Moderator: Ras
Correct, the first "50 Elo better" tuning reports have already appeared...gerold wrote:Some will still do it with great results.rvida wrote:But it works like placeboHoudini wrote:Ben, modifying the hex parameters in the executable should have no effect on the engine output.
It would be fun NOT to tell anyone that changing the hex string does nothing.
And then watching reports about successful "settings" with ridiculous elo gains all over the fora...
Now you have spoiled the fun
Sorry Robert,Houdini wrote:Ben,
The "HoudiniConfig" tuning should no longer work with Houdini 3.
In due course (could take a couple of months) I intend to release a special "Tuning" version of Houdini 3 with an officially supported configuration tool for modifying internal evaluation parameters.
Regards,
Robert
Can you give us one good reason why HoudiniConfig isn't supported any more? Can it be that most of this +50 ELO in H3 just comes from some published H2-setting (or at least inspirations out of them)?Houdini wrote:I intend to release a special "Tuning" version of Houdini 3 with an officially supported configuration tool for modifying internal evaluation parameters.
Most probably the case....SchachProfi wrote:Ben, you can simply change eval-parameters runtime when the engine is loaded but currently not calculating (waiting for UCI-commands), you just need a better debugger for it.
Can you give us one good reason why HoudiniConfig isn't supported any more? Can it be that most of this +50 ELO in H3 just comes from some published H2-setting (or at least inspirations out of them)?Houdini wrote:I intend to release a special "Tuning" version of Houdini 3 with an officially supported configuration tool for modifying internal evaluation parameters.
A bug is a bug. There have been times we released a version of Crafty with a bug that we didn't realize existed. While authors try to release a version without bugs, we can only test so many things, and a simple code addition can break something else without realizing it.Dr.Wael Deeb wrote: Most probably the case....
He's a talented programmer,no doubt about that,but.................
He's that kind of person that like to twist the reality mainframe to benefit from the current situation as much as he can.......
Do you even believe that he didn't know about the bug related to the x32 & x64 before the sales of Houdini 3 kicked off
People are not idiots regards,
Dr.D
Only a fool can believe that he didn't do anything but simple tunning to gain the Elo difference in Houdini 3....Peter Skinner wrote:A bug is a bug. There have been times we released a version of Crafty with a bug that we didn't realize existed. While authors try to release a version without bugs, we can only test so many things, and a simple code addition can break something else without realizing it.Dr.Wael Deeb wrote: Most probably the case....
He's a talented programmer,no doubt about that,but.................
He's that kind of person that like to twist the reality mainframe to benefit from the current situation as much as he can.......
Do you even believe that he didn't know about the bug related to the x32 & x64 before the sales of Houdini 3 kicked off
People are not idiots regards,
Dr.D
I've also stated that I believe Robert did more than just tuning to get the increase from version 2 to 3. It's also quite known my opinion of Robert, so if I am defending him, then believe it.
Peter
It is clear that you hate him and I believe that most of the houdini improvement (that seems to be more than 50 elo based on most lists) is not based on changing houdini personalityDr.Wael Deeb wrote:Most probably the case....SchachProfi wrote:Ben, you can simply change eval-parameters runtime when the engine is loaded but currently not calculating (waiting for UCI-commands), you just need a better debugger for it.
Can you give us one good reason why HoudiniConfig isn't supported any more? Can it be that most of this +50 ELO in H3 just comes from some published H2-setting (or at least inspirations out of them)?Houdini wrote:I intend to release a special "Tuning" version of Houdini 3 with an officially supported configuration tool for modifying internal evaluation parameters.
He's a talented programmer,no doubt about that,but.................
He's that kind of person that like to twist the reality mainframe to benefit from the current situation as much as he can.......
Do you even believe that he didn't know about the bug related to the x32 & x64 before the sales of Houdini 3 kicked off
People are not idiots regards,
Dr.D
Robert seems to be the top man improving the top engine in elo. More power to Robert.Dr.Wael Deeb wrote:Only a fool can believe that he didn't do anything but simple tunning to gain the Elo difference in Houdini 3....Peter Skinner wrote:Dr.Wael Deeb wrote: Most probably the case....
He's a talented programmer,no doubt about that,but.................
He's that kind of person that like to twist the reality mainframe to benefit from the current situation as much as he can.......
Do you even believe that he didn't know about the bug related to the x32 & x64 before the sales of Houdini 3 kicked off
People are not idiots regards,
Dr.D
A bug is a bug. There have been times we released a version of Crafty with a bug that we didn't realize existed. While authors try to release a version without bugs, we can only test so many things, and a simple code addition can break something else without realizing it.
I've also stated that I believe Robert did more than just tuning to get the increase from version 2 to 3. It's also quite known my opinion of Robert, so if I am defending him, then believe it.
Peter
Then again,he's the current ultimate opportunist in our elite computer chess community.....
Dr.D
I thought his comment was rather clear:gotogo wrote:Now that is very funny
Robert if the config doesn't work then why would you release a tuning tool? Again I see you bend over backwards to help the advancement of chess.
In due course (could take a couple of months) I intend to release a special "Tuning" version of Houdini 3 with an officially supported configuration tool for modifying internal evaluation parameters.
