On-line engine blitz tourney February

[Joost Buijs]

The monthly on-line engine blitz tourney for February will take place on:

Saturday February 22, 2:00 PM EST (Boston time), 20:00 CET (Amsterdam time)

To connect:

winboard -zp -ics -icshost nightmare-chess.nl -icshelper timeseal -fcp ENGINE.exe -fd ENGINEFOLDER -autoKibitz

(for UCI engines, add -fUCI)

People without an account on the server can submit a request for a (new) account to admin@nightmare-chess.nl

Tourney format: 9 rounds Swiss 5/1

Our tournament manager will be: Eendje (Lisebeth) from ICC

[Daniel Anulliero]

Cool!
Isa will be here of course

[Joost Buijs]

Everything went smoothly.

rofChade did very well and ended on par with LczThinker (Fat Fritz on a 2080ti).
Ofcourse I'm happy with the third place.

Tourney results:

Code: Select all

:Tourney Players: Round 9 of 9 
:
:     Name              Rating Score Perfrm Upset  Results 
:     ----------------- ------ ----- ------ ------ ------- 
:  1 +rofChade          [2898]  8.0  [3009] [ 120] +20w +09w =03b +13w +04b =02w +06b +07w +10b 
:  2 +LczTinker         [2653]  8.0  [3006] [ 762] +25w =04b +10w +06b +03w =01b +08w +13w +09b 
:  3 +NightmareX        [2785]  6.5  [2784] [ 172] +15w +05b =01w =08w -02b +20w +13b =04b +11w 
:  4 +BelFish           [3018]  6.0  [2795] [   0] +19w =02w +09b +07b -01w =13w +11b =03w =06b 
:  5 +atomGoldbar       [2519]  6.0  [2516] [   0] +27b -03w +12w +21b -07w -08b +15w +22b +17w 
:  6 -WaspX             [2864]  5.5  [2751] [  77] +21w +10b =13w -02w +11b +07b -01w =08b =04w 
:  7 +Goldbar           [2759]  5.5  [2670] [   0] +12w +11w +08b -04w +05b -06w =14b -01b +21w 
:  8 +Marvin            [2699]  5.5  [2658] [ 136] +22w +26b -07w =03b +10w +05w -02b =06w =13b
:  9 +WaDuuttie         [2621]  5.0  [2524] [   0] +24w -01b -04w +14b -13b +17w +20b +12b -02w 
: 10 +atomNightmare     [2560]  5.0  [2527] [   0] +17b -06w -02b +19w -08b +16w +21w +14b -01w 
: 11 +rpiArminius       [2357]  5.0  [2538] [   0] +23w -07b +16w +15b -06w +22b -04w +19b -03b 
: 12 +Barbarossa        [2267]  5.0  [2361] [  39] -07b +23w -05b +25w =14w =19b +24b -09w +22w 
: 13 +Amoeba            [2722]  4.5  [2678] [ 219] +16w +14w =06b -01b +09w =04b -03w -02b =08w 
: 14 +RookieMonster     [2313]  4.5  [2383] [ 223] +28w -13b +22w -09w =12b +15b =07w -10w =16b 
: 15 +Myrddin           [2294]  4.5  [2321] [ 209] -03b +27w +24b -11w +26b -14w -05b +18b =19w 
: 16 +Spartacus         [2253]  4.5  [2282] [ 277] -13b +28w -11b -22w +25b -10b +18w +27b =14w
: 17 +Joker             [2068]  4.5  [2317] [ 780] -10w -21b +28b +18w =19b -09b +27w +20w -05b 
: 18 +RavenEngine       [2500]  4.0  [2056] [   0] -26w -22b +20w -17b +28b +23w -16b -15w +25b 
: 19 +Bliep             [2300]  4.0  [2242] [   0] -04b +25w +26w -10b =17w =12w =23b -11w =15b 
: 20 +Floyd             [2300]  4.0  [2279] [   0] -01b +24w -18b +26w +23w -03b -09w -17b +27w 
: 21 +Nemeton           [2300]  4.0  [2311] [   0] -06b +17w +25b -05w -22b +24w -10b +26w -07b 
: 22 +Isa               [2250]  4.0  [2281] [ 303] -08b +18w -14b +16b +21w -11w +28b -05w -12b 
: 23 +microMax          [1921]  3.5  [2094] [ 458] -11b -12b +27w +24w -20b -18b =19w -25w +28b 
: 24 +Abbess            [2084]  3.0  [2029] [   0] -09b -20b -15w -23b +27w -21b -12w +28w +26b
: 25 +KingSlayer        [2089]  2.5  [2070] [   0] -02b -19b -21w -12b -16w +27b =26w +23b -18w 
: 26 +Skiull            [2011]  2.5  [2076] [ 528] +18b -08w -19b -20b -15w +28w =25b -21b -24w 
: 27 +Micah             [2027]  1.0  [1828] [   0] -05w -15b -23b +28w -24b -25w -17b -16w -20b 
: 28 -Skipper           [1726]  0.0  [1758] [   0] -14b -16b -17w -27b -18w -26b -22w -24b -23w 
:
:     Average Rating    2398.5

[abulmo2]

Thank you for organising the tournament. I watched it with my 7 year old son who was a very enthusiastic supporter of Amoeba
Congratulations for the two winners.

[hgm]

Weird: KingSlayer embarrasingly bungled a drawn KPK ending against Barbarossa. And this despite the fact that it does have some dedicated evaluation code for that end-game, which recognizes the classical drawn positions.

[d]8/8/8/8/3kp3/8/4K3/8 w - - 4 85

Instead of taking opposition here it played Kf2? I suspected that against all odds I must have goofed in programming the draw recognizer.

But I just tested, and cannot reproduce the problem. In this position KingSlayer almost instantly plays Kd2!, after an 28-ply search (its depth limit), with a 2-ply PV leading to a recognized draw. This PV is already found at 2 ply, after searching 28 nodes. Only at 1 ply it prefers Kf2. Even when I go back all the way to the position where Bararossa offers Rook trading, it sees the draw practically immediately (and this is the last move where it takes measurable thinking time), and defends perfectly from then on with instant moves.

In the game it kept insisting it was a draw for several moves after it was lost:

85. Kf2 {-0.52/28} Kd3 {+13.00/20} 86. Ke1 {-0.53/28} Ke3 {+13.15/20} 87. Kd1 {-0.52/28} Kf2 88. Kd2 {+13.19/17} e3+ {+13.07/14} 89. Kd1 {-0.53/28} e2+ {+99.88/13} 90. Kd2 {-10.40/16} e1=Q+ {+99.88/13}

In moves 86 to 89 it keeps insisting on a draw up to the instantly reached 28 ply, even though promotion should have been within the horizon in any decent search. (The score at 88. Kd2 is almost certainly a Barbarossa score that has overwritten KingSlayer's true score/depth, because it was received with lag from the ICS.) This suggests it must have been playing from the hash table. Havingthe engine think about a series of preceding moves, in an attempt to reproduce what was in the hash table, does not reproduce the error, though.

My best guess is that somehow the hash table got corrupted with draw scores when the strong side played a repetion in a winning position. But without being able to reproduce it, this will be hard to investigate.