Upgrade to Chess.com Premium!

What does "Mate in 2" or "Mate in 3+" really mean on TT?


  • 22 months ago · Quote · #1

    ModularGroupGamma

    Consider this problem:

    http://www.chess.com/tactics/server.html?id=191374

    This is not a mate in anything (at least not less than 10 moves), yet it is tagged as such.  Moreover, the "definition" of mate in 3+ given in the help topics is:

    • Any puzzle/position that ends with mate in three or more moves.

    This is ambiguous at best.  Does it mean merely that the puzzle (including the opponent moves) happens to end in mate, or does it mean the usual definition, i.e. it is a forced mate?  I think it is very confusing and deceptive to tag this a "mate in 3+" problem.

    Also, no one has commented on a previous thread I started asking why TT does not include castling possibilities, which can have an effect on the correct solution to a problem:

    http://www.chess.com/forum/view/suggestions/does-tt-display-castling-possibilities

  • 22 months ago · Quote · #2

    Bubatz

    A tag may be wrongly given (I didn't check the position), but why is the tag itself ambiguous? "mate in n+ moves" means that there is a forced mate, but it takes at least n moves.

  • 22 months ago · Quote · #3

    ModularGroupGamma

    The tag is ambiguous because it does not include the word "forced".  Technically, if you take the definition as it stands, this is a mate in 3, because it is a puzzle that ends with mate, and is 3 moves long.  The definition should say,

    • Any puzzle/position that leads to a forced mate in 3 or more moves

    You might say this is implied in the terminology, but not everyone may know this, and this might be the reason for incorrect tags.

    Another reason I mention this is because there was a "puzzle of the day" the other day titled "Mate in 2" that was not, and this was pointed out, and the title never changed, and no staff even bothered to comment on the thread, "our bad, you're right, it's not mate in 2", which leads me to wonder what their definition actually means.

  • 22 months ago · Quote · #4

    Bubatz

    Yes, I'd say this is semantically implied by terminology as "mate in n" in puzzle books always means "forced mate in n".

    I concede that there's a problem with people setting incorrect tags, but this is a problem throughout the web.

  • 22 months ago · Quote · #5

    Escapest_Pawn

    Agree that "mate in 3+" is so open-ended as to be meaningless (unless perhaps synonymous with "eventually winning").  Another complaint with your reference problem is that the "loser" does not play optimally after the "winner's" first move, which is a common problem in TT, and often done to illustrate a tactical point.  I think a more general point (or question) is what is the practical or learning purpose of tags?  If seen before doing the puzzle, they are unfair hints.  If seen after doing the puzzle, they are trivial.

    I think I have seen TTs that involve castling. Perhaps a tag saying whether or not castling is still permissible would be useful.

  • 22 months ago · Quote · #6

    Here_Is_Plenty

    This really cleared things up for me.  I thought mate in 3 was a threesome and mate in 3+ was an orgy.

  • 22 months ago · Quote · #7

    ModularGroupGamma

    Escapest_Pawn said: "Another complaint with your reference problem is that the 'loser' does not play optimally after the 'winner's' first move, which is a common problem in TT, and often done to illustrate a tactical point."

    I don't have a problem with this so much, as long as the less than optimal play is not too far short of optimal.

    "I think a more general point (or question) is what is the practical or learning purpose of tags?  If seen before doing the puzzle, they are unfair hints.  If seen after doing the puzzle, they are trivial."

    I thought maybe they use the tags to collect statistics for people so they can see what types of tactics they do well on, and which they need to improve?

  • 22 months ago · Quote · #8

    Escapest_Pawn

    I thought maybe they use the tags to collect statistics for people so they can see what types of tactics they do well on, and which they need to improve?

    This may well be the case and such had not occurred to me.


  • 22 months ago · Quote · #9

    Kphlash

    I just looked at the puzzle. It is in fact mate... Im not sure what the issue is. Sometimes the puzzle does not show the last move or two as there are 2 equally good moves that lead to mate in the same amount of moves, so the puzzle ends instead of giving an "alternate correct", but this is not one of them.

    I am kind of confused- did you not see the pawn on b3? For the most part the tags are correct. If you feel they are incorrect, use the "analysis board" and look at the lines showed. Play the moves and see what it shows. Then give a thumbs down to tags you KNOW are incorrect. If you are not sure, do not change them.

  • 22 months ago · Quote · #10

    ModularGroupGamma

    Kphlash,

    The analysis showed by the chess.com engine does not show mate:

    • ({12:+5.56} 40. Rb8 Nb1 41. Rxb6+ Kxb6 42. Qxb4+ Ka7 43. Qd6 Rb7 44. Qxe6 Nc3 45. Kg4 a5 46. Kf5 Rxb3 $18)

    I also plugged the position into Houdini and it gave about +6 eval after about 10 moves, but no forced mate.  Mate seems only to be forced following the blunder 40. ... Rb7 ??.  Of course, maybe the analysis is wrong, and I entered in the position incorrectly on Houdini.  If you have a proof of the mate, I'd be interested to see it.

    I can't vote on tags, because I don't have a premium membership.

  • 22 months ago · Quote · #11

    Kphlash

    The puzzle itself is a mate in 3 due to the blunder Rb7 that allows for the Qc5#. If the person played more accurately (which is a complaint I hear alot "why didnt XX play XX instead") the puzzle wouldnt be possible.

    All those other lines are generally better than the ones played (the ones played are mistakes, in order to present tactics- Rb7 in this case). If the other moves are played after Rb8 like Nb1, which is the only move shown in the analysis at that point (also the one you posted) white would be up 5.5pts roughly- houdini and the chess.com analysis are correct. 

    This is the full analysis sheet. Correct? You posted what black should have done after Rb8, just below that, is what actually happened. The computer did not analyze that line because Rb7 is a stupid move, but that stupid move is the point of the puzzle.

    ({12:+5.18} 39. ... Kb5 40. Rb8 Nc2 41. f5 exf5 42. Rxb6+ Kxb6 43. Qxf5 Nxd4+ 44. exd4 Rc3+ 45. Kf4 Rxb3 46. Qe6+ Kc7 47. Qxd5 $18)
    ({12:-0.01} 39. ... Qa5 40. Qe8+ Kb7 41. Rb8+ Ka7 42. Ra8+ Kb6 43. Qb8+ Kc6 44. Qe8+ Kb6 $10)
    ({12:+4.32} 39. ... Qa7 40. Ra8 Qb7 41. Qe8+ Kb6 42. Rb8 Qxb8 43. Qxb8+ Kc6 44. Qe8+ Kd6 45. f5 exf5 46. Qd8+ Kc6 47. Qf6+ $18)

    40. Rb8

    ({12:+5.56} 40. Rb8 Nb1 41. Rxb6+ Kxb6 42. Qxb4+ Ka7 43. Qd6 Rb7 44. Qxe6 Nc3 45. Kg4 a5 46. Kf5 Rxb3 $18)
    ({12:+2.58} 40. f5 Ka5 41. Rb8 Rb7 42. Rxb7 Qxb7 43. fxe6 Nb5 44. e7 Nc7 45. e8=Q Nxe8 46. Qxe8 $18)
    ({12:+0.01} 40. Rd6 Rc6 41. Rd8 Rc7 $10)

    40. ...Rb7
    41. Rxb7 Qxb7
    42. Qc5++

    Basically, the tags show that the puzzle is a mate in 3 because, due to the mistakes made by your opponent, it is a mate in 3. not what it would have been with better play. Also, in regards to the castling, There are a few puzzles where castling is the solution. 

    I think the main point here, which Retgruvie is pretty famous for stating on those puzzles "do not worry about the moves you opponent makes, or what they could have done differently/better, the point of these puzzles is to recognize a mistake and capitalize on it the best possible way" sometimes, it is a piece capture, sometimes mate, and sometimes it is a simple defensive move because they missed a golden opportunity.

  • 22 months ago · Quote · #12

    ModularGroupGamma

    Kphlash,

    I think you misinterpret my complaint.  My complaint is not that black played less than optimally.  My complaint is that the tag is incorrect.

    "The puzzle itself is a mate in 3 due to the blunder Rb7 that allows for the Qc5#."

    Except this completely re-defines what the traditional definition of "mate in 3" is.  "White mates in 3" means, from the original position of the puzzle, white can force mate in 3 moves, but not 2.  The term does not mean, "white can force mate in 3 moves, at some future unspecified position of some variation, provided black blunders or plays some particular move."

    It is true that with the less than optimal play of black that presents itself in this interactive puzzle, a position is reached which is mate in 3.  But if this puzzle were presented in a chess column in a newspaper with the heading "White to move and mate in 3", there would be many letters to the editor correcting them.

    So I have a question.  If a particular TT puzzle has less than optimal play that blunders into a mate in 1, is it correct to tag this puzzle with "mate in 1"?  Is the starting position, after 1. f3, a "mate in 1" puzzle for black, because black can play 1. ... e5 and white has the opportunity to blunder into fool's mate??

    This really underscores the fact that the "definitions" given in the help feature are ambiguous.  Some people say the definition means what I think it means, and the tag is wrong, and someone else says the definition means something else, and the tag is correct.  Obviously someone is mistaken.

    It would be nice if some clarification were provided by staff, so that we all knew what we were talking about, and so that we knew if 99% of chess players in the world had to revise their definition of "mate in n" when looking at TT tags.

    "If the person played more accurately (which is a complaint I hear alot 'why didnt XX play XX instead') the puzzle wouldnt be possible."

    No, the puzzle would be possible, it just wouldn't be this particular puzzle.  In fact, it should be possible to add another puzzle to TT that does continue with the best black response (which I believe is actually 40... Qxb8) although that puzzle would be much easier to solve, since 41. Qxb8 is easier to see.  Who knows, that puzzle might already exist in TT.

    Please understand me, I know why the blunder 40... Rb7 was chosen as a line... it makes a more interesting puzzle!  But it does not make it a "mate in 3" puzzle.  The most that can be said is that the puzzle follows a line that leads to a "mate in 3" position, but that is not what most chess players (at least not those who grew up learning tactics puzzles in actual books, not online) would understand as a "mate in 3" puzzle.

    "Also, in regards to the castling, There are a few puzzles where castling is the solution."

    That wasn't my concern.  My concern was that there are tactics problems where whether castling is possible or not affects the outcome of the analysis.  This is why all chess engines have options to check off for castling when entering a position.

  • 22 months ago · Quote · #13

    Kphlash

    Gotcha. Yes, I'm aware the puzzle is mistagged as it is not a forced mate. I thought you said the puzzle didn't end in mate (and just missed it).

    The tags are often incorrect, but you cannot see the tags until after you solve the puzzle (unless there is an option I missed) so, even though the tag is not correct in a true puzzle sense, it is more of a hint to those who originally solved it incorrectly.

    This "mistagging" as well as the castling issue you are concerned about seem to be solved with the same answer- It is the way the puzzle is designed.

    The puzzles are designed for you to make the best move, regardless of the opponent. I forget the exact formula, but one of the Moderators explained it once- they are coded, the lower numbers (not rating) the 1st move is a mistake by your opponent, the higher numbers, the 2nd move is the mistake. All the puzzles are based on a mistake.

    Just do the puzzles the best you can, and dont over analyze the tags- play them as if it were a real game.

    If you have a suggestion or a comment, go here http://support.chess.com/ and contact the staff in the bottom left hand corner.

    Hope that helps solve your problem better. 

  • 22 months ago · Quote · #14

    ModularGroupGamma

    Kphlash,

    I'll do that. The castling issue is not addressed by just saying "that's the way the puzzles are designed", though. Whether castling is allowed or not can change what the solution is. Not providing such information can make it impossible to determine the best move.


Back to Top

Post your reply: