Upgrade to Chess.com Premium!

Inconsistent computer analysis

  • DK81
  • | Jul 18, 2012 at 9:20 AM
  • | Posted in: DK81's Blog
  • | 259 reads
  • | 2 comments

I submitted a recent game for computer analysis with the first four moves as follows:

1.  f4   d5

2.  Nf3   Nf6

3.  d4    c5

4.  e3   Bg4

The suggested alternative to 4. ….. Bg4 was 4 ….. Nc6

Later I submitted a second game for computer analysis with the same initial moves (but in a different sequence), and the suggested alternative to 4. ….. Bg4 was 4. ….. cxd4

Thus, the computer analysis seems to be inconsistent, treating identical positions differently.  Comments are awaited!

Comments


  • 4 months ago

    DK81

    Thank you, Black_Knight, for your reply nearly six months ago concerning my blog on the topic of inconsistent computer analysis.  Over 200 people have now viewed the comments, but no-one else has taken the trouble to add anything.

    Your comment about the computer's mode of operation was interesting.

    I discovered early on that the computer analysis was most useful for missed tactics and combinations, and sometimes of little or no value near the end of a game.  I omitted the follow-on moves in my blog item in order to keep the entry as short as possible, so I was not really interested in a single move isolated from follow-up moves.

  • 10 months ago

    Black__Knight

    Sometimes the alturnatives are so close in the computer's evaluation, that the computer could go back and forth between move alturnatives in just milliseconds. It doesn't surprise me to say the least. Not to mention, chess.com is always trying to improve on it's computer evaluations.  

    Use the computer analysis for missed combinations and tactics, not for suggestion on single moves because the computer does not think like humans. 

Back to Top

Post your reply: