View previous topic :: View next topic |
Author |
Message |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Sat Mar 01, 2008 2:42 pm Post subject: problem 2823 was a mate in 4 not mate in 3 |
|
|
I subjected the position to chess computer analysis and it is a mate in 4 and not 3 as you stated. |
|
Back to top |
|
 |
mrmip Administrator

Joined: 05 Nov 2005 Posts: 230 Location: Finland
|
Posted: Sun Mar 02, 2008 1:54 am Post subject: |
|
|
Ummph - I do not to know, what to say? I still think it is mate in 3 - computer or no computer.
You know there are three kind of computers - those, which can count and those which cannot. Unless of course it is made by IBM - "The Four letters You can count on!".
Still I would like to see the four move solution the computer gave? _________________ There are only three kinds of chessplayers - those who can count and those who cannot.... |
|
Back to top |
|
 |
papadoble Master
Joined: 15 Jan 2006 Posts: 85 Location: Yerres, France
|
Posted: Sun Mar 02, 2008 4:42 am Post subject: |
|
|
In my experience a standard chess software such as Fritz - at least the 8 version running on chessbase database, is very bad for problems. If it sees a #4, it stops here and does not search for shorter mate. I have also noticed that it does not foresee defences based on underpromotion.
There is also the special case of e.p. key, which I know mrmip is fond of, this computer softwares definitely do not see. _________________ "Si tu vois un bon coup, ne le joue pas!!" - Philippe Bichu (after Em. Lasker) |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Sun Mar 02, 2008 1:50 pm Post subject: |
|
|
this computer specifically searched for mate in 3 but could not find it. I then told it to search for mate in 4 and it found it. It commences with c2c3.
Your solution had black exchanging rooks which is clearly not forced. |
|
Back to top |
|
 |
mrmip Administrator

Joined: 05 Nov 2005 Posts: 230 Location: Finland
|
Posted: Mon Mar 03, 2008 12:36 am Post subject: |
|
|
Thanks for the explanation. Clearly the computer does not search through all alternatives. Its suggestion c2-c3 indeed mates in 4 moves, but the given solution mates in 3 as reguired.
If you look to the 'Navigate solution' you will see that the alternative of exchanging the rook is also covered and results mate in 3 as well.
I suggest you put the first move and any possible black defence move into the computer and look for mate in 2. Surely it can find that ? _________________ There are only three kinds of chessplayers - those who can count and those who cannot.... |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Mon Mar 03, 2008 1:50 pm Post subject: |
|
|
As I stated before, your 3 move mate solution is contingent on black exchanging rooks. If black decides not to exchange rooks there is no mate in 3. |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Mon Mar 03, 2008 4:40 pm Post subject: |
|
|
I see th a2a3 move if black rook goes elsewhere. I will put position in th computer again at different stages and see what if finds. |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Tue Mar 04, 2008 1:10 pm Post subject: |
|
|
I resetted the computer to its default algorithm (it has multiple search algorithms) which is the strongest, and it found the mate in 3 in a second. So it was different algorithm that prevented it fidning mate in 3 but it found mate in 4. |
|
Back to top |
|
 |
mrmip Administrator

Joined: 05 Nov 2005 Posts: 230 Location: Finland
|
Posted: Wed Mar 05, 2008 4:03 am Post subject: |
|
|
So. We can regard this as settled then ?
Against all odds the Puzzle is mate in 3 afterall. _________________ There are only three kinds of chessplayers - those who can count and those who cannot.... |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Wed Mar 05, 2008 2:56 pm Post subject: |
|
|
and mate in 4 dont forget |
|
Back to top |
|
 |
mrmip Administrator

Joined: 05 Nov 2005 Posts: 230 Location: Finland
|
Posted: Thu Mar 06, 2008 7:01 am Post subject: |
|
|
No,no I will not forget - But here is a philosophical dilemma:
If a problem is mate in 3, can it also be a mate in 4?
Because a problem inherently reguires both parties to make very best moves, then when white makes a correct first move (key) the problem terminates after white's 3rd move and s/he never gets to play the 4th move - so the mate in 4 is impossible - No? _________________ There are only three kinds of chessplayers - those who can count and those who cannot.... |
|
Back to top |
|
 |
fianchetto Expert
Joined: 26 Sep 2007 Posts: 13
|
Posted: Thu Mar 06, 2008 4:42 pm Post subject: |
|
|
is the question always posed to find a mate with best possible moves, or simply to find a mate? With best possible moves the shortest path is reached, but if not best possible then any path is possible as long as it is forced, and in this case the mate in 4 is possible since it is forced moves from black's point of view, even if white did not choose the best possible first move. |
|
Back to top |
|
 |
mrmip Administrator

Joined: 05 Nov 2005 Posts: 230 Location: Finland
|
Posted: Fri Mar 07, 2008 12:12 am Post subject: |
|
|
Perhaps not always, but it certainly is customary. If stipulation says "Mate in 3 moves" then finding a mate in 4 or more moves has no relevance to solution.
I guess one could ask simply "Find a mate" in which case any mate - regardless how many moves it took - would be a correct answer.
But here the stipulation was clear : Mate in 3 moves. _________________ There are only three kinds of chessplayers - those who can count and those who cannot.... |
|
Back to top |
|
 |
|