Noomen KI Fianchetto H6.02 vs. K11.2.2 30m+30s 16-core

[jdart]

Up until the 1940's, the Fianchetto was not only the main line KID, it was practically the only KID. The modern theory of it starts in the 1950's.

But I don't think it is the most dangerous line for Black. The Orthodox (E94-E99) is really hard to defend and chess engines particularly have a lot of trouble getting the kind of counterplay Black is supposed to have, because it can take a couple of dozen moves to generate that.

The Saemisch used to be considered the main antidote to the KID, but it has lost some popularity. But some lines such as E81 are still potent.

--Jon

[tpoppins]

+15 Elo +/-24 (BayesElo)
+42 Elo +30/-26 (Elostat)

As much as I agree on the +42 Elo, from pure statistics theory,
I wonder about your BayesElo +15 Elo, which is a world apart !... How can that be ?

I was puzzled by this as much as you are, Ernest. As far as using Bayeselo (and about statistics in general) I'm no better than a well-trained monkey. I can only check that no PEBCAKs crept in during the execution of the commands as per Bayeselo's usage documentation:

Code: Select all

version 0056, Copyright (C) 1997-2007 Remi Coulom.
compiled Jan 30 2007 20:30:07.
This program comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it
under the terms and conditions of the GNU General Public License.
See http://www.gnu.org/copyleft/gpl.html for details.
ResultSet>readpgn H602-K1122 Noomen KI Fianchetto.pgn
100 game(s) loaded, 0 game(s) with unknown result ignored.
ResultSet>elo
ResultSet-EloRating>mm
00:00:00,00
ResultSet-EloRating>exactdist
00:00:00,00
ResultSet-EloRating>ratings
Rank Name                          Elo    +    - games score oppo. draws
   1 Houdini 6.02 Pro x64-popcnt    15   24   24   100   56%   -15   82%
   2 Komodo 11.2.2 64-bit          -15   24   24   100   44%    15   82%
ResultSet-EloRating>

If you can reproduce the results with this PGN on your end then perhaps it's worth asking Kai Laskos and Dann Corbit in the General forum. The former is our resident statistics guru, the latter was an active participant in the Bayeselo development thread and I presume is familiar with its code; neither looks in here much, apparently.

[tpoppins]

Thank you for the input, Jon, very interesting. Someone should write a book on the history of the KID, it would be a fascinating read.

Which lines in E81 do you mean? There are quite a few: with Bg5; Nge2 vs. ...c5; the Byrne Variation (with ...c6 and ...a6); and the ...c5 gambit vs. Be3.

[jdart]

I am not that up on this but I think the 7. Nge2 lines are the main line these days, that is:

1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. f3 O-O 6. Be3 c5 7. Nge2

(7. cxd5 is not bad here).

and now Black's best is probably .. Nc6. The Nbd7 lines are not doing so well I think.

--Jon

[MikeB]

+15 Elo +/-24 (BayesElo)
+42 Elo +30/-26 (Elostat)

As much as I agree on the +42 Elo, from pure statistics theory,
I wonder about your BayesElo +15 Elo, which is a world apart !... How can that be ?

I was puzzled by this as much as you are, Ernest. As far as using Bayeselo (and about statistics in general) I'm no better than a well-trained monkey. I can only check that no PEBCAKs crept in during the execution of the commands as per Bayeselo's usage documentation:

Code: Select all

version 0056, Copyright (C) 1997-2007 Remi Coulom.
compiled Jan 30 2007 20:30:07.
This program comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it
under the terms and conditions of the GNU General Public License.
See http://www.gnu.org/copyleft/gpl.html for details.
ResultSet>readpgn H602-K1122 Noomen KI Fianchetto.pgn
100 game(s) loaded, 0 game(s) with unknown result ignored.
ResultSet>elo
ResultSet-EloRating>mm
00:00:00,00
ResultSet-EloRating>exactdist
00:00:00,00
ResultSet-EloRating>ratings
Rank Name                          Elo    +    - games score oppo. draws
   1 Houdini 6.02 Pro x64-popcnt    15   24   24   100   56%   -15   82%
   2 Komodo 11.2.2 64-bit          -15   24   24   100   44%    15   82%
ResultSet-EloRating>

If you can reproduce the results with this PGN on your end then perhaps it's worth asking Kai Laskos and Dann Corbit in the General forum. The former is our resident statistics guru, the latter was an active participant in the Bayeselo development thread and I presume is familiar with its code; neither looks in here much, apparently.

I'm as interested as you are. I received guidance years ago to use "mm 1 1" and the command "covariance" - my results (the error bars specifically) are easily duplicated with any current version of bayeselo - would be interested in any expert guidance on this matter. I read on the web ( much later) that when playing positions from both black and the white side, "mm 0 1" should be used. But I am certainly no expert and not able to to say what is correct or why it is correct.