Originally posted by Kepler
What sort of data would you like? I have the whole record of the session plus notes I made as I went along.

If it's text, and you can post it someplace for us to visit would be interesting.

P-

Originally posted by Phlabibit
You shouldn't be evaluating ANY book/data positions. Each game should start evaluation from the first move out of data.

P-

But what think make it harder in this case is to know what was book in 1965. And one other thing that Berliner strongly hints on the System is that he doesn't use books to choose his moves he does it using his methodology.

Originally posted by Kepler
I know the methodology is flawed. I have said so in the past,
The opening moves were not analysed so are not a factor.

Sometime I'll likely redo the whole thing with better methodolgy but whether I post it here or not is another matter. Pearls before swine and all that.

No, i was not criticizing you in any manner, i was just unclear on this point.

Originally posted by Phlabibit
It would be very tiresome to evaluate all moves of a game to see what is forced.

yes i suppose if you're lazy and terminally dullminded, then of course examining the moves of one game to see which moves are forced and end up being engine match moves by default would surely be a problem of humongous proportions. Surely you jest my man.

Originally posted by adam warlock
But what think make it harder in this case is to know what was book in 1965. And one other thing that Berliner strongly hints on the System is that he doesn't use books to choose his moves he does it using his methodology.

If we're getting this info to compare to engine users of today I don't see how knowing what books WERE used affects anything. This is games played before engines, modern data should be disregarded, and compare results of strong players NOT using engines to strong players of today who may be using engines.

P-

Originally posted by eldragonfly
Says the exgame mod/wannabe who never got off the ground.

What's wrong with my evaluation, or just in the mood to take a run at me for kicks?

If you take out all forced moves, 85% may drop to 82% and the cheat percent might also drop by 3% consistently. You're making extra needless work.

P-

Originally posted by Phlabibit
What's wrong with my evaluation, or just in the mood to take a run at me for kicks?

Too simplistic, dull, arbitrary, basically worthless, not insightful in the least, etc.

Originally posted by Phlabibit
You shouldn't be evaluating ANY book/data positions. Each game should start evaluation from the first move out of data.... even checking a bit forward to be sure it didn't transpose back into known territory.

P-

I started analysing after the book was left. I checked further on for transpositions before starting the serious analysis, which I started from the first move out of book. Unfortunately it is difficult to know what was book back in 1965, especially considering that databases were not in use. My solution to the problem was to install the biggest book I could find for HIARCS and let it decide when the book had been left. In essence, HIARCS will not calculate if its book is turned on and it finds a move for the current position in its book.

Originally posted by eldragonfly
Too simplistic, dull, arbitrary, basically worthless, not insightful in the least, etc.

Perhaps you plan to run 2 games on any given user to determine if they use an engine, so you have time for your needless effort to search the pgn for 'forced moves'.

I'm running a battery of 20 or more games looking for an average. The average I'm used to is based on removal of data only. If I got some strange results (like I said before) I might evaluate that game for forced moves.

P-

Originally posted by adam warlock
Really nice post and effort. got my rec.

But one thing I'd like to say. I was used to think CC more strategical/positional and classic timed OTB chess as more tactical. Just like Berliner says in his book: On CC tactical erros does not abound and then is all about milking the opponent till he's smothered. One other thing is that I think it is pretty ...[text shortened]... inish: if you go check the Estrin vs Berliner game just check out all the kibtzing.

In general where Berliner disagrees with HIARCS (or vice versa) the engine gives 5 or 6 different moves that are all scored as being very close in value. It appears that Berliner is just choosing one of several equally good candidates. In his book he mentions not giving the opponent any targets and watching out for any targets that the opponent provides. This seems to be the way he plays. Certainly, once the opponent has made a mistake and given him a target, Berliner is absolutely relentless in his pursuit of the advantage.

Originally posted by Phlabibit
Perhaps you plan to run 2 games on any given user to determine if they use an engine, so you have time for your needless effort to search the pgn for 'forced moves'.

I'm running a battery of 20 or more games looking for an average. The average I'm used to is based on removal of data only. If I got some strange results (like I said before) I might evaluate that game for forced moves.

P-

Interestingly, I applied the same methodology to one of the games in the "Thread That Was Removed" and then disregarded all the forced or obviously best moves. The match up percentage changed from 92% to 84%. I don't think I would get such an extreme change in any of the games I ran here.

Originally posted by Kepler
Interestingly, I applied the same methodology to one of the games in the "Thread That Was Removed" and then disregarded all the forced or obviously best moves. The match up percentage changed from 92% to 84%. I don't think I would get such an extreme change in any of the games I ran here.

Out of curiosity, how many total moves were 'removed' from evaluation?

P-

Originally posted by Phlabibit
If we're getting this info to compare to engine users of today I don't see how knowing what books WERE used affects anything. This is games played before engines, modern data should be disregarded, and compare results of strong players NOT using engines to strong players of today who may be using engines.

P-

I didn't mean to say we need to know what books were used but that we need was theory on their days. I'm thinking like this: imagine that in 100 years someone decides to compare that era engines with Berloners play and decide not to analyse what's theory on their days and just analyse nonbook moves. Quite possibly they'll only have to analyse some 2 or 3 moves from a 30 move games. Am I making sense? What I'm trying to say is since chess theory evolves and expands taking into accoun present day knowledge wouldn't seem very accurate to me when is making game analysis of past player's games.

One other thing I thought. What about shutting off the engine's opening book and analyse all moves? Would that be necessary? Would this be too time consuming? I'm not talking about analysing the games from people on this site like this but only doing this when analysing some past games were theory was clearly behind from what we have today.

Originally posted by Kepler
Certainly, once the opponent has made a mistake and given him a target, Berliner is absolutely relentless in his pursuit of the advantage.

Yeah, on his wins after he has some kind of advantage, as tiny as it migth be, it is just blow after blow. And either his opponents crack up or end getting a tied up endgame. He made switch to 1. d4.

Originally posted by Phlabibit
Out of curiosity, how many total moves were 'removed' from evaluation?

P-

As I recall, there were about 50 matches out of about 54 or 55 non-book moves. I removed another 5 as forced or obviously best. Altogether there were over 70 moves, 20 of which were book.