Protector 1.4.0 & Protector 1.5.0: RESULTS Are In!

[geots]

Each person can derive from it what he chooses, but it is not often results with really an untested version fall right into place as they were purported to be- with 50 games, 500 games or 5000 games. Except for slightly less in the elo figures- someone was "dead-on" accurate. Does it mean this would hold up under more testing- you be the judge. I keep getting reminded I am "just a tester."



Below conditions identical for both matches.




Inspiron 620 Intel i5-4 True Cores
Fritz 11 gui/Fritz 13 gui
1CPU/64bit
128MB hash
Bases=NONE
Ponder_Learning=OFF
Perfect 2012b.ctg w/12-move limit
4'+2"
Match=50 games



Protector 1.4.0:

Code: Select all

SP8-Core-i5, Blitz:4'+2"  2/9/2013  0

                         
Deep Sjeng c't 2010   +28    +17/-13/=20   54.00%   27.0/50
Protector 1.4.0 x64   -28    +13/-17/=20   46.00%   23.0/50

Protector 1.5.0:

Code: Select all

SP8-Core-i5, Blitz:4'+2"  2/9/2013  0

                         
Protector 1.5.0 x64   +28    +15/-11/=24   54.00%   27.0/50
Deep Sjeng c't 2010   -28    +11/-15/=24   46.00%   23.0/50

Note the scores were identical- of course which meant a 28x2 swing in elo.
Very interesting at any rate.




On my way to "Tornado Alley"- Toto calls,

george

[geots]

BTW, I don't know if it makes any difference or not, but he somehow managed to get what I would call a substantial increase in NPS.


george

[velmarin]

Yes,
But the credit goes to the author,
I've just compiled,
Even when compiling, the computer froze previous version, now normally compiled and loaded normally.

Merit of the author.

[geots]

Yes,
But the credit goes to the author,
I've just compiled,
Even when compiling, the computer froze previous version, now normally compiled and loaded normally.

Merit of the author.

Still, I may need to start calling you "the magician". You do what they say can't be done quite often.


Best,

george