@ReneDescartes - I think he quoted you not because it was your idea, but because you articulated it clearly.

Nickajack wrote on 03/31/18 at 04:40:32:

Actually the definition is what the thread is about. Your hippo example is good. Anti-theory can also be theory.

ReneDescartes wrote on 03/07/18 at 21:35:55:


The flaw in this comparison is that we're only comparing the ratings of those who chose to play FRC in the first place. Players disliking FRC would not have competed and would have no such FRC ratings to compare. We would have to require all players to compete in FRC before a fair or exact comparison could be made.

Another problem for some of us in this discussion is that 'opening expertise' is an ill-defined concept. To most it probably means being booked-up, knowing long variations in fashionable openings, or a great variety of lines, etc. But, what if it should mean just having the ability to carry out the first phase of game with skill, regardless of the extent of one's actual memorized knowledge? I suppose a better word to describe such skill would be 'understanding', rather than 'knowledge'.

For instance, what would we call a player who excels at playing a weird opening like the Hippo, or anything else along those lines, with great success? On one hand he is avoiding the opponent's main book prep, while he's employing his strong understanding of opening principles to outplay the opposition on his own turf, so to speak, where he has a lot more familiarity with the possible set-ups. Wouldn't such a player also be deserving of the 'opening expert' classification?

[Such a player also would be likely to play FRC quite well, at least no worse when compared to his regular chess.]

TopNotch wrote on 03/09/18 at 17:03:31:

This third group is the one we used to call "booked-up fish". Actually at one time it was me. I remember the turning-point game very well. Me at 1600 Elo was playing the white side of a Najdorf against a master. He made an inaccuracy at move 10. At move 16 it was +/- ("black does not have enough compensation for the Exchange"), but sadly it was also the end of my book. After that, except for one forced recapture, every one of my "own" moves was a mistake, and at move 22 I resigned because I was losing a rook, when I would be a piece down.

But there is more! In the post-mortem he admitted he didn't know where he went wrong. I showed him the correct 10th move, he said "Oh! Because..." and before I could get another word out of my mouth he played the theoretical variation and gave the assessment "good for black", which is the same thing Geller wrote in the book. Yes, it's true, I was not only booked-up, I was "overbooked". I played a second-rate line for white because I figured my opponent would know the best line and would not know the sideline. And at least in that calculation I was correct.

But there is more! Because even though I knew Geller's assessment, when I had looked at it at home I had had no idea how to continue for black (or white for that matter). I remember actually trying to analyze it for white and giving up because I didn't know what black should be trying to do. (This was before computers, obviously.) So basically I would have been helpless at any point I was out of book, whereas the master was able to re-invent the book on his own.

This one game really drove home to me that I was wasting my time studying sharp openings. Within a few months I revamped my repertoire to stuff like 1.b3 for white, 1.e4 Nc6 and 1.d4 e6 2.c4 Bb4+ for black. Basically I forced myself to learn the middlegame by playing non-theoretical openings where memorization would not help. I played a lot of random junk up until about 2300, at which point I had to knuckle down once again and get a decent repertoire. But in the meantime I was able to devote a lot of time to the middlegame and endgame. And the constant experimentation in the opening allowed me to eventually find the best openings for me, which are nothing like the razor-sharp lines I was playing as a junior.

TalJechin wrote on 03/08/18 at 00:08:39:

Don't drag the match Alpha Zero against Stockfish in this discussion! There are too many loose ends as it was purely an advertising campaign from Google.

To measure the impact of opening study you need to collect data and make some statistical analysis. I did for my article. It makes little sense to discuss this topic any further if you don't approach it more scientifically.

I guess we're talking about different things by now. To me an opening expert is someone who has specialised deeply into opening theory, keeping up with current development in a lot of different openings and analysing one's own novelties and so on - which is quite rare.

While opening knowledge is very common, most 1800 players knew 20 moves in the Ruy Lopez ML already in the 1980s, before the chess book flood. But here in Malmö, except for 2-3 GMs, there's only one player I'd say is an opening expert (and it certainly isn't me, btw). We've played in the same team a few times over the years, so I've seen him win some very quick games - and also some quick draws when the opponents were up to the challenge.

Maybe he only gains 50 Elo points compared to others - I'm not sure how to compare that in a fair way. The example with engines limps a bit, since engines don't really use their books to create their own novelties - maybe Alpha Zero did in a way by playing itself on a supercomputer, and the general opinion seems to be that it's more than 50 points stronger than Stockfish...



brabo wrote on 03/07/18 at 21:44:31:


For one thing, since the middlegames are quite peculiar in FR you could just as well say that C=Y and A+B=0
As someone said, chess is 99% tactics (though that's not mentioned in your simple formula) and the higher your rating the better you calculate...

A+B+C = X
B+C = Y
X = Y
What is A?
A = 0

Replace A with opening knowledge, B with middlegame skills, C with endgame skills, X = standard chess rating, Y = Fischer random rating

It is a very simplified mathematical model but easy to understand.

TalJechin wrote on 03/07/18 at 20:19:23:

brabo's point, as mentioned in an earlier comment, is that the more important opening theory is, the less Fischer Random ratings should correlate with regular ratings, but he observed that the ratings have a high correlation.

an ordinary chessplayer wrote on 03/07/18 at 16:02:27:

True but then you assume that people of the same strength have more or less the same opening knowledge. Reality is different. We do see big differences even when players have exactly the same rating. Therefore if opening knowledge plays a serious role then we should see this in Fischer Random. However it is not.