Website founded by
Milan Velimirović
in 2006

1:21 UTC
ISC 2021



Remember me

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

Rating lists


MatPlus.Net Forum Misc Construction Challenge
You can only view this page!
Page: [Previous] [Next] 1 2 3
(21) Posted by Bob Baker [Wednesday, Apr 10, 2019 10:59]

Geir, do you suppose anyone here is working on surpassing your record count of 32 consecutive ! moves? I'm certain I wouldn't have a chance of succeeding at that.

To give people a less daunting task to work on, I will present the other challenge you suggested.

Construct a legal drawn position with no promoted pieces in which the first 12 or more half-moves include no captures and are ! moves for both sides under the Nunn Convention.
(Read Only)pid=17191
(22) Posted by Geir Sune Tallaksen Østmoe [Wednesday, Apr 10, 2019 22:18]

Hard to tell. I don't know how many usually try the construction challenges that appear now and then, and of course we will never know for sure how many try them without replying.

The non-capture challenge is a tough one. I have a possible idea, but I haven't been able to make it work.
(Read Only)pid=17197
(23) Posted by Bob Baker [Thursday, Apr 11, 2019 06:13]

Geir, if you don't mind answering a personal question, have you lived in Norway all your life? I'm wondering because it seems to me your use of English is better than that of quite a few native speakers of the language.
(Read Only)pid=17198
(24) Posted by Geir Sune Tallaksen Østmoe [Thursday, Apr 11, 2019 17:49]

Wow, thanks a lot! Except for half a year in Australia, I have lived in Norway my whole life.
(Read Only)pid=17199
(25) Posted by Hauke Reddmann [Saturday, Apr 13, 2019 16:45]

With teh Internets, you don't have to go anywhere :-)

On another point:

(= 2+3 )

This is the longest Nalimov EGTB sequence of "only" moves in the sense that all
alternatives by White throw away the win, but all by Black merely lose faster.
(Source: Rosie M. on ChessStackExchange/ Guy Haworth)
(Read Only)pid=17206
(26) Posted by Bob Baker [Saturday, Apr 13, 2019 22:17]

Unfortunately for our current challenge, the Nunn Convention doesn't give a ! to a move by Black when there is just one "best" losing move. But this is still an interesting position to know about.
(Read Only)pid=17208
(27) Posted by Bob Baker [Sunday, Apr 14, 2019 06:44]; edited by Bob Baker [19-04-14]

Let's call a move a semi-only move if it's the only winning move for a winning player or the only most protracting move for a losing player. Then the position Hauke gave is a candidate for the greatest number of consecutive semi-only moves for both sides from an initial position.

I checked that position and found that, on the way to mate in 50, the first non-semi-only move occurred on the 19th full move, when White had two winning moves. So the current record, based on that position, is 36 consecutive half-moves that are semi-only moves for both sides.

I'm not surprised that the record number for semi-only moves in a six-man won position is greater than Geir's record for ! moves in a drawn position. I wouldn't be surprised if the semi-open record were even greater in a seven-man position from the Lomonosov EGTB. But I wonder what Geir could do with this record and an unrestricted number of men. It's not a task I would have the slightest idea of how to tackle.
(Read Only)pid=17209
(28) Posted by Hauke Reddmann [Sunday, Apr 14, 2019 10:53]

To my best (and Guy Haworth's) knowledge the position still is a record for 7 men.
Cf. record review at
(Read Only)pid=17210
(29) Posted by Bob Baker [Sunday, Apr 14, 2019 15:26]; edited by Bob Baker [19-04-14]

Hauke, I found the position in Guy Haworth's database and I believe there is an error in the move sequence given. The 18th full move is given as
18. Rb2 Ne1. I believe the correct move is 18. Rb2 Ne3+ since Nalimov shows, for Black's 18th moves, Ne3+ loses in 32 and Ne1 loses in 20. Can you check with Guy Haworth to see if I am correct or have misunderstood the definition of the '' moves for Black? If I am correct, the sequence ends here because on White's 19th move both Kc5 and Kc6 are winning moves. Thank you.
(Read Only)pid=17211
(30) Posted by Geir Sune Tallaksen Østmoe [Sunday, Apr 14, 2019 19:52]

I hope you don’t expect too much from me with the new challenge, Bob. :) I have no idea of how to approach such a task, other than database mining. But I suppose you can use the existing 6-piece position and a couple of moves in the beginning.

(= 4+5 )

(Read Only)pid=17213
(31) Posted by Bob Baker [Sunday, Apr 14, 2019 23:09]; edited by Bob Baker [19-04-15]

Geir, you have already set a new record for this task with 40 half-moves, up from 36 half-moves for the position in Guy Haworth's database. The annotation there claims 46 half-moves, but I think it is mistaken, as I told Hauke in my previous post.

[Edit] Geir, I just realized that there is a form of the challenge which makes the Haworth database solution valid. Black must always have a unique most protracting move, but is not required to play it. If there is a suboptimal move that will give a longer sequence of unique moves, it may be played. (I had assumed Black is required to make the "best" move.)

In either case, your position is still good for a new record.
(Read Only)pid=17214
(32) Posted by Bob Baker [Monday, Apr 15, 2019 07:39]

I thought I should post my position for the task with no captures in case the record isn't broken for a while.

Each half-move (both sides) is a ! move (Nunn Convention)
No captures are allowed
White to play and draw

(= 2+3 )

1. Rb1 d3 2. Kg4 d2 3. Kf3 Kd3 4. Ra1 e2 5. Ra3+ Kc2 6. Ra2+

Ply count = 11
(Read Only)pid=17216
(33) Posted by Hauke Reddmann [Monday, Apr 15, 2019 23:12]

Bob - no mistake, but metric. You are talking about
the obvious distance-to-mate metric, but there are
tons of others, depending on e.g. minimizing the time
to conversion (=capture) or on the 50 move rule.
This should even stand somewhere in the article.
If not, cf. also:
(Read Only)pid=17225
(34) Posted by Bob Baker [Tuesday, Apr 16, 2019 02:49]

Thanks, Hauke.

Geir, here is something completely off the subject which I thought might interest you as a native of Norway. I follow the athletic events of my alma mater, Oklahoma State University. Their men's golf team is ranked #1 in the country but only finished 4tlh in their latest tournament. Their top golfer, Viktor Hovland, and their head coach were both absent because Hovland was invited to play as an amateur in the Masters, just won by Tiger Woods. His coach went with him and was his caddy. Hovland won the silver cup as the low-finishing amateur. He also made history by becoming the first-ever Norwegian to compete at the Masters.
(Read Only)pid=17228
(35) Posted by Bob Baker [Thursday, Apr 18, 2019 06:45]; edited by Bob Baker [19-04-18]

Geir, I showed your prelude to Conrady's winning-AUMS to Guy Haworth, who had a comment in his database that Conrady's was the longest one known to him. He agreed that Conrady remains the longest-known 5-man winning-AUMS, and that Ostmoe is now the longest-known 9-man winning-AUMS. That would make yours the longest currently known with any piece count.

I also showed him your 32-ply-count drawing-AUMS. He agreed it works, and doesn't think he had a maximal drawing-AUMS at all. One comment he made was "It would be interesting to know where 8/4p3/2k5/8/2P3p1/p7/P3P2P/7K w - - 0 10 came from." All I could tell him is that is not in HHdbV, and I don't know if it just turned up in your line, or you started with it and worked backward. But I said I would ask you.

I asked Guy if AUMS for Absolutely​ Unique Move Sequence is commonly used among problemists. I hadn't seen it before Hauke directed me to his database. It is not included at, but I have suggested it as an addition. They will add acronyms that they find to be in common use in a discipline, and I gave a pointer to his database as an official reference. I haven't heard back yet. If they need more references, do you know of any?
(Read Only)pid=17231
(36) Posted by Geir Sune Tallaksen Østmoe [Thursday, Apr 18, 2019 11:01]

I haven’t heard the term AUMS before, nor have I seen any such task before this forum thread.

I did not take the pawn endgame from anywhere, I just thought a pawn race in a pawn endgame would probably be a good place to maximize the number of «only moves». I think I started with something like this,

(= 3+3 )

then found too many problems with this type of setup, so I reduced the number of moves a bit,

(= 3+3 )

then I added a pawn on c4 to make Black’s king moves unique. This still doesn’t work for two reasons - White can draw by moving his king over to c2 to lock Black’s king in, and Black can afford to lose a move because queen versus a-pawn is a draw. This was the most difficult part to solve, but I eventually found that adding the e-pawns solved both issues.

(= 5+4 )

And then I worked backwards from here, adding as many silly exchanges as I could. :)
(Read Only)pid=17232
(37) Posted by Geir Sune Tallaksen Østmoe [Thursday, Apr 18, 2019 11:05]

BTW, thanks for the information on Hovland. I know about him, but I don’t follow golf closely, and Norwegian media does not seem very interested in golf either. They keep pretending cross country skiing is the world’s second largest sport.
(Read Only)pid=17233
(38) Posted by Bob Baker [Saturday, Apr 20, 2019 01:21]

Geir and Hauke,

I want to thank Hauke for referring me to Guy Haworth's database of endgame records. There is something I thought I should clear up. At comment #25, Hauke showed a position by Conrady from the database. Hauke said he was told that Black has to avoid moves that lose faster. In my discussions with Guy, I have learned that this is incorrect. The requirement for uniqueness only applies to White's moves. Black can (and at some points must) make moves that lose faster than others, and can even have more than one move choice. But each move by Black must be optimal, in the sense that no other move could lead to a longer sequence of White's unique winning moves. Otherwise, it would not produce the record-setting sequence for a position with seven or less men.
(Read Only)pid=17235
(39) Posted by Arpad Rusz [Wednesday, May 1, 2019 20:31]

The 5-piece record position seems to have 16 plies:

This is shorter by 2 plies but the play is more interesting:

I've found them with the help of Syzygy Endgame Explorer (SEE), my tool written for mining Syzygy tablebases.
(Read Only)pid=17261
(40) Posted by Bob Baker [Thursday, May 2, 2019 10:27]

Here is Árpád's 16-ply position:

(= 2+3 )

1. Kc6 g5 2. Kd5 Kg6 3. a4 Kf5 4. a5 g4 5. a6 g3 6. a7 g2 7. a8=Q g1=Q 8. Qc8+ Kf4

Árpád's tool did not disallow "only legal" moves or captures, but by chance, all of these moves are '!' moves and include no captures, so this position breaks my 11-ply record for the capture-free case. I found several "almost drawing" moves for White in the sequence. For example, 4. Kd6 loses and allows Black to mate in 172, or to mate in 87 if the 50-move rule is suspended.

I asked Árpád for his opinion of requiring '!' versus unique moves, which prevents some trivial perpetual check cases. His reply was, "I don't think only legal moves should be excluded. Instead, I propose to put perpetual checks in a different group."

If Geir is going to set a new record for the capture-free case, it will take something larger than a 5-man position.
(Read Only)pid=17262

Page: [Previous] [Next] 1 2 3

MatPlus.Net Forum Misc Construction Challenge