[Retros] FW: distance-PG-

[Yoav Ben-Zvi]

To get something new without greater complexity I suggest to exchange the
mate condition with e.p. capture: 

Shortest sound SPG with e.p. capture that has a)Maximal, b)Minimal distance 

My intended solutions are 4.5 moves with distance 12+6*SQRT2=20.485
(/9=2.276) for a) and 8+4*SQRT2=13.657 (/9=1.517) for b).

It should be possible to verify (or cook) with the files generated by
François.

 

An open-ended version:

Sound SPG with e.p. capture that has a)Maximal b)Minimal average distance
per move. My preliminary candidate for a) has 5.0 moves with average
distance of 2.755 and for b) my candidate has 5.0 moves with average of
1.124  

 

Yoav

 

From: Retros [ <mailto:retros-bounces at janko.at>
mailto:retros-bounces at janko.at] On Behalf Of Francois Labelle
Sent: Thursday, September 08, 2016 12:20 AM
To:  <mailto:retros at janko.at> retros at janko.at
Subject: Re: [Retros] distance-PG-

 

roberto osorio wrote:



----------------------------------------------------------------------------
------------------------------------

"find a sound PG performing a  distance of SQR1125 ending in checkmate"

----------------------------------------------------------------------------
------------------------------------


sqrt(1125)/sqrt(5) = 15, so we're looking for 15 single knight moves.

I found 31 such sound PGs with my computer, for example:

r1b1kb1r/ppN1pppp/7n/8/8/8/PPP1PPP1/RNBQK2R in 7.5 moves
1. Nf3 Nf6 2. Ne5 Ne4 3. Nxd7 Nxd2 4. Nxb8 Nxf1 5. Nc6 Nxh2 6. Nxd8 Ng4 7.
Ne6 Nh6 8. Nxc7#

Amusingly I found exactly 1 mate PG in 7.5 moves with one knights-only
solution, and one which isn't:

rnb1kb1r/p1N1pppp/8/8/8/8/PPP2PPP/RN1QKBnR in 7.5 moves:
1. Nf3 Nf6 2. Ne5 Ne4 3. Nxd7 Nxd2 4. Nc5 Nb3 5. Nxb7 Nxc1 6. Nxd8 Nxe2 7.
Ne6 Ng1 8. Nxc7#
1. d4 b6 2. Bh6 Nxh6 3. Nf3 Nf5 4. Ne5 Nxd4 5. Nxd7 Nxe2 6. Nxb6 Qd5 7. Nxd5
Ng1 8. Nxc7#

So if we're looking for a composing challenge with a unique "solution", then
this can be one.



Nicolas Dupont wrote:

Thanks François. The shortest I'm able to produce yet is NC#(38,0,0) in 9.0:


1. Pa2-a4   Ph7-h5     2. Pa4-a5   Th8-h6     3. Pa5-a6   Th6-f6 
4. Ta1-a5   Tf6-f5     5. Ta5-e5   Pd7-d5     6. Ph2-h4   Dd8-d6 
7. Th1-h3   Dd6-g6     8. Th3-e3   Pf7-f6     9. Pf2-f3   Dg6-g3# 

Maybe there are not too many capture-free mate PG (sound!), so that you
might compute all of them till such a bound? 


I don't have the numbers but I suspect that the fraction of games that are
capture-free is still too large to provide a significant speed-up. At least
not enough to compensate the exponential increase from 6.0 to 9.0 moves.

    François

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist1.pair.net/pipermail/retros/attachments/20160909/798e5359/attachment.html>