is such a bid is OK - taking into account vulnerability and the level of competition. For the 5000 simulation runs, though, I presume that the Lebensohl response would not be in one of the simulations? The 5000 simulations would be able to take into account the frequency of a flat hand, such as robot has, as well as those card distributions in which there is an excellent fit. Would you be able to run such a simulation? Only through that would we be able to tell the average outcomes of a double and judge your hypothesis that my hand was "not even remotely close to a double". Best regards.

Mike

[smerriman]

Grossly wrong description?! I don't see anything wrong in the description.

Playing lebensohl means that its direct bids of 3 of a new suit show extra values. It does not have extra values.

All weaker hands bid 2nt lebensohl forcing you to bid 3♣, exactly as the description says. It then either passes, or signs off in 3♦ or 3♥, depending on what its longest suit is. It happens to hold the worst possible shape where its longest suit is 3 cards, but it has to choose one..

Perhaps you could argue that letting the opponents play (and maybe make) 2♠x is better than anywhere it runs, but your double is the only clear error here.

[msheald]

Thank you for your note. Respectfully, I'm afraid that I disagree with both your points, but thank you for letting me consider your position.


Here is the Lebensohl response. "With 0-7 points 2NT is bid forcing a relay of 3♣. This is either passed or corrected to another suit.
With 8-11 points suits are bid at the 3 level." This did not fit robot's hand.

Again, the only way to judge your hypothesis is to run 5000 simulations. Would you be able to do that? Thank you and Best regards.

Mike

[mycroft]

So, where did you get that explanation?(*) It sort of fits my agreement, *but my partnerships are conservative against preempts* (and certainly would not consider an effective 10 count with bad shape and wasted spades anything but a pass of 2♠).

Instead, did you look at what your partner thinks the bid means? "Lebensohl -- forces 3♣ by partner - 9- total points". Note, not 0-7, 0-9 (counting shape). And what does a Christmas hand(**) with no suit and four dead spades have? I'd say "less than 9 total points", myself.

And why the difference? Because many people play that a double of 2♠, especially in balancing seat, is basically "opening values" rather than my conservative "count on 15 or so". So they have to say that "values" is bigger than I do (or is appropriate after a 15-17 NT, if (*) is correct).

You have put partner to the guess with this horrible hand. Partner made their best guess - best because "if it's doubled, we might be able to scramble to our best fit doubled. But we're not doubled yet." Partner was wrong this time - and would have been wrong with a more normal double, like 7 AQT8 AT95 QJ73 or 7 AQT8 AT954 QJ7. But would have been spectacularly right opposite 7 AQT8 QT7 AJ954.

If you choose not to play these hands against robots, I don't blame you. They're awful, and you're going to get a bad score. And since nobody is going to be affected if you abandon the hand, I don't care.

But the robot has their bid, according to the system they play - and that you are required to play if you choose to play with the robot and want them to have a clue what's going on. I won't blame them for trying; I definitely won't blame them for not working out you have a 1♥ overcall that you couldn't make over 2♠.

(*)My guess is that it's a description of "Lebensohl after interference of our strong NT", not "Lebensohl after double of opponents' weak 2". If that's the case, then note the caveat: "I know partner's got 16ish Walrus points". If you want that to apply to this case, then your doubles should show "15 points in support of any non-spade suit, or a huge hand that can override partner's warning".
(**)For those who haven't heard the joke before: okay, it's more an Epiphany hand, but still; "We Three Kings".

[mycroft]

In re: simulation, I can run one, easy. I can find out what the par result is for you vs what you'll score in 2♠ by passing.

But *even if action is better on average than defending*, I can't tell you if double is right, because building the logic of partner's direct pass and response to the balancing double, never mind third-hand's second round responses, is way more than I'm willing to do when it's clear that even if action is right, partner will never be able to guess correctly after a double (even with a better hand-for-me than this one).

[smerriman]

msheald, on 2025-July-27, 09:05, said:

Here is the Lebensohl response. "With 0-7 points 2NT is bid forcing a relay of 3♣. This is either passed or corrected to another suit.
With 8-11 points suits are bid at the 3 level." This did not fit robot's hand.

All of your previous examples in this thread (and others) have been the robot holding a hand which did not match the alerted description. I therefore assumed by 'grossly incorrect description' you meant the same thing here. Under that assumption, you were completely incorrect; the robot alerted that it had 9 or less total points, and was making a bid which asks you to respond 3♣, which is exactly what it had. This is a fact, not something you can disagree with.

It sounds like instead you meant something completely different:

a) that you believe the best way of playing 2NT is that it should be alerted as 7 or less total points, not 9 or less total points (this doesn't seem 'grossly' different, but OK)
b) as a result, you believe someone that played the convention your way should make a positive response with this hand. What response are you saying you wanted made - 3♣, showing a good hand with clubs? Or just pass, knowing it might make but being the least evil? Though then I don't know why you'd be focusing on the description of 2NT.

I don't see how a positive 3♣ would improve your outcome at all, so I may still be misunderstanding; if these are not your points, perhaps you could be a bit more specific about what you're trying to say.

As for the double, the standard definition of a double - which robots agree with - is that it shows shortage in spades, and sufficient length in the other three suits, hoping partner had a trap pass but willing to deal with playing in partner's longest suit otherwise (initially; with some hands you may overrule this later). If double is the correct bid for you, this means this is a grossly incorrect definition of the double - you have length in spades, shortness in diamonds, and would far prefer hearts to a minor. What is your definition of the double?

Running a large simulation will not help answer these questions in any way, because it is extremely complex how all followups to all bid sequences will work. The simplest way is to create a bidding poll; if 100% of players disagree with you including a lot of experts, it doesn't prove that you are wrong, but strongly suggests it.

I've assumed MPs (don't think the scoring will matter) and posted one for your action here. Once I know what you thought was wrong on GIB's side, I can create a separate one for that.