Merciless and Brutal!

[geots]

AMD Phenom(tm) II X6 1090T Processor x6 @ 3.20 GHz - 8.00MB Memory
Windows 7 Home Premium Edition (Build 7600)
Fritz Benchmark:
Speed: 23.33
KNS: 11196
GUI: CB Rybka 3
Book: Perfect 2009- 10 moves
Hash: 256
TB : ON
Ponder: OFF
Blitz:5' 0


Code:

1 Houdini 2.0c Pro x64 3195 +75/-24/=51 67.00% 100.5/150
2 Deep Junior 12.5.0.3 UCI 3110 +24/-75/=51 33.00% 49.5/150

Houdini: 1 CPU

Deep Junior: 6 CPU


AMD x6 W7 Games.rar

Here is one more from the same chess forum.

Miodrag- all the people here who would rather slit their throat than believe what we are showing them- they are all very fine people- very intelligent people. They just have no conception of these engines being this good. They know not what they are dealing with. It's scary- these are such monsters.

So please have patience with my friends here- and I know you have been so far. This is just a bitter pill for them to swallow.



Best,

george

Hi George,
you are great.Loved your posts and your advice.
Yes , I am patient and I agree with you that those people are very fine and intelligent people.
Their problem is that for two years since Ippolit and it's family including Houdini showed up they had constantly to swallow the bitter pills and don't want to except the fact that these engines are like supersonic jets while the engines like Fritz, Shredder and Junior are biplanes trying to go supersonic.That can be done only in their dreams.
Look at their reactions.When they don't have the arguments they start immediately with the insults.Just look at the post from H.G. Muller.
If they don't insult than there are excuses.Loved your great response to those excuses.
You are right that it is scary how strong and good these engines are.I have been using them from the time Robbolito, IvanHoe and Houdini were released and had the chess time of my life.
If they don't want to use them that is their problem but to call the engine that is loosing against Houdini 103-7 and 40 draws World champion is really funny. On the other side they don't want to have the tournament where only the strongest engines would participate independently of the names they are being called.
I am glad that you are also having great time with these engines.
I am trying to stay away from discussions on this forum because so much fertilizer is being thrown here that I don't want to get dirty.
I hope that you have some kind of protection.
Cao !

One last thing, Mio. My email address is geotsp8@yahoo.com Please email me and give me yours so I can add it to my address book. Then we can talk privately.

And look, I dont want to be misunderstood. If Fritz 13 had won- I would have posted that real quick as well. The way I like to do things, when I am told one program is not that strong- why argue. I go and run them and find out for myself. Im not much with computers compared to these highly proficient guys here. But I can tell you one thing. I know how to test and get true results. When in a 100 game match- IvanHoe , in wins and losses- beats Fritz 13: 40 - 18- anyone who tells you that is not enough games to get a true reading is nuts. Now it may be correct that Fritz might make a better showing along the line- but I can promise you from watching that if you continued from here and played to 50,000 games- Fritz would NEVER get within 10 games of IvanHoe. And this is not even considered a strong IvanHoe version. Hell, it's not even a version. It's one of those CM personality deals- per se.

You know why I do this. Not to embarrass anyone. I actually thought it might be close matches because I picked one of IvanHoe's weaker engines. I do it because I have been told a thousand times that this is just a hobby. And we should do what we enjoy. And that is all I am doing. It is most fun when you are running matches that involves an engine you are not quite sure about- but want to find out. I cannot imagine anyone begrudging that.


Best,

George/ and send me your address

[hgm]

Look at their reactions.When they don't have the arguments they start immediately with the insults.Just look at the post from H.G. Muller.

Insults? I would think 'stupid' was actually a very mild and polite way to put it. How would you call people that do not even understand what a certain championship is for?

My uncle just bought a number of Formula I racing cars, put my cousins in it that had just obtained their driving licences, and had them race a few laps against his Porsche, which he was driving himself. And guess what, he beat them by multiple laps. So he is now boasting that Formula I racing are a sham, these cars not being good enough to match up to a simple Porsche.

Now is he stupid or not?

I actually think he is stupid beyond belief...

I wonder if you can see why...

[solis]

I could try to go down to your level and call you the idiot but I wont .

[hgm]

Very wise, because randomly calling people idiots would not make you look very smart. You would have to give a motivation to prove they are idiots, and I am pretty sure that would be tad difficult for you.

Nevertheless, now you have introduced the word 'idiot' in this discussion, it is of course a good question whether people thinking that a contest to reveal the best programmer / car driver is a contest to reveal the strongest blitz chess program / fastest car on a straight road should qualify as idiots, or that they are merely moronically stupid...

Of course you might wat to argue they are very smart, right? There is no accounting for standards...

[Uri Blass]

AMD Phenom(tm) II X6 1090T Processor x6 @3.80 GHz 8.00 MB Memory
Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
Fritz Benchmark:
Speed: 28.33
KNS: 13599
GUI: CB Rybka 3
Book: Perfect 2009-10 moves
Hash: 256
TB: ON
Ponder: OFF
Blitz:5' 0


Code:

1 RobboLito 0.09 x64 3100 +28/-12/=20 63.33% 38.0/60
2 Deep Junior 12.5.0.3 UCI 3100 +12/-28/=20 36.67% 22.0/60

Robbolito; 1 core

Deep Junior: 6 cores

AMD.x6 Testing.rar




A good old fashioned ass whipping. (Who is the World Champion)

The version of Junior that played in the world championship was not the commercial version of Junior and one of things that Shay bushinsky said is that the programmers of Junior made improvements.

I can add that you did not use the special book of Junior and
in world championship Junior use book that help it to go for positions that it plays relatively better so it can win against engines that are better with
Perfect 2009-10 moves book.

The bottom line is that the programmers of Junior believe that Junior have chances also against the strong engines that did not play in tournament conditions like WCCC.

They are not going to refuse to play Houdini or Rybka in these conditions
if you find a sponsor for that event.

Uri- let's end this one quick:


1. Use the version that was in WCCC.
2. Use whatever book you like
3. Use whatever time control you like.
4. Let Junior run with learning = ON, tho that wont matter
5. Give Junior as many cores as you like- but no more than 8 cores.
6.Have the match played wherever you like.
7. Just tell RobboLito when and where, and he will be waiting with 1core when you show up.
8. Get it on- then read em' and weep.


george

1)I cannot use the version that was in WCCC because I do not have it.
2)I believe that the version that the Junior team used in WCCC used more than 8 cores.
3)I did not claim something about the level of Junior that won WCCC and I do not know if you are right or wrong in claiming that Robbolito one core beat the private version of Junior on 8 cores when Junior use its own book but only that the results that you posted do not prove your claim.

[Milos]

1)I cannot use the version that was in WCCC because I do not have it.
2)I believe that the version that the Junior team used in WCCC used more than 8 cores.
3)I did not claim something about the level of Junior that won WCCC and I do not know if you are right or wrong in claiming that Robbolito one core beat the private version of Junior on 8 cores when Junior use its own book but only that the results that you posted do not prove your claim.

There are possible, probable and (almost) impossible things in life.
WCCC Junior version on 8 cores with own book beating Robbo (with own book also) is in possible but improbable category.
The difference here is around 100 Elo, lets take a lower bound of 80 Elo only.
2 more cores and faster machine at this level of scaling could account for maximum 20 Elo. The book can account even for 50 Elo, however, there are also good books for Robbo/Ippo, so this would not make more than 20 Elo difference. LTC vs. 5'/game could only provide 20 Elo even though I strongly doubt that since no one ever presented any valid statistical proof that for any 2 strong (over 3000 Elo) engines the difference between 5'/game and 2h/game can be more than 20 Elo (since it is 20 Elo difference it would require at least 1000 games match, and I'm not talking here about rating list but individual match statistics). I repeat no one ever!
This leaves us with 20 Elo stronger private version of Junior (and that's 20 Elo vs. Ippo family engine not weak engines from WCCC it was tuned for).
So it is possible but not so probable (have in mind that I took all worst case assumptions for Robbo).
It would be good that for a change once you refute someones claims also present your own valid arguments instead of the usual blah, blah story.

[hgm]

It is common practice, even automatic, to say Junior when it is actually Amir Ban whose name might even be unknown to those using or talking about his engine. Same with a lot of other engines. Everyone knows that there is a man behind every engine, no need to call them (even me) stupid just because of a substitution.

Indeed, that is not what makes them stupid. What makes them stupid is that they use the fact that some other version of that program (under entirely different circumstances) is beaten by another program as an excuse to make derogative remarks about the people behind it and the quality of the championship in general. It shows as much lack of understanding of the purpose of the tournament as calling marathon runners second-rate athletes because they go slower than bicycle (and actually thought that such a brilliant idea you actually went through the trouble of trying it out)...

[hgm]

There are possible, probable and (almost) impossible things in life.
WCCC Junior version on 8 cores with own book beating Robbo (with own book also) is in possible but improbable category.

That is what Vas thought too, when he tricked FIDE into awarding $100,000 for a match against an opponent he had a 150-Elo lead on. But Zappa won...

[geots]

AMD Phenom(tm) II X6 1090T Processor x6 @3.80 GHz 8.00 MB Memory
Windows XP 64 Bit Professional Service Pack 2 (Build 3790)
Fritz Benchmark:
Speed: 28.33
KNS: 13599
GUI: CB Rybka 3
Book: Perfect 2009-10 moves
Hash: 256
TB: ON
Ponder: OFF
Blitz:5' 0


Code:

1 RobboLito 0.09 x64 3100 +28/-12/=20 63.33% 38.0/60
2 Deep Junior 12.5.0.3 UCI 3100 +12/-28/=20 36.67% 22.0/60

Robbolito; 1 core

Deep Junior: 6 cores

AMD.x6 Testing.rar




A good old fashioned ass whipping. (Who is the World Champion)

The version of Junior that played in the world championship was not the commercial version of Junior and one of things that Shay bushinsky said is that the programmers of Junior made improvements.

I can add that you did not use the special book of Junior and
in world championship Junior use book that help it to go for positions that it plays relatively better so it can win against engines that are better with
Perfect 2009-10 moves book.

The bottom line is that the programmers of Junior believe that Junior have chances also against the strong engines that did not play in tournament conditions like WCCC.

They are not going to refuse to play Houdini or Rybka in these conditions
if you find a sponsor for that event.

Uri- let's end this one quick:


1. Use the version that was in WCCC.
2. Use whatever book you like
3. Use whatever time control you like.
4. Let Junior run with learning = ON, tho that wont matter
5. Give Junior as many cores as you like- but no more than 8 cores.
6.Have the match played wherever you like.
7. Just tell RobboLito when and where, and he will be waiting with 1core when you show up.
8. Get it on- then read em' and weep.


george

1)I cannot use the version that was in WCCC because I do not have it.
2)I believe that the version that the Junior team used in WCCC used more than 8 cores.
3)I did not claim something about the level of Junior that won WCCC and I do not know if you are right or wrong in claiming that Robbolito one core beat the private version of Junior on 8 cores when Junior use its own book but only that the results that you posted do not prove your claim.

Uri, you can dig back in threads till hell freezes over, and you will not find where I said RobboLito beat the private Junior version on 8 cores.

It was stated that Robbo on 1 core kicked the shit out of the latest and strongest commercial Junior- with Junior using 6 cores. I saw the results and the pgns.

Number 1- Try to stick to the truth.

Number 2- You mentioned the private version in WCCC- not me. That was a convenient excuse because you knew when you said it you had no access to it.

Number 3- This aint the first rodeo I've been to. I know nothing about the private version- but I can tell you what I do know. I got vaccinated good when a "private version" of Shredder won the WCCC way back when, and I suckered myself into buying the "World Championship Program", which turned out to be an astounding 8 elo stronger than the previous version I had also bought.

Number 4. It matters one hell of a lot what strength a program is that wins the WCCC- because he will slap "the same world champoinship version that won WCCC" on the next version he sells. Now I have been bit twice- so I don't fall for that shit. But what about the hundreds of people who don't know rating lists from dump trucks. Is this fair to them?

Number 5. Back up to what I did not say. I did not say RobboLito on 1 core beat the private version of Junior using his own book and 8 cores. What I DID SAY was that Robbo on 1 core, giving the private version of Junior its own book and 2 extra cores, equaling 8- and you pick the time limit:

I have money-how much?- enough to match whatever you come up with- that says Robbo 1 core will stomp the living shit out of Junior and his whatever version on 8 cores. Did not say it had, only that if it took place it WOULD happen.

Number 6. there is no number 6. I am thru trying to discuss this subject with people who seem to relish living in the past. The future is staring squarely in your face- and instead of realizing it and trying to compensate by getting back to programming so you have a chance to be the ass kicker instead of the receiver of the kicks- you would rather spend your time making excuses for why the capitalists- except for Richard and Tord- can't cut it ag. this level of competition. I would advise to quit making excuses for outdated ideas and get back to work.


I'm outta here- you guys feel free to come up with excuses- call me a liar, and generally just stew and bitch.

bye

[Milos]

That is what Vas thought too, when he tricked FIDE into awarding $100,000 for a match against an opponent he had a 150-Elo lead on. But Zappa won...

Personally I believed that you had at least a basic understanding of statistical significance of match results. You know error bars and stuff.
Unfortunately you seams not to be better in understanding simple things than any layman basement tester... Pretty sad stuff actually .
95% significance error bars in 10 games of that match were 22.5% which is 170 Elo. If you'd just be able to understand this elementary school math... But I guess I'm asking too much.