I know 0 is mathematically even but

[thepossum]

I know I was shouted down and attacked last time I mentioned an issue like this so I refer the readers to other threads for the unfounded abuse I am likely to get here

 

However it is ridiculous for anyone (including GiB) to bid 5NT (even number of keycards and void) when you (or GiB) have zero keycards.

 

The whole point of Blackwood is to avoid unmakeable slams

 

Excuse me for expecting the occasional bit of common sense bidding

 

Fortunately this time it occurred on Hand 1 of a Free MP tournament so it didnt waste my precious time and dollars, unlike other occasions

 

PS the modification to the code would be quite easy

 

In psuedo code its something like

 

If (currentbidder=ME) then

If ( partner.previousbid()=="4NT" and

Not(isQuantitative(partner.previousbid())) and

me.myhand.numberofkeycards()==0)

then

 

case (ME.biddingsystem().blackwoodtype)

1430: ME.mycandidatebids={"5D"}

else: ME.mycandidatebids={"5C"}

end case

 

ME.bypass_simulation_algorithm/thinking_to_save_processor_time/brain_and_partners_good_mood_time_and_money()

ME.bid(select best candidate bid from(ME.mycandidatebids))

return

 

else

ME.run_all_the_rest_of_your_code_or_bidding_system()

return

 

end if

 

else

ME.waitformyturntobid()

end if

[hrothgar]

I suspect that

 

1. ME.bid(select best candidate bid from(ME.mycandidatebids)) is significantly more difficult than you might think

2. There will be some significant losses when that 5♦ response that you think promises an Ace actually has a void

 

Please note: I think that having GIB always bid 5!D with zero keycards and a void is a much more reasonable ask

 

Also, I can't help but note your repeated complaints when GIB shows the Queen of trump in keycard sequences when, in fact, it actually holds extra trump length. (The standard definition)

 

What you asking for here is for BBO to change the meaning of keycard away from standard and towards and idiosyncratic implementation that you prefer.

 

this feels inconsistent

[etha]

use jack you can switch off extra length= q trumps.

[pescetom]

There's quite a bit of of detail to void showing replies that is not fully standard and should be agreed between partners.

Some do not show with zero keycards, some change the responses when there is a known void (5NT = 1 keycard, etc.), some use 6 trumps to show a void in a higher suit, and so on.

The important thing as always is that both partners are singing off the same sheet, which in the case of GIB simply means knowing what it will do (in this case, show an undisclosed void with any even number of keycards).

[HardVector]

It's a nice try to improve the bots, but there is one important aspect to showing voids in a Blackwood situation...judgement. You show the void if you feel that it is a USEFUL void. If you don't feel it is useful, you don't show it and just respond keycards. Having 0 keycards, unless you have been able to show a bust hand already, you may deem the negative bid to be more informative than the void. I'm not sure this is something you can teach the bot.

[pescetom]

It's a nice try to improve the bots, but there is one important aspect to showing voids in a Blackwood situation...judgement. You show the void if you feel that it is a USEFUL void. If you don't feel it is useful, you don't show it and just respond keycards. Having 0 keycards, unless you have been able to show a bust hand already, you may deem the negative bid to be more informative than the void. I'm not sure this is something you can teach the bot.

 

Judgement but also meta-rules, which are effectively agreements about judgement - for instance, if I already showed my void as an ambiguous singleton/void splinter, do I still show it as a void, and if so is it considered a known void?

[nullve]

It's a nice try to improve the bots, but there is one important aspect to showing voids in a Blackwood situation...judgement. You show the void if you feel that it is a USEFUL void. If you don't feel it is useful, you don't show it and just respond keycards. Having 0 keycards, unless you have been able to show a bust hand already, you may deem the negative bid to be more informative than the void. I'm not sure this is something you can teach the bot.

The robot could just sim whether the void is useful before answering.

[DozyDom]

I agree in general, yeah. Forcing to slam when you have no keycards, not really a great shout as a rule. Opposite a particularly strog hand, or where you've already limited your hand extensively, sure, but it's often enough a sad mistake that I feel like GIB should avoid it more - I can't comment on the code necessary though.

 

Note though that you haven't received any abuse about this problem, even if some people disagree as to the solution.

[steve2005]

There may be times when it is correct to show a void with no keycards.

However, Gib is not programmed such that it can make such a decision correctly (even with sims)

Better for Gib to never to show a void with 0 KC.