suokko

suokko & Derric I know I haven't posted too much in a while. But I'm still lurking. We have a new biding system that will be 16 days old on Saturday. You might get a chance to laugh when bid some funny contracts.

I play inverted raises that can have 4M side suit without any strength limiting problems. There is one big issue of course. Continues are completely artificial based on relay structure that is not something that I would encourage any IA player to try. The advantage over inverted minor denying the side suit are minimal compared to cost of system accidents doesn't pay in long run. I just wanted to raise the fact it is possible to have good inverted minor raise structure with ability locate major fit and limit strength. But the cost isn't worth it for about 95% of bridge players.

True. I'm not actually claiming that singleton leads are bad in every case. I do them my self against 3NT too. But in my option frequency of GIB leading singleton is a lot higher than winning frequency would be. I also created hypothesis that GIB actually doesn't read accurate meaning of pass when generating hands for the lead simulation. I actually tough for long moment if my view might be worse than what GIB does in table. But my conclusion was that in this case GIB does worse than I do.

I just saw one of those weird looking singleton leads from gib. I started wondering what could make gib to see that lead as most likely way to set 3NT. After a short moment I have started to suspect that gib might not take into account partner's passes when simulating the hand. For example pass over 1NT limits a lot chances that singleton lead would be the setting lead because many hands with long good major would have bid. But if the simulation code doesn't check hands to be consistent with passes in auction then that would sometimes give huge probability advantage for singleton lead against 3NT. Of course I can't know if that is the reason why simulation seems to favour singleton leads a bit too much. But if that is the case then it should help many situation where gib does simulations after some passes if there was filtering code to remove hands that don't match passes in the table.

Why not bid 2♦ as opener with minimum and no extra length in majors? To me 2NT sounds like creating a game forcing situation in depend of meaning of 1♦-1♥;1NT-2♣;2♦-3♣. But even in that biding I would assume 2♥ to be only nonforcing bid by responder. NMF should promise 5+M or FG values. With that definition you can simple simple offer to stop to 2M when not having FG values. I would like to combine that with all direct bids above the responder's major as natural invitational NF bids. But I would not assume partner to understand that thinking without discussing it first. (Above 3M splinters)

Splinter, please! That helps to see where tricks are coming.

Many don't do same decission in table and when presented something as problem. It is natural to suspect that first tough coming for problem isn't the correct one and people take longer to think for the solution.

Ace asking is evil convention. It is very hard to try to teach youngesters not to use it after they have learned it from start as only slam exploration method. But anyway. LTTC is part of "standard" cue biding that I try to teach :)

The friend detection sounds good idea to me. Another group that should be allowed extra thinknig time are new players. New players often have to think a lot longer even in situations that are automatic to experienced players. If new player gets kicked out of table because of slow play it hugelly reduces likely hood that the person continues the hobby. For everyone else it is simple minor annoyance. Also going out of computer to door or something may take easily over 90 seconds. Is there some way to prevent kicking during that time? That makes me thinking kick should be vote between other players instead of automatic. Automatic feature can't coup with all cases where a kick would just cause bad feelings. If you think 90 second thinking break is completely out of question I think you aren't actually ever defending part-scores with skill. IMO option most of part-score defense require very long thinking breaks to weight between possible defenses early in hand. I can easily play in game that average 4-5 minutes per board in live bridge. That is even the ussual speed of play. But if the session has multiple hands were I need to decide between different defenses early in hand I can easily raise the average to 7 minutes.

How about keeping the warning but instead of automatic kick offer all other players option to kick the stuck player. That would prevent issues when playing with friends where I expect regulary someone to take over 90 seconds thinking break. That kind of feature of course requires some UI changes to mak a boot button appear when timer hits that 80 seconds or what ever is the boot time. If all other players click the boot button then host would be booted out of table.

+1 I was also looking if I could set or read the dealer/vuln/board number for the hand that I'm generating. When I read the documentation I understood that they won't work correctly for BBO dealer. Problem is that for standard alone dealer those settings are simple affecting output information for deal printing. Even if setting deal number would be supported that wouldn't be exactly what I would like to do. For me the most usefull feature would be to have dealer(compass)/vuln(compass) syntax support for the conditional syntaxt. That would allow deal script to modify the condition based on dealer and vulnrability information. I suppose that shouldn't be hard to add to the BBO dealer but you are very busy with other more important stuff.

Hi Fred, It would be nice if connection code would be moved to a service that allows bbo to keep connection active while multitasking in Android. Implementation needs to extend Service class and start the connection code in the service. More details can be found from Service documentation

From permissions it looks like you try to acquire wake lock manually. There is simpler and probably more robust way of acquiring wakelock trough window manager. That doesn't require permissions even so wakelock permission can be droped. http://stackoverflow.com/questions/2039555/how-to-get-an-android-wakelock-to-work

If I try to kibitz a tournament that disallows kibitzing I start to get messages from the tournanament but I stay kibitzing in MBC table. I also get all messages from MBC table that leaves messages from the tournament as only visible effect of the bug to user. At least I haven't yet noticed any other side effects. This is with flash client v47m.

Didn't fix the problem. :(

Problem is: I I reveice Japanese characters flash bbo starts to use a lot of cpu all the time while frames take many seconds to render. I noticed problem when I was playing in BBO japan tournament. Same time firefox eats a lot of memory (from ~200M to ~800M). Characters that caused the problem showed up as squares (missing from my font) instead of Japanese glyps. More details: I'm running 32bit Debian Linux with Flash 10.1.85.3. My locale is set to en_US.UTF-8. I'm now testing if upgrade to 10.1.102.65 will fix the problem.

My partner just added dealer rules (windows client). Bug triggered after change in hands 45 and 47. No idea if that is the reason.

Not even dealer. Now it happened after 40 boards without dealer limitations :/ I could monitor the problem with wireshark and see if there is difference in communication.

Doesn't matter. I have had opps passing, GIBs interfering and today even host biding. It has happened when I'm host, partner is host or kibitzer is host. My partner is using windows client. No. He is using windows client. I didn't report this when I first saw because I couldn't find pattern. Only relevant configuration that I can think of is that if table is limiting who is dealer somehow. Maybe even switcheroo might be reason. But I haven't yet found the pattern.

I'm still biding with my partner :) But problem is very random. It apears maybe once in 30 hands in average. But might happen multiple times in short time.

I noticed that Flash bbo somehow randomly doesn't so hand after biding ends in partnership biding. Same deal is not recorded in My result either so I can't see the hand without having 2nd client kibitzing my self. Is there some bug thread somewhere still? I couldn't find it.

Even if there was flash support flash is horrible slow on ARM and inteface is not the best for touch screen. You can actually run bbo in N900 but it is slow and hard to touch anything correctly in screen because of size.

minimum for acol 2 is invitational hand (5 losers). So ranges do actually overlap. for me this hand is still invitational after 1S response. But when playing without an unamed Italian convention I would force to game with 3C.

I'm at least interested what is the style of other people. I'm even open to change my personal style based on commens here. :) But for me opening bid with this hand depends a lot more on the system details and partners option.

Mikeh, in the OP alludes to this. Phil outright would open 1NT. And you, deep down, know it is right to open it 1NT. Maybe, It's a 1NT opener. I know nobody downgrades at your level, but you could lie about seeing one of the Jacks, in the postmortem. Me, I have no reputation to worry about, so it is easy to open 1NT. Looks like routine 1NT for me. An d I have reputation of upgrading a lot. :) In fact 1nt with 4333 18 HCP has produced an Ave- and 3 close to tops in pairs. I started doing that in the last fall.