What is New in Toga II 1.3x4?

[mclane]

if the new version always gets more points than the old version in a tournament, its quite obvious that the new version is better. you don't even have to understand chess to find this out. a look in the ranking-table shows this.

[Terry McCracken]

i have not been able to set up a tournament where toga1.2.1a and 1.3x4 participates where the new version is not ahead of the old version with at least 2 or 3 points and even more.

its not possible !

maybe something is broken with the way you test (and with the way the elo-list guys test).

Maybe it's you who has the problem how to test, not the rest of the testers?

Has that ever occurred to you?

[Uri Blass]

i have not been able to set up a tournament where toga1.2.1a and 1.3x4 participates where the new version is not ahead of the old version with at least 2 or 3 points and even more.

its not possible !

maybe something is broken with the way you test (and with the way the elo-list guys test).

Maybe it's you who has the problem how to test, not the rest of the testers?

Has that ever occurred to you?

Another possibility is that there is a difference between tests with ponder off and tests with ponder on.

The rating lists that I know are based on ponder off games.

Uri

[mclane]

exactly. even on my SLOW laptop AMD2400+ toga is doing a lot faster than on terry's machine. so IMO the results are strange enough.

btw. blueberry is weaker than default toga.

[Terry McCracken]

i have not been able to set up a tournament where toga1.2.1a and 1.3x4 participates where the new version is not ahead of the old version with at least 2 or 3 points and even more.

its not possible !

maybe something is broken with the way you test (and with the way the elo-list guys test).

Maybe it's you who has the problem how to test, not the rest of the testers?

Has that ever occurred to you?

Another possibility is that there is a difference between tests with ponder off and tests with ponder on.

The rating lists that I know are based on ponder off games.

Uri

I test with ponder off.

[mclane]

no - i have enough machines at home to make sure that what i say is true.

if toga 1.3x4 is weaker, how do you explain those results:
http://216.25.93.108/forum/viewtopic.ph ... 5&start=20

[mclane]

i have 5 or 6 or whatever machines and do test ponder on or off, depending on what i want to test.

but in ALL tournaments toga 1.3x4 is better than 1.2.1a .

in the same way in all tournaments Gambit-tiger is better than normal tiger.
with ~ the same amount of points.

see e.g. here
http://f50.parsimony.net/forum200336/messages/8068.htm
http://f50.parsimony.net/forum200336/messages/8063.htm
http://f50.parsimony.net/forum200336/messages/8042.htm
http://f50.parsimony.net/forum200336/messages/8012.htm
http://f50.parsimony.net/forum200336/messages/8013.htm
http://f50.parsimony.net/forum200336/messages/8009.htm
http://f50.parsimony.net/forum200336/messages/8018.htm

[Dr.Wael Deeb]

no - i have enough machines at home to make sure that what i say is true.

if toga 1.3x4 is weaker, how do you explain those results:
http://216.25.93.108/forum/viewtopic.ph ... 5&start=20

I personaly think that there's something definitely wrong with Pavel's test,he change the opening books from time to time,an odd opening books like 1.d4 e5? and so on....
I also think that his hardware configuration is broken.....

[Tony Thomas]

i have 5 or 6 or whatever machines and do test ponder on or off, depending on what i want to test.

but in ALL tournaments toga 1.3x4 is better than 1.2.1a .

in the same way in all tournaments Gambit-tiger is better than normal tiger.
with ~ the same amount of points.

see e.g. here
http://f50.parsimony.net/forum200336/messages/8068.htm
http://f50.parsimony.net/forum200336/messages/8063.htm
http://f50.parsimony.net/forum200336/messages/8042.htm
http://f50.parsimony.net/forum200336/messages/8012.htm
http://f50.parsimony.net/forum200336/messages/8013.htm
http://f50.parsimony.net/forum200336/messages/8009.htm
http://f50.parsimony.net/forum200336/messages/8018.htm

Adding up all the games, you still got less games for each engine than me.

[mclane]

of course

http://f50.parsimony.net/forum200336/messages/8014.htm