What the computer chess community needs to decide

[lkaufman]

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

LMR at PV nodes is -5 Elo. Is Don really so clueless?
And you believe what Chris writes???

Regarding LMR at PV nodes, Critter reported a huge gain from it, SF uses it and gained points when it was increased recently, and Komodo uses it and benefits from it. I don't know if it helps the Ippo family, we don't work on that code. But Houdini did get a big speedup at the same time as Ivanhoe introduced it. Coincidence? Seems unlikely. As for what Chris writes, I have no reason not to believe him. Aside from the question of the derivation of Ippo, do you dispute anything else he said?

[Christopher Conkie]

I never said I had an engine or ever that did try to create (or adapt) one, did I?
Seems you are running out of arguments.

Arguments? What arguments? I was just answering your question.

What would you expect me to argue about? It was allready established with absolute certainty about 200 postings ago that only a complete asshole would think all engine authors copy code, rather than writing it from scratch. So what is left to argue?

The best post of the day. I must say I did laugh out loud.

There is no substitute for quality. What a super-dupa post.



Chris

Oh so funny, indeed one the best posts ever. LOLLLL

I rest my case so you have more time to play the chess engine authors with your silly engines all written from scratch.
Good luck.

"Protestations of impartiality I shall make none. They are always useless and are besides perfect nonsense, when used by a news-monger."

-William Cobbett, (In the first issue of Porcupine's Gazette, 4 Mar,'Address to the public'.)

2nd best post methinks......



Chris

[Don]

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

LMR at PV nodes is -5 Elo. Is Don really so clueless?
And you believe what Chris writes???

For us LMR at PV nodes is significant, perhaps 20 - 30 ELO but I don't remember off the top of my head. We have tried several times to turn it off and try for more aggressive LMR and other things in non-pv nodes but in each case we have failed. LMR at PV nodes is clearly a good thing.

Ivanhoe is a really good program, it would only take a few changes such as this to explain the modest improvement or the first houdini over Ivanhoe.

[Houdini]

Regarding LMR at PV nodes, Critter reported a huge gain from it, SF uses it and gained points when it was increased recently, and Komodo uses it and benefits from it. I don't know if it helps the Ippo family, we don't work on that code. But Houdini did get a big speedup at the same time as Ivanhoe introduced it. Coincidence? Seems unlikely. As for what Chris writes, I have no reason not to believe him. Aside from the question of the derivation of Ippo, do you dispute anything else he said?

Larry, the speculation about Houdini from yourself and Don is out of line and a complete disgrace.

The use of LMR in PV nodes in Houdini pre-dates the appearance of the same feature in Ivanhoe, as can easily by verified by the disassembly wiz kids of the block.
This feature was not introduced in Houdini 1.5.

Robert

[Adam Hair]

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

LMR at PV nodes is -5 Elo. Is Don really so clueless?
And you believe what Chris writes???

At this point, do you think that anybody believes what you write?

[Milos]

For us LMR at PV nodes is significant, perhaps 20 - 30 ELO but I don't remember off the top of my head. We have tried several times to turn it off and try for more aggressive LMR and other things in non-pv nodes but in each case we have failed. LMR at PV nodes is clearly a good thing.
Ivanhoe is a really good program, it would only take a few changes such as this to explain the modest improvement or the first houdini over Ivanhoe.

LMR at PV nodes as is in Ivanhoe is -5 Elo. With little better implementation you can gain +10 Elo. But that's it.
You can't gain 50-60 Elo by bugfixes. As a matter of fact there are no negative Elo bugs in Ivahoe...
And a speed-up in terms of NPS of Houdini vs. Ivanhoe is less than 10% in x64, meaning less than 10 Elo...

[Adam Hair]

Ivanhoe smp was more buggy than a termite hill.

Chris

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

Everything in Ivanhoe was copied into Houdini progressively.

As Ivanhoe got something.....so did Houdini.

Houdini is nothing more that the dull non-swearing face of Ivanhoe,

Think of it as "their marketing".



Chris

Are you sure about which direction the ideas were being copied?
There is something I have noticed, but I have been uncertain about
whether Ivanhoe was copying Houdini or if Houdini was copying
Ivanhoe.

[Houdini]

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

I personally believe it is due to bug fixes, and some work on evaluation, mostly endgame stuff. However it's likely there were a lot of other minor changes.

In any program there are about 100 changes you COULD make that have very little consequence one way or the other, and I'm sure that would have been done too. That would be one way to "customize" the program and make it seem like it was your own.

Don, the idle speculation about Houdini from yourself and Larry is out of line and a complete disgrace.
Please stop this nonsense.

Robert

[Christopher Conkie]

Ivanhoe smp was more buggy than a termite hill.

Chris

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

Everything in Ivanhoe was copied into Houdini progressively.

As Ivanhoe got something.....so did Houdini.

Houdini is nothing more that the dull non-swearing face of Ivanhoe,

Think of it as "their marketing".



Chris

Are you sure about which direction the ideas were being copied?
There is something I have noticed, but I have been uncertain about
whether Ivanhoe was copying Houdini or if Houdini was copying
Ivanhoe.

The direction is Ivanhoe into Houdini. The reason why I say this is the timelines. If you look closely you will discover that one Ippolit derivative gets replaced by another. For all of these authors not one continues when they are "superseded" so to speak.

I and others I know who that are like-minded are looking for anything new. We search for the otherwise undiscovered (at no matter what level). It is because of this that we can see this syndrome develop. We help with a GUI. A GUI needs engines. I suppose you could say that we watch things as they "develop".

At this current moment.....the most palatable Ippolit that has been conjured (both engine and person) is Houdini.

Of course you will not see any of these "authors" at an official tournament. The reason for that is they will never be able to explain "their" creation to the other programmers there. Conversely, that is also the reason why they will never be allowed in to those places.

We can put dates to everything. Exact dates. We are watching you see.

Hope this explains it.

May I go now?

Gotta go tae the dancin'...... So many women so little time......



Chris

[tomgdrums]

Were the bugs in Ivanhoe ones that would affect the Elo rating on a single core? I wonder whether bug fixes alone could account for the higher Houdini rating. Also, the rating jump in Houdini 1.5 corresponds in time to the addition of LMR at PV nodes in Ivanhoe. Assuming this was copied into the latest Houdini, this could account for the large speedup and much of the Elo gain. What do you think?

I personally believe it is due to bug fixes, and some work on evaluation, mostly endgame stuff. However it's likely there were a lot of other minor changes.

In any program there are about 100 changes you COULD make that have very little consequence one way or the other, and I'm sure that would have been done too. That would be one way to "customize" the program and make it seem like it was your own.

Don, the idle speculation about Houdini from yourself and Larry is out of line and a complete disgrace.
Please stop this nonsense.

Robert

Why is it out of line?