Website founded by
Milan Velimirović
in 2006

5:57 UTC
ISC 2020



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum General Number of all "4 men only"
You can only view this page!
(1) Posted by Hauke Reddmann [Thursday, Dec 11, 2008 11:06]

Number of all "4 men only"

For simplicity, let's restrict to orthodox n# (since this is,
apart from studies*, the only thing implicitely contained in the
Nalimov tables). Obviously, the number of positions is smaller than
64*63*62*61*5*10. (And this is only the "dumb" approximation,
I'd knew a handful of ways to bring the number down.)

This means it *might* be possible to at least scan all 4 men
positions in the Nalimov and check if the key to shortest
n# is dual-free.
(Yes, an orthodox n# may have more than one solution if
the author insist so. Restrict to orthodox orthodox then :-)

Can it be done?
Should it be done?
What do you expect for the result? Which percentage will
Speckmanns collection already cover?


* I have a file with "all" <4 men studies, and I must say,
Longcat is loooooong.
(Read Only)pid=2979
(2) Posted by Dmitri Turevski [Thursday, Dec 11, 2008 15:21]

I did a similar thing couple of years ago, but abandoned it completely because it is boring (its boring level is over nine thousand). The whole approach looks like a deadend to me because it hits the "total pieces on board" limit too fast. I think, problem composing algorithm should search for combinations of motives/effects and check whether they are feasable rather than analyze ready setups. My pastime is working on it, but i doubt i have something ready soon, not to say ever.

Nevertheless, here's a twomover from my "bruteforce" attempts,
hardly original, but i like it:
(= 3+2 )
1... h5 2. Qf6#
1. Kf4? zz 1... h6 / Kh6(Kh5) 2. Qf7 / Q(x)h7# 1... h5!
1. Se6! zz 1... h6 / h5 / Kh6(Kh5) 2. Sf4 / Qg7 / Qg5#

Back to your questions:

>Can it be done?

>Should it be done?
No idea.

>What do you expect for the result?
I expect one ending his days digging through the pile of the resulting junk. It is not being dual-free what makes an aestetic value, not at all. I think one should search for more specific patterns like switchbacks, or rundlaufs, or underpromotions to find something interesting in his lifespan.

>Which percentage will Speckmanns collection already cover?
My guess is that it would be a rather small percentage. However, if you manage to leave the junk out somehow, then this percentage can be surprisingly high (i guess again).
(Read Only)pid=2980
(3) Posted by Michael McDowell [Friday, Dec 12, 2008 09:41]

The diagram adds a worthwhile try to a problem by Herbert Hultberg, Tidskrift for Schack 1928, which starts the white king on e4 and the knight on g7.
(Read Only)pid=2982
(4) Posted by Torsten Linß [Monday, Jan 12, 2009 23:22]

> Nalimov and check if the key to shortest n# is dual-free.
Why Nalimov? His tablebases are not designed to answer this kind of questions efficiently.

> Can it be done?
No, because there is no definition of a "correct problem".

> Should it be done?
You'll always find somebody who is willing to waste his time on anything...

> Which percentage will Speckmanns collection already cover?
What's Speckmann's collection? A collection of FMO problems?

> What do you expect for the result?
- Have a look at the problems in the collection and you'll know what to expect of problems of this kind.
- Or are you asking for a number?

BTW, do you understand the following problem?
Ke5 Db1 - Kh6 Tc3, #12
1.Db4 Te3+ 2.Kf4 Te8 3.Kf5 Tg8 4.Dh4+ Kg7 5.Dg5+ Kh7 6.Dh5+ Kg7
7.Dg6+ Kf8 8.Df6+ Ke8 9.Ke6 Tg6 10.Df6*g6+ Kd8 11.Dc2 Ke8 12.Dc8#
[Why not Db7 in the initial position?]
(Read Only)pid=3062
(5) Posted by Hannu Harkola [Tuesday, Jan 13, 2009 10:43]

If Db7 then it is 9# by 1.Kf6.
(Read Only)pid=3063
(6) Posted by Hauke Reddmann [Tuesday, Jan 13, 2009 17:27]

@ Torsten: Yes, Speckmann did a collection of orthodox FMO. E.g. look here:

Do I understand the #12? Depends on the definition of "understand".
I have ELO 2300 and thus it's no wonder I took a peek into the new
TB defenses in the KQ/KR ending - you'll never know :-) The trick is
to position the Q such that the R can't pester you with checks from
far behind. Thus I understand the "what", but the "how" might still
be above my ELO. (Luckily any game will be either a rapid where the
opponent doesn't have enough time either, or a long game, where I always
have one hour more on the clock - if I make it into the endgame :-)

(Read Only)pid=3064
(7) Posted by Kevin Begley [Thursday, Jan 15, 2009 09:24]

Also, if Db7, there's 1.Qg2! with a trivial mate in 11, which I assume went unstated because of the obvious duals.
(Read Only)pid=3067

No more posts

MatPlus.Net Forum General Number of all "4 men only"