|
(v) Vitaly Medintsev & René J. Millour (France) |
Original Fairy problems |
No.1347 
|
No.1347 Vitaly Medintsev |
Solution: (click to show/hide) |
|
white Qa2 Pe5a3g3 Ke4 Bb8
black Kh1 Bg8
hs#8 (6+2) C- |
|
|
|
|
|
No.1347.1 Vitaly Medintsev & René J. Millour |
Solution: (click to show/hide) |
|
white Qb3 Pe5 Ke4
black Kh1 Bg8 Sa6
hs#8 (3+3) |
|
Q blocks B for WK to cross d5 and B blocks Q for BK to cross d5! Helpself well conceived. Is Pg3 needed?
Pg3 is a cookstopper.
For example: 1.Kf3 Kg1 2.e6 Kf1 3.e7 Ke1 4.Kg2 Be6 5.Kh1 Kf1 6.e8=Q/R Bc8 7.Bh2 Bh3 8.Qg2+ Bxg2#
Nice King chase !!
Yes g3 is needed to avoid a mate with king on h1
True!
Hello Vitaly,
Pg3 stops a cook by Bb8, Pa3 is another anti-cook, these 3 white units are not brilliant in this problem.
Is the following proposal suitable? White Ke4 Qb3 Pe5 Black Kh1 Bg8 Sa6.
1.Qe6 Kg2 2.Kd5 Kf3 3.Kc6 Ke4 4.Qb3 Bc4 5.Kb7 Kd5 6.Ka8 Kc6 7.e6 Bd5 8.Qb6+ Kxb6#.
A question of taste: with bKh5/6 and bBf7, a non-linear bK-march seems also possible. 1.Qe6 Kg5 2.Kd5 Kf4 etc.
Advantages: Sa6 controls b8 and prevents cooks, switchback on b3, miniature 3+3.
Too long for the computer, as you wrote! Is it correct? That is the question!
What is your opinion?
Best wishes!
Hello René,
thank you for suggested version.
Please, give me some time to analyze this opportunity. I’ll tell my opinion, in a while.
Well René, on this moment I’m optimistic about the proposed version 🙂
Aleksey Oganesian started testing it on popeye engine under MoveNumbers option. After 3 and half hours the first output line has appeared – Qb3-a2. This means that the problem has no short solution including hs#7.5 Then he stopped the engine.
I’ve tested last 6 moves of author’s solution. No cooks. So, I think this version might be correct – highly likely 🙂
I’m gonna ask Julia to publish this version as a joint problem. And I prefer linear King chase in view of better usage of chessboard space as well as funny geometry in the play of both Kings. Hope, you don’t mind.
Best regards!
It’s a little pity that BS is static piece but the advantages listed by you are such attractive!
Vitaly,
Many thanks for the good news!
The non-linear route was just a possibility. I also prefer the linear one, for the same reason.
Instead of 1…Kh1-g2, a bS-move (for example 1…Sb8-a6) would be dramatic: cooks! Sa6 must be static!
I am glad you are satisfied. You speak of joint problem. Many thanks, I am proud, and honoured, to sign with you!
All good wishes!
I have decided to still try a computer check of this problem (a beautiful 6-piece version), based on Vitaly’s information from 30.11.
I used Popeye 4.63 (I had seen much slower results from later versions).
I got first move to be started in 1 hour 50 minutes _and did not stop the calculations_.
The first move e5-e6 was checked in 60 hours. I was tempted to stop, but have not done it. And then move Ke4-f4 was calculated in only 7 more hours. I think that a possibility of earlier promotion could have lead to much bigger tree. Anyway, e5-e6 is the only move to make a faster promotion, so it is quite possible that two more weeks will be enough for the result.
If anyone wants to join he may start with startmovenumber 15 or more.
Georgy,
true, the time of testing dramatically depends on Popeye version.
I was testing several problems with stipulations H# and HS# on many different Popeye versions. The fastest is Popeye WINDOWS-32Bit-Version 3.70. It solves problems much faster (up to 1,7 times) than latest Popeye versions, including 64Bit-Versions. Tested on both CPU brands: Intel and AMD!
Another strange fact is the order of possible introductory moves (displayed under Movenumbers option) differs radically from one version to another when solving the same problem…
The problem 1347.1 is C+ (92 hours).
Fantastic!
Thank you Georgy!
Just great! Have updated the page, under the solution of 1347.1: (C+ by Popeye 4.79, tested by Georgy Evseev, 92hr.)
C+ by Popeye 4.63
Such wonderful news for Vitaly and me!
I am really very happy!
Many thanks, Georgy, for these 92 hours!
I tested with Popeye v.4.63 from position (R.Millour version) after 1.Qe6 Kg2 as HS#7 and it solved in 3hr.52 mins!
This is just for information!
Thank you!
I am a bit surprised by the choice of the authors.
Black king h6 with black bishop on f7 (if correct!) as suggested by René-Jean, seems better. It adds that the black king cannot go through f8, so that his route is less mechanical.
You are right, Jacques and I understand why you are surprised 🙂
But the difficulty of solution is not a predominant criterion for me.
I repeat: I prefer linear “King chase” (© Seetharaman Kalyan) in view of better usage of chessboard space as well as funny geometry in the play of both Kings.
I tested this problem with new Gustav 4.1a (by Olaf Jenkner), elapsed time 03 : 11 : 04,33 h (Hashtables 45055 MB). This is better than 92 hours 🙂