[blml] Positronic brain [SEC=UNOFFICIAL]

Tim West-Meads twm at cix.co.uk
Fri Oct 6 14:43:00 CEST 2006


Richard wrote:
> In my opinion, given that there is a logical
> alternative to 7H of Pass, it does not matter that
> both the AI and the UI seem(1) to promise identical
> values.

Parsing UI is a two-step process.  The first step is do identify the 
content value of the extraneous action, ie. "What do I know about 
partner's hand from the EA that I do not know from the legal auction?" 
and then "What action does that information suggest over another?"

If partner alerts when you expect him to alert or thinks at point in the 
auction when you expect him to have to think (e.g. you've just made an 
undiscussed call and you know he'll need time to work out the meaning) 
the content value is zero.  The content is also zero if the AI subsumes 
the UI.

> The UI still demonstrably suggests that 7H will be
> more successful than the other logical alternative
> of Pass, so under Law 16 it is irrelevant that the
> AI also demonstrably suggests that 7H will be more
> successful than Pass.

That doesn't matter if the AI and UI *promise* identical values.
It might matter if it's a case of probability.  e.g. The AI makes it 70% 
likely that pard has the H queen, the UI also makes it 70% likely that 
pard has the HQ *but based on different logic*  thus the combined 
"identical" AI/UI make it 90% likely that pard has the HQ.

Tim




More information about the blml mailing list