43280/43258

[tomohawk]

Look identical, no?

[revenant]

(43280/43258)

Perfect, now I have a chance to play 1. Nxc6? under two different problem numbers!

Fortunately I've memorized 1. Bxg4 winning and just play it automatically whenever this position reappears -- which now that you mention it, did seem an oddly frequent occurrence of late...

No matter how many times I play it through, I never quite get how white comes out a piece up for free.

[tomohawk]

A trick I use for first inspection when playing long capturing sequences is to see if the player who captured first also captures last. Generally, if that happens, the sequence is good for that player.

[drahacikfm]

Interesting rule, I never thought of that.  Doesn't work though if you are capturing a bunch of pawns while your opponent is capturing pieces!

[richard]

Yes the ending positions are identical.  The generator uses FEN positions to detect duplicates and as this position was reached using different move sequences in both games the half move clock ended up being different and so the duplicate was allowed (strictly speaking the positions are different but you'd only notice if you had a tactic which required 48 moves with no pawn advances or captures :-) ).

Richard.

[dougy]

Also problem 43249.  No wonder I get this problem so often.

[richard]

Hi all,

Wow, just checked and there are a total of FOUR of these in the set, all arrived at with different move orders, a common  blunder I guess.

I also checked how often this occurs in general and there are 7 other positions that have a duplicate arrived at by transposition, but they all have only 1 duplicate.

Regards,
Richard.

[drahacikfm]

Richard,

The FIDE definition of a position for deciding if there is a 3-fold repetition does not take into account the number of moves until the 50-move rule applies.  It only takes into account the position of all the pieces, who is to move, and whether castling or en passant rights exist in one position and not the other.

Probably you should eliminate duplicate problems (based on the position the User must solve, not the position before the problem's "pre-move").  Or if you don't want to delete problems, at least filter new problems to not duplicate existing ones.

[drahacikfm]

(strictly speaking the positions are different but you'd only notice if you had a tactic which required 48 moves with no pawn advances or captures :-) ).

In the debate over whether some problems are "strategic" instead of "tactical", that might qualify as a strategic problem ... 48 moves in a row with no captures or pawn moves

[revenant]

What was the opening & ECO code in the 4x-dup'd problem?

In the words of the nabbed bank robber asking Clint Eastwood as "Dirty Harry"...   "I gots to know!"

(i.e., Did he fire six shots or only five?)

[drahacikfm]

Sicilian, Accelerated Dragon.  One way to get it:  1.e4 c5 2.Nf3 g6 3.d4 cxd4 4.Nxd4 Bg7 5.c4 Nc6 6.Be3 Nf6 7.Nc3 d6 8.Be2 0-0 9.0-0 and then the mistake 9...Ng4.

Many transpositions are possible, such as d6 before Nf6, etc.

[richard]

I'll be preventing this type of problem from being added in the future (by simply looking at only the position part of the FEN when looking at duplicates). I'll also disable the current most recent duplicates of each of the 8 duplicated positions when as soon as I have time to do so.  I'd like to test the disabling on my test machine first and I don't have access to it right now.

When I do rememove them I'll try and remember to post the full move list of the duplicates in the 43280/43258 series. If I remember correctly some of them had different move sequences from move 1 and transposed into the sicillian later.

Regards,
Richard.