calm01

Doubling even a strong 1NT for penalties is a fairly common occurrence. So i was somewhat surprised to find what my 2S bid meant to GIB was not showing any bridge logic. My left hand Robot opponent started the bidding: !NT double pass 2S all pass. My Robot partner penalty double apparently promised 16+ HCP. I wanted to play in 2S but my 2S bid apparently meant: 11- HCP; twice rebiddable S' 12- total points Surely there are two problems here: I am unlikely to hold 10 HCP and if I did hold 9 or 10 points I would not be making a signoff in 2S. The descriptions of the meaning of the Robot double and the my 2S bid are not mutually consistent. Which one or both need correction for this reasonably common situation?

GIB already has the key najor suit support-with-support tool, namely Jacoby 2NT. So why not support with support? Bidding 1S in response holding a known 10 card heart fit is asking for trouble. I would not like to be in guess mode when the opponent bid 4 or 5 clubs without knowledge of the massive heart fit opposite. Please update GIB to use its support-with-support tools. Just raise hearts or employ Jacoby or splinter as standard technique when a 9 card or bigger fit is known.

junyi_zhu, What you say about the lack of few a basic bridge guidelines in GIB and its tendency not to no double for penalty very often, its unwillingness to bid NT to limit its hand and its poor slam bidding jugdment etcetera all resonates with my experience of GIB as a bidding partner or bidding opponent. This all makes very grim reading. I could add to your comments on GIB mis-bidding two suiters to its frequent mis-responding to a partner (GIB or human) showing a two-suiter. What seems to be missing is being able to communicate or recogne distributional information. GIB also does not always support with support (especially for hearts) losing much of the advantage of a 5 card major system. This sounds like a recipe for abandoning all hope. However, from what Ari has written in many recent posts, it seems that correcting a few of the more commonly occurring bidding descriptions and ensuring GIB understanding corresponds to its descriptions will make a very significant improvement in the GIB bidding experienced by human partners. Also its database or similar situations may help to improve its bidding over time. After all it is not a bad declarer. And adding some signalling in defence will help a lot. So there is much to go for. In summary it seemed to me like a "wouldn't start from here" situation but Asi has gven me hope for a better future. What do you feel?

Bbradley62, I agree about the lack of self-consistency inherent in the GIB's options. The approach you suggest makes bridge sense to me - and is one I actually employ with one particular partner - it has worked reasonably well for over a decade. Ari, Thanks for your prompt analysis and reply. I accept that GIB is selecting from a poor range of options in this situation. So the problem is not of GIB programming, but the options it has been provided with have little inherent bridge logic. Based on playing with over 20 partners and six completely different bidding systems, there are probably only three broad but practical approaches to responding to strong bid opening after having given a first negative/relay depending on support/strength combination: First let us assume that 2D is a first negative based on a point count/control agreement then a second response to a suit rebid by the strong hand probably needs to be based on something like one of three broad styles: - when less than say any one King or the Queen of partners suit, support with 2 good/any 3 cards or if less support give second negative, - with say at least any one King or the Queen of partners suit, a) support or bid own reasonable 4/5 card suit (4/5 and definition of reasonable according to style), or B) support or show stops/features for 3NT when partners suit is a minor, or c) support or give second point count/control answer. The choice of a,b,c is a matter of partnership style and partnership perception of frequency/utility. If the first 2D response is just a relay then the strength requirements for an own suit bid are obviously higher but the bridge logic is basically the same. I appreciate that bridge logic may be without the programmers scope but if GIB is to improve, both the bridge logic needs to be sound (self-consistent and take account of frequencies) and the programming needs follow the options. Here the problem appears to be in the bridge logic not in the programming. For convenience I repeat your descriptions of the options available to GIB after 2C 2D 3C. "it is not always easy to choose a bid with a 100% accurate description. For example, in your auction 2♣-P-2♦-P;3♣-P-? here are the descriptions of several possible bids: 3♦ Cheaper minor -- 4- total points 3♥ 4+ H; 5+ total points; forcing to 3N 3♠ 4+ S; 5+ total points; forcing 3N 5- H; 5- S; 7-12 HCP 4♣ 3+ C; 3-5 total points" Assuming you agree there is little bridge logic in these options, what is the best way forward for all four of 2C 2D then 2H/2S/3C/3D by the strong hand. Will a review of the bridge logic behind all the GIB bid descriptions pay equally high dividends in terms of effort and improvement is GIB bidding?

GIB held Q63 105 Q8532 1054 The bidding began pass 2C pass 2D (by GIB) pass 3C pass 3S (by GIB) ... The description of the 3S bid made by my GIB partner was: 4+ S; 5+ total points; forcing Obviously GIB' understanding of the 3S bid differs markedly from the description of the bid. I repeat my request for the descriptions of GIB bids to corrrespond with GIB's undestanding. Until then the GIB bidding will be flaky at best. It matters less if the descriptions make some or little bridge sense - GIB must act in a manner consistent with its bid descriptions or a reputation of poor bidding will remain and GIB bidding will continue to underperform its declarer play. what do you feel about the issue of GIB bid descriptions not being respected?

We currently just need to accept that at the current status, GIB is a very poor partner when defending and just laugh it off. I do. Hope you can see the funny side of the following. GIB often makes random signals and seems to ignore high for encouraging, low for encouraging. high-low for a even number wanting a ruff, reverse length signals for ruffing ... It is not just the lack of much signalling or signalling capacity or awareness be it attitude or count, GIB makes random leads too. Even the Rueful Rabbit (RR) would not find a lead of x from AQx in a side suit against 5Hx to allow me to make an impossible contract. Also RR would not lead a sude suit singleton holding AKQxxxx hearts with the opponents ended in 4D in an auction suggesting I had zero points (the actual case) let alone the Ace of the singleton suit. To make you fall about laughing. my charmingly erratic robot partner did not even have a trump to ruff with! We need, perhaps, to accept that GIB cannot defend very well at the moment but will get better over time. I believe correcting GIB bidding must come first else we will be defending more often than we should when partnering a charmingly erratic bridge computer profram.

Arigreen - your response to "Passing of a Forcing Bid" included: "The underlying problem was that 3S did not promise club support, so GIB didn't know that it could fall back by rebidding 4C. When I updated the definition of 3S to promise 4-card club support, the problem went away. Now GIB chooses among 3N and 4C, picking 3N more often than not." Based on this most helpful reply, it appears the robots seem to mostly honour the coding that lies behind - and corresponds to - the description of each bid made by its partner. So if the descriptions are not appropriate for the situation, but the corresponding encoding is consistent with the descriptions, it is not the robot making poor bidding judgements but that the the robot is following poor encoding that corresponds to poor descriptions. This is another of many similar problems but occurs more often than the issue raised in "Passing of a Forcing Bid". Here is one of many similar examples I could give: 1C pass 1D 1S X pass pass pass The robot almost always seems to pass for penalty. The double is described as "3+C; 2- D 17-21 HCP biddable S ...". If the double means this, the robots pass is not unreasonable. But the description is possibly not reasonable bridge. When opening one of a minor when playing a 5-card major system, holding 4 in major is one of the most frequent situations. Holding 4 of the opponents suit one can - if one wants to - pass, rebid 1NT or 2NT and double with the other major. Trap passing at the one level is rarely useful. So a more useful meaning for a double in this sort of situation is 4 in other major and better than sub-minimum hand (12+) and either tolerance (2-3) of your partners suit or if not holding tolerance, say 15+ HCP as then we hold the clear balance of points. A more credible rebid for the actual description given would not be double but be 2NT or 3NT or if unbalanced pass, support partner or rebid clubs or reverse as appropriate. Fixing the inappropriate description and the encoding that goes with it would stop the robot passing these common and necessary other 4-card major bids for penalties. This leaves me with a nice 'warm feeling' about the GIB bidding potential. If the descriptions make bridge sense and the encoding corresponds to the descriptions, the robots could bejave as a good club player or better. May i suggest that a comprehensive and independent review is undertaken of: - the bridge sense of the bidding descriptions, - the correspondence of encoding to the descriptions, - development of measures of the consequent improvements in robot bidding. One such measure would be the improving ratio of robot-robot scores in IMPS compared to robit-human scores in IMPS when three robots are at a table.

Could a simple check reduce the size of these occasional bidding failures? For example: When a combined fit is available (at least 8 diamonds are implied) and over 28 points are held (here at lest 29 are implied) the robot must bid to game in NT or a suit. This would at least get the partnership to game when a slam is on and avoid the programming embarrassment of ending in a part score in a slam hand.

Playing with 3 robots the following losing behaviour is frequently seen by a defending robot. It is quite amusing to see, as it is a defensive error often indulged in by human players. The trump suit is spades. You are West with a spade holding of K105 with AQJ8 of spades in dummy (North), A spade is led by declarer through you towards dummy. What do you do? It is rarely right to play the King! But people do and the robot often does also! If the suit is trumps, declarer might intend to play the Ace to draw a second round of trumps to prevent a defensive ruff. Or if the suit is not trumps declarer may have a singleton and cannot afford a loser in the suit. Or declarer may be desperate to get a guaranteed entry to dummy for a more critical other suit finesse or suit development or avoidance play. Also declarer may have a singleton and intend a ruffing finesse on the way back. Another reason to withhold the King is to make declarer waste an entry to hand to take a second finesse. So it normally demonstrates a death wish to play the King as a defender in this situation. I have dubbed this strategy of playing the King "Group Captain Syndrome". In the Second World War the attrition rate for fighter pilots was so high, many preferred the certainty of death by flying into a mountain side to the uncertainty but high probability of death by continuing missions. A Group Captain was a senior commissioned officer in the Royal Air Force. Is it about time the robots defence was demoted to that of a more junior and more sane rank?

Thanks for your prompt response. How about a response to the problem under the heading "bidding failure?"

The robot bidding makes an extraordinary turn by passing a bid described in its own desriptions as forcing. Playing with 3 robots, the bidding was: 1Club 2Spade 3Spade all pass. A robot partner opened a Club with 9542 K10 2 AQ2 A83 Following the 2spade jump over-call, I responded with a bid of 3spades holding: K AJ984 K9 KQJ54 Checking its meaning before making the bid (subsequently confirmed by using the hovering of the mouse over the 3spades bid), I was somewhat surprised to find a forcing bid passed. Should the bidding be consistent with the word "forcing" in the descriptions ir the word "forcing" be removed from the descriptions? Regards, Ric (calm01)

I enjoy playing against 3 robots - the service works well and the erratic behaviour of the robots gives them a sort of hunan rouch. However sometimes their extreme behaviour in the bidding exceeds the capabilites of the Rueful Rabbit. The bidding was: pass pass pass 2NT pass pass pass. I bid 2NT with AK95, A86, K7, KQJ10. My robot partner passed with an 11 point hand including 2 aces and a good 6 card suit. The robot hand was 1064, Q, AJ10942, A93. Taking the kind view, this smells of bidding on the basis of an analysis from a portfolio of similar hands and the corresponding historic results. Such an approach is doomed to failure as bidding on what partner has promised and known combined assets is probably the only sensible way forward unless millions of hands are available for analysis on each and every distinct bidding sequence and vulnerability. Perhaps it is time to review the robot bidding appriach and start again with a more sensible crew/approach. What is your view of the best way forward? Regards, Ric (userod calm01)