The other solution to part b
Impossible-to-reach position with the least amount of pieces?
Sort:
Nov 17, 2020
0
#23
Well done, @infestationPit and @Travkusken! The point is that all 3 solutions have a symmetrical "try" (near-miss) that fails due to a possible prior promotion.
Nov 17, 2020
0
#24
Rocky64 wrote:
Well done, @infestationPit and @Travkusken! The point is that all 3 solutions have a symmetrical "try" (near-miss) that fails due to a possible prior promotion.
These are also illegal
MARattigan wrote:
Rocky64 wrote:
Well done, @infestationPit and @Travkusken! The point is that all 3 solutions have a symmetrical "try" (near-miss) that fails due to a possible prior promotion.
These are also illegal
Not this again...
The Black queen is residing on de4.5.
Haven’t been reading the post, but uhhh... zero pieces. I mean that’s illegal. Or three. Kings on board plus a pawn on The first
Nov 17, 2020
0
#28
Mako_Cat wrote:
Haven’t been reading the post, but uhhh... zero pieces. I mean that’s illegal. Or three. Kings on board plus a pawn on The first
If you read my original post, I don't consider clearly illegal board positions valid, ie something that just does not conform to the rules of chess. For example, the rules of chess mandate that when a pawn reaches the final rank, it has to be promoted to a piece that's not a pawn nor a king. Therefore it would be rules-breaking for there to be a pawn on the last rank.
Likewise having more or less than two kings (of different colors) on the board would be an illegal position because there rules do not support this in any way, shape or form. There have to be exactly two kings, and they have to be of opposite colors, and a king cannot be removed from the board, and nothing can become a king. Else it's not even theoretically a valid position.
A position fulfills my idea when there's nothing inherently wrong with it according to the rules of chess, but it's merely impossible to reach from the standard starting position. The position could theoretically be valid rules-wise (there are exactly two kings, they are not both in check at the same time, and so on), it's just that it's impossible to reach via legal moves from the starting position.
Having too many of a particular piece (eg. ten queens of the same color) would theoretically fulfill this definition, but I consider that too trivial and "cheap" of a solution.
Nov 17, 2020
0
#29
DjVortex wrote:
Mako_Cat wrote:
Haven’t been reading the post, but uhhh... zero pieces. I mean that’s illegal. Or three. Kings on board plus a pawn on The first
If you read my original post, I don't consider clearly illegal board positions valid, ie something that just does not conform to the rules of chess. For example, the rules of chess mandate that when a pawn reaches the final rank, it has to be promoted to a piece that's not a pawn nor a king. Therefore it would be rules-breaking for there to be a pawn on the last rank.
Likewise having more or less than two kings (of different colors) on the board would be an illegal position because there rules do not support this in any way, shape or form. There have to be exactly two kings, and they have to be of opposite colors, and a king cannot be removed from the board, and nothing can become a king. Else it's not even theoretically a valid position.
A position fulfills my idea when there's nothing inherently wrong with it according to the rules of chess, but it's merely impossible to reach from the standard starting position. The position could theoretically be valid rules-wise (there are exactly two kings, they are not both in check at the same time, and so on), it's just that it's impossible to reach via legal moves from the starting position.
Having too many of a particular piece (eg. ten queens of the same color) would theoretically fulfill this definition, but I consider that too trivial and "cheap" of a solution.
FIDE defines an illegal position in art 3.1.3
A position is illegal when it cannot have been reached by any series of legal moves.
which, together with the proviso that the moves could be played in accordance with the remaining articles (including art 2.3 - the starting position) corresponds with normal usage.
So to ask for a legal position that can't be reached from the starting position is to ask the impossible.
The things you say make a position clearly illegal do so only because it's possible to prove that they can't be true of a position reached by a series of legal moves from the starting position. This depends entirely on the starting position. You can't have a pawn on the first or last rank because there aren't any in the starting position and no series of legal moves will produce one. Each player must have exactly one king because they do in the starting position and no series of legal moves will result in either player gaining or losing one. You can't have multiple pieces on a single square because the starting position has no such cases and legal moves preserve that status.
But what is "clearly" illegal is ill defined. If you want the question to be well defined you have to list what counts as "clearly". E.g. you say both kings may not be in check, but not that the king of the side not having the move may not be in check. Does that mean that you will accept such positions? Those in my post #33? (The second position in #33 can legitimately occur in a game if Black is in process of playing Qe4-d5, for example, but wouldn't be legal under the definition because a series of legal moves from the starting position would leave all the pieces occupying squares.)
Nov 17, 2020
0
#30
Oddly enough, endgame tablebases can help us to decide if some positions (7 pieces or fewer) are illegal or plain nonsensical. While the software plays perfect chess, it's not smart enough to detect illegal positions and it would happily analyse, say, a position with WPg2 and WBh1. However, it rejects positions with properties like these as "invalid": missing K, adjacent Ks, both Ks in check, P on 1st/8th rank. So a "legitimate" illegal position is one that's accepted by tablebases, but a nonsensical or invalid one is not.
Nov 17, 2020
0
#31
Rocky64 wrote:
Oddly enough, endgame tablebases can help us to decide if some positions (7 pieces or fewer) are illegal or plain nonsensical. While the software plays perfect chess, it's not smart enough to detect illegal positions and it would happily analyse, say, a position with WPg2 and WBh1. However, it rejects positions with properties like these as "invalid": missing K, adjacent Ks, both Ks in check, P on 1st/8th rank. So a "legitimate" illegal position is one that's accepted by tablebases, but a nonsensical or invalid one is not.
________________________________________________________________________________________
You could point to the definition here (section "Legality" at the bottom of the page)
https://kirill-kryukov.com/chess/nulp/method.html
but that leaves something to be desired. It misses out for example that the piece placements should be on squares and no more than one piece should occupy the same square and that e.p. captures should be restricted to a single file. It also says a position is illegal if the side to move can capture the opponent's king, but art. 1.4.1 of the FIDE handbook says, "Leaving one’s own king under attack, exposing one’s own king to attack and also ’capturing’ the opponent’s king is not allowed ", so this would never be the case. It should rather say, "if the side not to move is in check". The coining of the phrase "Number of Unique Legal Positions" there is particularly unfortunate referring neither to unique positions in a normal sense, nor, necessarily, legal positions.
As far as tablebases are concerned they contain only positions that are won for one side or other. It's up to the software that queries the tablebase to use the same legality conditions that were assumed in the tablebase. But this can give inconsistent results. E.g.
if the above position is queried here
https://syzygy-tables.info/?fen=R3k3/8/3N4/7Q/8/8/8/4K3_b_-_-_0_1
it says "Invalid position", whereas if I query it through Wilhelm in the Nalimov tables it gives mate in 3. Whether or not it exists in the Syzygy tables I can only guess - it could be the site that is censoring the position. Illegal double checks are accepted by syzygy-tables.info.
Of course you can't get positions like those in #33 through any computer interface anyway, even though the second could legitimately occur during play.
PromotionZone wrote:
This is one with 5 moves, but it's still illegal because the 2 pawns are blocking the bishop's exits, so technically it couldn't have moved.
In a similar way:
If you notice carefully, the 3rd rank is entirely guarded by the pawns, so technically the king should not be able to get to the 1st rank.
Not illegal, white could have promoted to a bishop
Correct me if I'm wrong, but this should be an easy solution.
Nov 17, 2020
0
#34
Chessguy149 wrote:
Correct me if I'm wrong, but this should be an easy solution.
Consider yourself corrected. Try the board the other way up.
MARattigan wrote:
Chessguy149 wrote:
Correct me if I'm wrong, but this should be an easy solution.
Consider yourself corrected. Try the board the other way up.
Wdym? This position is impossible to get to
Oh nvm, I forgot about promotion. Dang.
Nov 17, 2020
0
#37
Epiloque wrote:
wtf #33
Illegal positions satisfying @Rocky64's conditions.
Nov 17, 2020
0
#38
MARattigan wrote:
Epiloque wrote:
wtf #33
Illegal positions satisfying @Rocky64's conditions.
Nah, I asked for illegal positions, but those are not even positions.
MARattigan wrote:
Chessguy149 wrote:
Correct me if I'm wrong, but this should be an easy solution.
Consider yourself corrected. Try the board the other way up.
Aaaaargh! I was going to say something very similar until I read this.
Rocky64 wrote:
MARattigan wrote:
Epiloque wrote:
wtf #33
Illegal positions satisfying @Rocky64's conditions.
Nah, I asked for illegal positions, but those are not even positions.
Ironically enough, I cannot find a phrase in the FIDE handbook (or at least I failed to spot it) which states that "a piece cannot occupy more than a single square at any turn".
This is in contrast to Chinese chess where a piece must occupy the intersection of two lines (i.e. occupying up to four squares at once).
My first thought for part a was king on b8 and rook on a8 but that's a legal position