# Computer Chess Club Archives

## Messages

### Subject: Re: Counting & Encoding Any Chess Position in 157 bits

Author: Les Fernandez

Date: 07:29:37 11/12/99

Go up one level in this thread

```On November 11, 1999 at 17:44:22, KarinsDad wrote:

>On November 11, 1999 at 03:47:19, blass uri wrote:
>
>[snip]
>
>>>Hello Ratko,
>>>
>>>Listen I have been following this very interesting thread since it started and
>>>am involved in another thread which has also generated some interest.  Appears
>>>to be some different ideas then what I have seen here.  One of the gentlemne is
>>>interested in max of promoted pieces.  ie I remembered seeing the thread that
>>>you can have 8 queens of one color and 4 of the other.  Can someone tell me what
>>>the max #'s are for rooks, knights and bishops.
>>
>>The max number of promoted pieces is 8 for both sides(queens,rooks,bishops or
>>knights)
>>
>>Every capture of a piece can generate 2 passed pawns(1 white and 1 black) and
>>there are 14 pieces in the initial positions except kings(7 white and 7 black)
>>so it is not a problen to capture 8 of them.
>>
>>Uri
>
>This is a bit of a oversimplification of the issue.
>
>There are multiple ways to promote. Each of which gain material for one side and
>lose for the other.
>
>For example, in a capture of a piece in a locked pawn structure, both sides can
>gain a promoted pawn as you state. However, one side gained an entire piece for
>the pawn whereas the other side gave up a piece in order to promote a pawn to a
>potentially different piece (or the same piece). Hence, one side promoted and
>the other really didn't.
>
>The worse case promotion scenario is where pawns capture pawns. With just 4 pawn
>captures of other pawns, 12 pawns can be promoted to pieces. This results in the
>addition of 12 pieces and the removal of 16 pawns as opposed to the addition of
>16 pieces and the removal of 8 pieces and 16 pawns in your example; or an
>effective equivalent of the addition of 8 pieces and the removal of 16 pawns.
>
>Note: The above assumes a schema where it takes the same number of bits to store
>a knight as it does a queen. Hence:
>
>+8 pieces - 16 pawns takes fewer additional bits than
>+12 pieces - 16 pawns
>
>
>PS. Les, still waiting for the link to that other discussion.

Hi KD listen are you familiar with the current 192 bit format being used?  If so
KD and if you should visit that other thread can you please psot how the current
scheme works.  They have shown interest in understanding the current scheme but
I wasnt familiar with it.

Thanks

Les

```