[Retros] Retro Go

[Francois Labelle]

As for proof games in Go, the longest possible SPG in Go is only 2 moves 
long, which isn't encouraging. Nonetheless, I programmed my computer to 
search for shortest A-to-B proof games. I didn't expect much, but to my 
surprise my computer found some surprisingly long problems. The most 
impressive are two 24-move long problems on a 4x4 board. To get nice 
diagrams I published the problems on Sensei's Library at 
http://senseis.xmp.net/?ShortestAToBProofGames4x4 . I managed to solve 
the first problem with a friend in maybe 30 minutes, but we gave up on 
the second one. So I think they're hard, but still humanly solvable 
(hint: use the fact that the solution is unique, but of course you 
already know that).

If you want to practice with easier problems, I also published 3x3 and 
3x4 problems with some of the solutions on 
http://senseis.xmp.net/?RetrogradeAnalysis .

Here are the maximum lengths achievable for various board sizes 
according to my computer:

1x1: 0
1x2: 1
1x3: 7
1x4: 8
1x5: 10
1x6: 11
1x7: 12
1x8: 14
1x9: 15
1x10: 16
1x11: 18
1x12: 19
1x13: 20

2x2: 7
2x3: 9
2x4: 14
2x5: 17
2x6: 20
2x7: 23

3x3: 14
3x4: 19
3x5: 24

4x4: 24

     François

On 02/12/2015 12:15 PM, Alain BROBECKER wrote:
>  Dear retro friends,
>
> I didn't thought it was possible to make retrograde analysis in Go,
> but apparently it is (problem not solved):
> http://senseis.xmp.net/?RetrogradeAnalysisProblem1
>
> Best regards,
> Alain
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist1.pair.net/pipermail/retros/attachments/20150730/9a590724/attachment.html>