Move number bug in game viewer prevents ChessBase importing PGNs from Chess.com

I found a serious bug that misnumbers the moves in the game viewer inside articles and blogs.  It's hard to explain, so I'll just go with an example.  One of many infected posts is the recent, excelleng "A Travelling Chess Player, Part 2", by IM Bryan Smith.  I'll be posting some screenshots below, but please go to Smith's article to see for yourself.

This bug comes in 2 flavors.  Type A is shown below, and is pretty obvious:

In the sub-variation beginning with 20.Rbd1, look at the move 23.Rxc1.  After 23.Rxc1, there is a sub-sub-variation (23.Qd7 etc), and then it picks up back at 24...bxc1=Q.  Uh, FYI Chess.com, that should be 23...bxc1=Q.  It's not that I really care that much, but ChessBase does.  When I try to import the PGN, I just get a truncated game fragment.  And the bug is right there in the game score for everyone to see.  It's also in the PGN, have a look:

OK, moving on to Type B of the same bug, which is harder to spot.  From the same variation in the same game:

Believe it or not there is a move numbering error in the 23.Qd7 sub-sub-variation as well.  This one looks just hunky-dory in the game viewer, but check out the mess in the PGN:

It reads 23.Qd7 { This is the best defense. } 24...Rec8 24.Be4 Qc4.  That should be 23...Rec8, after the comment, and this also confuses ChessBase.

This problem doesn't interfere with my enjoyment of following the game in the viewer.  But it is a major pain if I want to import the game into my databases.  You can find many of these gamescores online, but not with the player annotations you often get with the Chess.com version.

Can someone out there please confirm this?  I submitted this bug twice to site support.  The first time they said thanks for pointing it out, but then didn't fix it.  The second time, they said they can't reproduce it, that I'm the only one with the problem.

Here is a screenshot of my database after importing this game:

thanks for letting us know! where are you generating these diagrams - in the diagram editor for blogs, forum posts, etc? or from http://www.chess.com/analysis-board-editor.html ?

erik,

These are screenshots of the diagram editor/viewer for blogs, etc., in IM Smith's article: "A Travelling Chess Player, Part 2".  I'm going to guess that when you put a game in a blog via the diagram editor, and you're just pasting the PGN file rather than entering all the moves and comments manually, that the editor somehow renumbers the moves according to its own algorithm.

It seems to assign bogus move numbers when highly-nested variations pick back up after a branching point.  For example, a level 3 sub variation branches to level 4 sub-sub, and then gets back to level 3 with the wrong move number.  That is the Type A.  Type B is when the move number is wrong after a comment, but in the PGN only, the game viewer looks OK.

The best way to catch these bugs is to get a heavily-annotated game with comments all the way out to level 4 and 5 variations, and keep importing into ChessBase until it comes through clean.

thank you!