No, not in my opinion.

To me, Quackle is a multimodal tool -- like a Swiss Army knife, you only
expect certain aspects of it to do a good job on certain tasks -- and if you
address the right function to the right task it does a wonderful job.  You
wouldn't write to the Swiss Army people complaining that the saw blade did a
rotten job of opening your heineken.

In this case, you used the tool correctly, Stu -- you switched to a direct
sim to get a confirmation of your intuition.  The tool worked as expected.

This is not to say the static leave tables or other heuristics can't be
improved -- but Speedy Player and Championship Player are not, by
definition, going to give the same answer as a full sim, not because they
need improvement, but because they are answering a different question.

By the way: ARREEDE#

-John

On Fri, Feb 29, 2008 at 12:55 PM, <[EMAIL PROTECTED]> wrote:

>  Everyone--
>
>         Another example when Quackle missed an obvious clear choice.  Set
> to OWL06, it generated the best choice for the opening rack ADEEERR as
> REREAD at 8g,  A quick inspection told me that the best leave from that rack
> would be ADER, so the play should involve EER.  REE was far down the list of
> choices, and I decided that ERE at 8h would be best, to avoid both a vowel
> next to a DLS and front hooks on the DWS row, and typed it in to sim.  The
> "Generate Choices" option placed it 11th, about half a point below the 10th
> choice, and a valuation some 3 points lower than REE, which I
> don'tunderstand at all.
>
>         In simulation, ERE quickly rose to the top,.so I stopped after a
> little over 500 iterations and proceeded with the game.  Does this signal
> that some adjustment of the tile values, etc., of Quackle is needed?
>
> Stu Goldman
>
>
> 
>

Reply via email to