M ANSARI wrote:And what agenda would that be? If it is to note that all the Ippolit based engines are derived from Rybka 3 and not an original work, then you are of course correct. I would hardly say that finding the truth is "an agenda".
Once again you are proven wrong. But you seem to get upset when you make unsubstantiated claims, and someone comes with data that contradicts your account. Why are you surprised that someone would actually check a statement you make and try to validate it?!
A few days ago, I pointed out to him that he makes too many wild guesses, presenting them as facts.
He replied that he has his head in his ass, or something similar.
M ANSARI wrote:One more thing ... the reason I realized that the position that you posted will be solved by RL and would simply need more time is because when I looked at the position it was similar to many positions where I mentioned that RL is missing code that makes it play very poorly in some positions. Telling you what it is would simply make the cloners life much easier. So why don't you look at the position and try to figure out why RL takes so much time and R3 does not.
Why do you pretend that you know something about the RL and specially R3 code, when you have no clue at all?
You are the same guy that claimed R3 has its binary encrypted. For most ppl at this site this is sufficient to never take seriously anything you say...
Where did I make the claim that I know anything about code in either RL or R3. I wouldn't have a clue about a binary if it hit me in the face. My observation was that RL was somehow searching faster and thus playing stronger. That the advantage against R3 was due to faster search and not a better evaluation. That this speed advantage COULD be due to some obfuscation or encryption used that might slow down the program. I also mentioned that this speed could be due to removal of a lot of knowledge code that might not be relevant in fast interval games. I did that observation not by studying the code of RL or R3, but by analyzing the games played by these 2 engines, and by trying to figure out why one engine plays faster but with almost identical evaluation. Strelka was cloned from older Rybka, so it would not take a lot of imagination to guess that Vas would try to make some scheme to protect his code from cloning again. I know that there are obfuscation competitions where people try to obfuscate code the best to win. I did also mention that although Vas agrees with the removal of some code of R3 to make it run faster, he never gave me an answer regarding anti cloning methods.
Again why are we straying from the subject above. You know the part about data showing that R3 and RL share the same weaknesses. Did you even go through the post?
M ANSARI wrote:If you are drawing a blank I suggest instead of doing silly statistics that you look into the actual chess being played.
Engines whose games you are "looking" have more than 3200 ELO on your hardware. No man on planet can take a win against them in even odds game. Most except few ppl on the planet have no clue for 99.9% of the time why engine is playing exactly what is playing.
And you base your assumptions on those completely statistically irrelevant 0.1% of the time when you actually have clue what and why engine played.
Are you serious ??? Do you really think that a human would have no clue what is going on? Who do you think tuned those evaluations? An engine might play like a 3200 ELO player (which so far does not exist in the human world) but it is not quite so simple. The ELO rating is inflated due to the extemely high ability of chess engines in tactics, probably around 3800 ELO. But engine play in static positional positions is probably less than 2400 ELO. So if you think that a human cannot see or understand a chess engine game, you have a LOT to learn. Go in the engine room on Playchess and have a chat with some of the more experienced players there. As you are chatting and both your engines are playing, he will be able to tell you that your +5.00 evaluation is bogus and that the position will end up being drawn. He will be able to tell you that the -.30 evaluation that you are seeing will end up being enough to win the game. They KNOW the weaknesses of their engines and they look at thousands of games where tuning an opening line might take advantage of a known weakness. You obviously have no clue what you are talking about. I suggest you stick to computer chess code and leave the actual chess playing part to others.
M ANSARI wrote:And what agenda would that be? If it is to note that all the Ippolit based engines are derived from Rybka 3 and not an original work, then you are of course correct. I would hardly say that finding the truth is "an agenda".
Once again you are proven wrong. But you seem to get upset when you make unsubstantiated claims, and someone comes with data that contradicts your account. Why are you surprised that someone would actually check a statement you make and try to validate it?!
A few days ago, I pointed out to him that he makes too many wild guesses, presenting them as facts.
He replied that he has his head in his ass, or something similar.
.
You obviously did not read the post correctly. I mentioned that YOU had your head up your ass because you obviously were not looking at what was written. It seems that you suffer from the same till now. Do you have anything useful to offer in this thread?
