Is GIB out of its mind?

[Yu18772]

http://tinyurl.com/8nnm5qq

 

What was 4H? With 20 opposite opening bid and support for his 6 card suit, why didnt GIB ask aces after 3♥?

http://www.bridgebase.com/forums/public/style_emoticons/default/rolleyes.gifYu

[georgi]

Seems good exceptional case, but GIB needs 21TP(having 20here despite all these controls) to gear up to 4N, as opener promised only 12 ( yet ).

 

Currently fixed explanation to be 17-20 instead of unlimited.

[Yu18772]

with a fit, even if opener is 12 the slam is probably on opposite 20.....

http://www.bridgebase.com/forums/public/style_emoticons/default/rolleyes.gifYu

[cloa513]

Seems good exceptional case, but GIB needs 21TP(having 20here despite all these controls) to gear up to 4N, as opener promised only 12 ( yet ).

 

Currently fixed explanation to be 17-20 instead of unlimited.

Except you are wrong- at the 2 double-tons add 2 more TP so its at least 22TP. 20HCP yes but isn't such a point the reason to have simulations- yet you don't use it ridiculous.

[georgi]

Except you are wrong- at the 2 double-tons add 2 more TP so its at least 22TP. 20HCP yes but isn't such a point the reason to have simulations- yet you don't use it ridiculous.

 

You can't accumulate HCP and TP in shortness.

 

Let's say you hold ♠QJ ♥KJ ♦AKQx ♣Qxxxx - 18HCP and how many TP?

 

According to your saying it's 20TP , correct?

 

 

What i mean under exceptional hand is it has 20HCP ( and still 20TP ) but too good just to follow the limit. The better continuation will be improved in one of the next updates.

[cloa513]

That's how GIB mind supposed works. You are still wrong. This given hand is ridiculous evaluation agreed but that's internal rules that GIB is supposed to follow. So you are saying GIB doesn't follow its internal evaluation so its not following its own rules. You are basically saying GIB doesn't follow any logic- no matter how bad. There are dozen of hands where the GIB evaluation has ridiculous evaluations of its hand but it still plugs through its TP evaluation (e.g. shortage in opponents suit) and using TP for NT bidding. That's GIB TP evaluation.

HCP is better described as Milton Work Index(MWI)- 4 for ace etc.

 

Indeed TP=MWI+ Distribution Points(DP).

 

Its in the system notes that DP is 1 for doubleton, 2 for singleton, 3 for void and 1 for extra cards in the long suit longer than 5.

 

There is no special aggregation.

 

GIB has a lot of logic using TP even it obviously nearly impossible to do real simulations using it. MWI is much better because there is 40 HCP in a pack.

Edited September 29, 2012 by cloa513

[Stephen Tu]

GIB undervalues long suits. If you want to devalue the honors in doubletons cancelling out the points for doubletons fine, but 6 cd suit with a 9 cd fit is worth 2 pts at least. If GIB evaluate this as 20 "total points" it has to be fixed.

[Bbradley62]

You can't accumulate HCP and TP in shortness.

Yes, you can and should, under specific conditions.

 

Even when I learned Goren's point count bidding 40+ years ago (and common methods have clearly gotten more liberal since then), singleton A, Ax and Kx were assigned both high card points and distributional points, whereas other shortnesses-with-honors were not. If Player A has a singleton A, he can win the first trick in that suit, just like Player B who holds Axx. But, Player A also can ruff the second trick, just like Player C who holds a small singleton and loses the first trick. So, Player A should reasonably count his holding as 4 HCP and 6 total points; Player B counts 4 HCP for his A and Player C counts 2 total points for his singleton.

 

While I certainly agree with Stephen that GIB should also be counting distributional points for long suits, GIB should, at the very least, count both HCP and distributional points for the three holding of A, Ax, and Kx. Baby steps...

[georgi]

I didn't say that's the perfect call.

 

I've just gave more information why GIB stopped to 4♥. And how many TP GIB finds for that hand.

 

Also i've mentioned that it will be improved.

 

I don't recall saying GIB doesn't follow its internal logic. GIB follows strictly particular rules and could deviate after simulation around other rules.

 

There might be not dozens, but thousands hands where GIB does something inaccurate, but it doesn't mean it bids randomly in general. I hardly believe there is at least one person posting in this GIB forum who never opens or overcalls with 4 cards, 1NT with 14HCP, 2NT with 18-19HCP, 2♣ with optimistic values and etc. Human makes assumption based on his knowledge and 6th sense, GIB based on rules and sometimes simulation.

 

Of course, that there are variety of counting and evaluation of any hand based on current calls, possession in hand as well.

 

The hand as example is stronger and therefore not well evaluated, but to be evaluated as such there must be information which GIB to match and consider. One hand is stronger based on shortness, another is stronger based on two fits, third based on just points and so on.

So it needs more then "Hey, bid slam here!". It needs followups, which will be added.

[Bbradley62]

HCP vs Total Points

Gib uses both old fashioned HCP (A=4,K=3,Q=2,J=1)) and “Total points” (HCP+3 for void, 2 for singleton, 1 for doubleton, but short suits containing an honor are reduced by 1 point). It sometimes uses “8421” points when cuebidding (A=8,K=4,Q=2,J=1). It will usually force to game if it thinks it has 25 Total Points between the two hands.

Its in the system notes that DP is 1 for doubleton, 2 for singleton, 3 for void and 1 for extra cards in the long suit longer than 5.

Did you make that last part up, or does it actually say this somewhere else in the System Notes?

[cloa513]

Did you make that last part up, or does it actually say this somewhere else in the System Notes?

Anyway what is really wrong is that the hand should be evaluated using 8421 points which is slam region hand evaluation- then its clear cut.