Please turn on javascript in your browser to play chess.
Only Chess Forum

Only Chess Forum

  1. Subscriber Proper Knob
    Cornovii
    12 Mar '12 11:26
    gambit05 recently asked me if i wouldn't mind doing a 20 game engine analysis of his games, he had no objection to the results being posted here. So here they are.

    20 games against 2000+ rated players each with 20+ moves out of book.

    Houdini 1.5a 32bit, 192 Hash, 20 ply

    Game 8346249
    { White: gambit05 }
    { Top 1 Match: 17/32 ( 53.1% )
    { Top 2 Match: 24/32 ( 75.0% )
    { Top 3 Match: 26/32 ( 81.3% )

    { Black: zonnebloem }
    { Top 1 Match: 23/32 ( 71.9% )
    { Top 2 Match: 26/32 ( 81.3% )
    { Top 3 Match: 31/32 ( 96.9% )


    Game 8246445
    { White: st40 }
    { Top 1 Match: 15/27 ( 55.6% )
    { Top 2 Match: 19/27 ( 70.4% )
    { Top 3 Match: 26/27 ( 96.3% )

    { Black: gambit05 }
    { Top 1 Match: 14/26 ( 53.8% )
    { Top 2 Match: 17/26 ( 65.4% )
    { Top 3 Match: 20/26 ( 76.9% )


    Game 7322971
    { White: gambit05 }
    { Top 1 Match: 22/53 ( 41.5% )
    { Top 2 Match: 37/53 ( 69.8% )
    { Top 3 Match: 45/53 ( 84.9% )

    { Black: kes29 }
    { Top 1 Match: 24/53 ( 45.3% )
    { Top 2 Match: 33/53 ( 62.3% )
    { Top 3 Match: 37/53 ( 69.8% )


    Game 8077392
    { White: bogoluzhin }
    { Top 1 Match: 22/37 ( 59.5% )
    { Top 2 Match: 29/37 ( 78.4% )
    { Top 3 Match: 32/37 ( 86.5% )

    { Black: gambit05 }
    { Top 1 Match: 19/37 ( 51.4% )
    { Top 2 Match: 26/37 ( 70.3% )
    { Top 3 Match: 34/37 ( 91.9% )


    Game 7507924
    { White: jl1000 }
    { Top 1 Match: 13/31 ( 41.9% )
    { Top 2 Match: 21/31 ( 67.7% )
    { Top 3 Match: 26/31 ( 83.9% )

    { Black: gambit05 }
    { Top 1 Match: 20/32 ( 62.5% )
    { Top 2 Match: 24/32 ( 75.0% )
    { Top 3 Match: 26/32 ( 81.3% )


    Game 7507652
    { White: jankrb }
    { Top 1 Match: 34/58 ( 58.6% )
    { Top 2 Match: 47/58 ( 81.0% )
    { Top 3 Match: 54/58 ( 93.1% )

    { Black: gambit05 }
    { Top 1 Match: 34/57 ( 59.6% )
    { Top 2 Match: 44/57 ( 77.2% )
    { Top 3 Match: 51/57 ( 89.5% )


    Game 8262390
    { White: gambit05 }
    { Top 1 Match: 19/24 ( 79.2% )
    { Top 2 Match: 21/24 ( 87.5% )
    { Top 3 Match: 21/24 ( 87.5% )

    { Black: jankrb }
    { Top 1 Match: 13/25 ( 52.0% )
    { Top 2 Match: 19/25 ( 76.0% )
    { Top 3 Match: 23/25 ( 92.0% )


    Game 7507870
    { White: gambit05 }
    { Top 1 Match: 9/22 ( 40.9% )
    { Top 2 Match: 15/22 ( 60.2% )
    { Top 3 Match: 19/22 ( 86.4% )

    { Black: jli100 }
    { Top 1 Match: 11/22 ( 50.0% )
    { Top 2 Match: 17/22 ( 77.3% )
    { Top 3 Match: 20/22 ( 90.9% )


    Game 8077391
    { White: gambit05 }
    { Top 1 Match: 13/22 ( 59.1% )
    { Top 2 Match: 14/22 ( 63.6% )
    { Top 3 Match: 15/22 ( 68.2% )

    { Black: bogoluzhin }
    { Top 1 Match: 17/22 ( 77.3% )
    { Top 2 Match: 19/22 ( 86.4% )
    { Top 3 Match: 20/22 ( 90.9% )


    Game 7508086
    { White: David Tebb }
    { Top 1 Match: 12/21 ( 57.1% )
    { Top 2 Match: 15/21 ( 71.4% )
    { Top 3 Match: 18/21 ( 85.7% )

    { Black: gambit05 }
    { Top 1 Match: 13/21 ( 61.9% )
    { Top 2 Match: 17/21 ( 81.0% )
    { Top 3 Match: 19/21 ( 90.5% )


    Game 8052611
    { White: gambit05 }
    { Top 1 Match: 21/40 ( 52.5% )
    { Top 2 Match: 31/40 ( 77.5% )
    { Top 3 Match: 39/40 ( 97.5% )

    { Black: paolo2504 }
    { Top 1 Match: 24/40 ( 60.0% )
    { Top 2 Match: 31/40 ( 77.5% )
    { Top 3 Match: 35/40 ( 87.5% )


    Game 5558219
    { White: caissad4 }
    { Top 1 Match: 18/40 ( 45.0% )
    { Top 2 Match: 25/40 ( 62.5% )
    { Top 3 Match: 26/40 ( 65.0% )

    { Black: gambit05 }
    { Top 1 Match: 14/41 ( 34.1% )
    { Top 2 Match: 26/41 ( 63.4% )
    { Top 3 Match: 35/41 ( 85.4% )


    Game 6980280
    { White: gambit05 }
    { Top 1 Match: 15/34 ( 44.1% )
    { Top 2 Match: 24/34 ( 70.6% )
    { Top 3 Match: 28/34 ( 82.4% )

    { Black: Caesare }
    { Top 1 Match: 25/35 ( 71.4% )
    { Top 2 Match: 31/35 ( 88.6% )
    { Top 3 Match: 33/35 ( 94.3% )


    Game 6980546
    { White: Caesare }
    { Top 1 Match: 21/34 ( 61.8% )
    { Top 2 Match: 26/34 ( 76.5% )
    { Top 3 Match: 31/34 ( 91.2% )

    { Black: gambit05 }
    { Top 1 Match: 9/33 ( 27.2% )
    { Top 2 Match: 18/33 ( 54.5% )
    { Top 3 Match: 23/33 ( 69.7% )


    Game 5558220
    { White: gambit05 }
    { Top 1 Match: 24/39 ( 61.5% )
    { Top 2 Match: 34/39 ( 87.2% )
    { Top 3 Match: 36/39 ( 92.3% )

    { Black: caissad4 }
    { Top 1 Match: 19/40 ( 47.5% )
    { Top 2 Match: 26/40 ( 65.0% )
    { Top 3 Match: 30/40 ( 75.0% )


    Game 7199951
    { White: jankrb }
    { Top 1 Match: 14/21 ( 66.7% )
    { Top 2 Match: 18/21 ( 85.7% )
    { Top 3 Match: 19/21 ( 90.5% )

    { Black: gambit05 }
    { Top 1 Match: 7/20 ( 35.0% )
    { Top 2 Match: 12/20 ( 60.0% )
    { Top 3 Match: 15/20 ( 75.0% )


    Game 6775647
    { White: jankrb }
    { Top 1 Match: 18/28 ( 64.3% )
    { Top 2 Match: 26/28 ( 92.9% )
    { Top 3 Match: 28/28 ( 100% )

    { Black: gambit05 }
    { Top 1 Match: 15/28 ( 53.6% )
    { Top 2 Match: 21/28 ( 75.0% )
    { Top 3 Match: 25/28 ( 89.3% )


    Game 5668753
    { White: FabianFnas }
    { Top 1 Match: 18/34 ( 52.9% )
    { Top 2 Match: 30/34 ( 88.2% )
    { Top 3 Match: 33/34 ( 97.1% )

    { Black: gambit05 }
    { Top 1 Match: 11/33 ( 33.3% )
    { Top 2 Match: 18/33 ( 54.5% )
    { Top 3 Match: 24/33 ( 72.7% )


    Game 6486326
    { White: gambit05 }
    { Top 1 Match: 10/24 ( 41.7% )
    { Top 2 Match: 16/24 ( 66.7% )
    { Top 3 Match: 19/24 ( 79.2% )

    { Black: FabianFnas }
    { Top 1 Match: 20/25 ( 80.0% )
    { Top 2 Match: 23/25 ( 92.0% )
    { Top 3 Match: 24/25 ( 96.0% )


    Game 5668486
    { White: gambit05 }
    { Top 1 Match: 12/26 ( 46.2% )
    { Top 2 Match: 16/26 ( 61.5% )
    { Top 3 Match: 19/26 ( 65.4% )

    { Black: FabianFnas }
    { Top 1 Match: 14/26 ( 53.8% )
    { Top 2 Match: 18/26 ( 69.2% )
    { Top 3 Match: 21/26 ( 80.8% )


    Batch Summary

    gambit05 (Games: 20)

    Top 1 match: 318/642 ( 49.5% )
    Top 2 match: 455/642 ( 70.9% )
    Top 3 match: 537/642 ( 83.6% )


    Opponents (Games: 20)

    Top 1 match: 375/648 ( 57.9% )
    Top 2 match: 499/648 ( 77.0% )
    Top 3 match: 567/648 ( 87.5% )
  2. Standard member gambit05
    Mad Murdock
    12 Mar '12 12:49
    Thank you very much. Plenty to think about this analysis now.

    First of all, when I separate wins, draws, and losses, I get the following values:

    6 wins

    Top 1 match: 122/242 ( 50.4% )
    Top 2 match: 174/242 ( 71.9% )
    Top 3 match: 207/242 ( 85.6% )

    4 draws

    Top 1 match: 75/124 ( 60.5% )
    Top 2 match: 97/124 ( 78.2% )
    Top 3 match: 110/124 ( 88.7% )

    10 losses

    Top 1 match: 123/278 ( 44.2% )
    Top 2 match: 174/242 ( 64.7% )
    Top 3 match: 207/242 ( 77.0% )

    It seems that I played most engine-like in my drawn games. In general, a lot of top 3 matchups from me and my opponents are >80%, or even >85%. I don't know how many engines I have played here, but I think the top 3 match of 85% as threshold for engine use might be a bit low.
  3. Subscriber Proper Knob
    Cornovii
    14 Mar '12 08:59
    Originally posted by gambit05
    Thank you very much. Plenty to think about this analysis now.

    First of all, when I separate wins, draws, and losses, I get the following values:

    6 wins

    Top 1 match: 122/242 ( 50.4% )
    Top 2 match: 174/242 ( 71.9% )
    Top 3 match: 207/242 ( 85.6% )

    4 draws

    Top 1 match: 75/124 ( 60.5% )
    Top 2 match: 97/124 ( 78.2% )
    Top 3 match: 110/124 ( 88.7 ...[text shortened]... played here, but I think the top 3 match of 85% as threshold for engine use might be a bit low.
    You have the percentages wrong for your losses. You have 278 and 242 for the number of moves, i'm guessing it's 278 -

    Top 1 match: 123/278 ( 44.2% )
    Top 2 match: 174/278 ( 62.6% )
    Top 3 match: 207/278 ( 74.5% )

    With regard to the thresholds, if the best players ever don't breach them i see no reason to move them.
  4. Standard member gambit05
    Mad Murdock
    14 Mar '12 16:06
    Originally posted by Proper Knob
    You have the percentages wrong for your losses. You have 278 and 242 for the number of moves, i'm guessing it's 278 -

    Top 1 match: 123/278 ( 44.2% )
    Top 2 match: 174/278 ( 62.6% )
    Top 3 match: 207/278 ( 74.5% )

    With regard to the thresholds, if the best players ever don't breach them i see no reason to move them.
    Thanks again. 278 should be the correct number. By the way, my total number of moves differ from your total number of moves by 2, if I remember correctly.

    I just find the top 3 match of 85% a bit low. With the 10 games I won or draw, I reach that value, and many games of my opponents in your analysis as well (some are probably engine generated, but not all!). And there should be a buffer, since I am far away from the best players. Just my thoughts when I saw the numbers.

    Do you know what numbers Magnus Carlsen gets?
  5. Subscriber Proper Knob
    Cornovii
    14 Mar '12 19:05
    Originally posted by gambit05
    Thanks again. 278 should be the correct number. By the way, my total number of moves differ from your total number of moves by 2, if I remember correctly.

    I just find the top 3 match of 85% a bit low. With the 10 games I won or draw, I reach that value, and many games of my opponents in your analysis as well (some are probably engine generated, but not ...[text shortened]... ayers. Just my thoughts when I saw the numbers.

    Do you know what numbers Magnus Carlsen gets?
    Here's a 20 game batch of Magnus Carlsen ran by Phlabbit -

    Analysis "Rybka 2.2n2 32-bit Hash:128 Time:30s Depth:12-16ply"

    { Magnus Carlsen (Games: 20) }
    { Top 1 Match: 484/842 ( 57.5% ) Opponents: 457/837 ( 54.6% )
    { Top 2 Match: 628/842 ( 74.6% ) Opponents: 599/837 ( 71.6% )
    { Top 3 Match: 698/842 ( 82.9% ) Opponents: 673/837 ( 80.4% )
    { Top 4 Match: 733/842 ( 87.1% ) Opponents: 720/837 ( 86.0% )
  6. Standard member gambit05
    Mad Murdock
    14 Mar '12 19:35
    gambit05

    Top 1 match: 318/642 ( 49.5% )
    Top 2 match: 455/642 ( 70.9% )
    Top 3 match: 537/642 ( 83.6% )

    Magnus Carlsen

    Top 1 Match: 484/842 ( 57.5% )
    Top 2 Match: 628/842 ( 74.6% )
    Top 3 Match: 698/842 ( 82.9% )

    So, he is clearly superior (if engine matchups are an indication for that) for the top 1 and top 2 choices, but not so for top 3.

    Interestingly, he had 200 more moves in the same amount of games. I wonder if he is going earlier out of book (or his opponents), or whether the games are longer.

    Anyway, how comes that a nobody like me is able to choose in 83% of my moves the top 3 match of an engine, just about as often as Carlsen (OTB)?
  7. Standard member hedonist
    peacedog's keeper
    14 Mar '12 20:16 / 1 edit
    Originally posted by gambit05
    gambit05
    Anyway, how comes that a nobody like me is able to choose in 83% of my moves the top 3 match of an engine, just about as often as Carlsen (OTB)?
    The % shows how often you play the move the computer likes. thats not the same as the "best" move. Carlsen is more likely to play the best move rather than just the computers choice. So his top 3 being lower than you seems ok to me.

    The first choice, on the other hand, will contain a lot of "only" moves that a super GM and computer will spot. You are less likely to see those.
  8. 14 Mar '12 20:33
    a computer analysis is nothing but a statistical test, no?

    it's not a test of how good your chess is. a computer can't do that.
  9. Subscriber Proper Knob
    Cornovii
    14 Mar '12 20:45
    Originally posted by gambit05
    gambit05

    Top 1 match: 318/642 ( 49.5% )
    Top 2 match: 455/642 ( 70.9% )
    Top 3 match: 537/642 ( 83.6% )

    Magnus Carlsen

    Top 1 Match: 484/842 ( 57.5% )
    Top 2 Match: 628/842 ( 74.6% )
    Top 3 Match: 698/842 ( 82.9% )

    So, he is clearly superior (if engine matchups are an indication for that) for the top 1 and top 2 choices, but not so for top 3. ...[text shortened]... to choose in 83% of my moves the top 3 match of an engine, just about as often as Carlsen (OTB)?
    It was a larger sample just because just because a lot of the games selected had a lot of moves. They're all listed in the RJHinds lunacy thread.
  10. 15 Mar '12 13:08 / 1 edit
    Originally posted by gambit05
    gambit05

    Top 1 match: 318/642 ( 49.5% )
    Top 2 match: 455/642 ( 70.9% )
    Top 3 match: 537/642 ( 83.6% )

    Magnus Carlsen

    Top 1 Match: 484/842 ( 57.5% )
    Top 2 Match: 628/842 ( 74.6% )
    Top 3 Match: 698/842 ( 82.9% )

    So, he is clearly superior (if engine matchups are an indication for that) for the top 1 and top 2 choices, but not so for top 3. ...[text shortened]... to choose in 83% of my moves the top 3 match of an engine, just about as often as Carlsen (OTB)?
    it must of been hard for your friend ? why didn't you analyse it all - your-self?
  11. Standard member gambit05
    Mad Murdock
    15 Mar '12 15:35
    Originally posted by tim88
    it must of been hard for your friend ? why didn't you analyse it all - your-self?
    I don't have the tools.
  12. Standard member gambit05
    Mad Murdock
    15 Mar '12 15:40
    Originally posted by Proper Knob
    You have the percentages wrong for your losses. You have 278 and 242 for the number of moves, i'm guessing it's 278 -

    Top 1 match: 123/278 ( 44.2% )
    Top 2 match: 174/278 ( 62.6% )
    Top 3 match: 207/278 ( 74.5% )

    With regard to the thresholds, if the best players ever don't breach them i see no reason to move them.
    I just realised were the error came from. I copy-pasted the lines from my win matchups, but didn't changed the number of moves for top 2 and top 3.

    It should be like:

    Top 1 match: 123/278 ( 44.2% )
    Top 2 match: 180/278 ( 64.7% )
    Top 3 match: 214/278 ( 77.0% )

    So, the percentages were correct.