wat opening theories would you recommend for someone who doesn't want to learn a lot of opening theories? Because that is what I am searching for

I think my post right before yours contains some examples of openings that avoid heaps of theory but don't concede much.

Other examples for White (with e4):

Against the Sicilian:

Alapin, Smith-Morra, Grand-Prix (though the GP gets slaughtered against 2200+)

Against e5:

Vienna, Exchange Variation of the Spanish, Center Game

Against the French:

Tarrasch (some theory but not as much as 2. Nc3)

Against the Kahn:

Lines with 6. Ne2

Against the Pirc:

I forget the name, but it involves a quick Bc4 and Qe2 with the idea of an early e5

As Black:

Against e4:

Petroff if you don't mind drawing as Black

Against d4:

I don't know, d4 is the hardest to play an offbeat opening against.

thx, what about for black?

I think a 2300 probably still beats an 1800 almost every time even if his opening is 1. Nf3 .. 2. Ng1 .. 3. Nf3. If you just play logical moves then there's no need for theory until you reach quite a high level of play! It's easy to blaim the opening when you get crushed in 15 moves or whatever - but I think more often it was just basic tactical/positional blunders that got you, and not a misevaluation of some subtle theoretic opening position.

I do alright in the botvinnik variation of the semislav (I think this is probably the most theoretic opening there is?) against some pretty decent opponents with little more theory than: play g2/Bg2, d5 is often a good way to respond to c5 and black can't usually afford to capture b2 because of the weakness of his king.

Also, I think everybody thinks everybody else knows lines much better than they do.

Another benefit to playing the mainstream theoretical openings is that it works great as a bluff. I've just recently started wandering into every main line I can usually half blind, and often have opponents that [apparently] play inferior moves just to get ME out of book. And all it does is give us a position where both of us are out of book, but I have a better position.

If you're specifically talking about giving other pieces the one-time option of moving like a knight, I think you're overestimating the difficulty computers would have with this. Keeping track of which pieces have used their N-moves is a matter of a single bit, and generating a knight move for each piece can be done with a single boolean operation on stored bitboards.

The big effect would be the large increase in move options at each ply (half-move), but even this isn't as nearly as big an effect as we might think at first glance. An average chess position has around 38 legal moves for the side to move. Simply generating moves for each side leads to a tree with 38^n nodes, where n is a half move. This gets huge very quickly, so that looking 3 full moves ahead would seem to involve examining 38^6 ~ 3 billion positions.

It's not that simple, though. The overwhelming majority of those positions are eliminated by pruning techniques and never evaluated at all, with zero effect on the outcome. These tree-pruning and searching techniques have effectively lowered the 38 to something like 6, so that 3 full moves might require looking at fewer than 50 thousand positions rather than 3 billion.

The same effect would be there for any extra moves introduced by new rules. Rather than adding 25 new move possibilities to the 38 already there, you might be adding 3 or 4 to the 6 that are there. Still a very large increase, but maybe not as exponentially explosive as humans might have hoped.

Under current chess rules, a 6-fold increase in computing power adds a half-move to a chess program's lookahead capability. Under your proposed rule change, the computer might need a 10-fold increase, instead. This might be enough to allow a handful of humans to catch up to computers...for a little while. But to what end? The race is over. The programmers and engineers won, as most people expected.

Caveat: The numbers I used should be fairly accurate, but they're from half-forgotten memory from when I was interested in all this. Still, the general argument is valid, I think.

About Sklansky's variant where the rook can move like a knight once per game - there's already a move in the rules of chess that has similar characteristics. It's called castling. Each side can do this atypical move once per game. So a chess engine deals with that by assigning value to retaining the option to castle. Similarly a chess engine would assign value to retaining the option to do the special rook/knight move. I don't see why this would give humans any special advantage.

Not necessarily. Computers are still significantly weaker than the best humans in shogi, which has drops.

The first suggestion (setting up your pieces as your first eight moves) has been tried, and is called "pre-chess." I don't know if it has been tried with computers, though probably it has.

The GrandPrix is not so bad, and is almost certainly at least equal for white. Of course this isn't anything to write home about but implying that it gets killed against someone over 2200 is a bit crazy. One example is Adams, who played it a decent amount just a few years ago and did fine with it. It's a perfectly fine opening IMO, I'm just not a big fan personally.

Might that be due to lack of development? I have no idea about Shogi, so I'm just wikiing it but I found this interesting article: http://en.wikipedia.org/wiki/Computer_shogi

It mentions that the Japanese Shogi Association is preventing professionals from publically playing computers without permission and restricting any fees they would receive from playing computers. And the only time they gave permission for a pro to play a computer publically was 4 years ago, when it was the world champion versus a computer rated as an amateur - and the computer had already back then ended up with winning chances!

Not exactly the most appealing atmosphere to try to develop software for. Maybe learning from the mistakes of chess.

The castling analogy is bad because it goes away quickly and it can not usually be used with devastating effect.

The reason why a one time option to use a rook like a knight or whatever can't be easily programmed results from the fact that you should often turn down the option because it is likely to be even more valuable later.

Obviously coming up with the best version of my idea should not be done lightly. Serious thought by experts is needed.

Right, knowing when to use the move is a judgment thing that humans would be better at than computers.

Perhaps even more interesting would be the option to, at any one point in the game, make two moves in a row as long as neither is a check or capture. This would but a large premium on the initiative and also make knight outposts a huge part of the game. (Because if there were, say, pawns on b4, c5, e5, and f4, and a white knight on d2, and black pawns covering all the squares a knight move away from d6, the white knight might decide to embed itself by using two moves to bypass the pawn guardians. It would then, of course, be trapped, but that's not such a big problem for knights as long as they're on a very good square and can't be dislodged.)

I thought of that option also but rejected it because I had guessed it would be too devestating even with your caveats. If I am wrong about that I like your idea slightly more than mine. Mainly because it seems more "dignified".

Well, without the ability to play two checks/captures in a row, I don't think it would be horribly devastating. It would just require both players to play more cautiously. Have you ever played bughouse? The rule change would make chess a tiny bit like bughouse, because creating a hole in your position would become very dangerous when at any moment your opponent might decide to embed a piece there with a double-move.

There is a reason why rybka is rated a good 400 points higher than the highest rated gm. Gm's today use engines such as Rybka to prepare for and analyze played games to see what was the best move in every situation, to see what they missed.

When somebody today plays a complicated position right after rybka's suggestion under suspicious circumstances, like for example leaving the table, they get accused of cheating, because making the moves that rybka prefers consistently in complicated positions w/o cheating is not done by anybody. period.

I think it was GM Igor Kurnosov who was accused not to long ago, by super GM Shakriyar Mamedyarov, because he played something like 15 moves in a row exactly as rybka would have played it, under suspicious circumstances.
Fact is, human Gm's are not even expected to find the best moves most of the time, i am ofcourse now talking about the hard positions.


Unless there comes another Nimzovich into the chess world, computers will be unbeatable in, say, a 6 game match.
Also by the part about Nimzovich, i mean that if the way of thinking about chess would change once again, computers would lag behind untill they had enough game material to surpass human player once again.

All this and i havent even mentioned the human factor. I can simply show to the Carlsen vs Radjabov game, Linares 2009. As a norwegian that was a hard day for me, Carlsen blundered away the tournament in the last round when he was in a winning endgame, in a simple position, for a GM atleast, not saying it was for me. AND he had loads of time on his clock.

http://www.chessgames.com/perl/chessgame?gid=1536707
47....Kf3 saddest moment of my chess watching career. Great game though.

this was written at 03:35 am so dont pick on the grammar plz.


I should also say that i read the op in about 10 seconds and, well... yeah. this is directed at the headline.

oh boy

Although I'm not directly advocating either system..

Against d4:

An offbeat defense would be either the

Knights Tango
1.. Nf6 and 2.. Nc6

or

The Albin-Counter Gambit

Also the ..a6 Slav used to be off-beat but is pretty much almost mainstream nowadays (at least at the top levels).

This seems like a cool way to play an offbeat line that is still respectable but allows you to avoid knowing any theoretical lines. But actually it's the opposite, it can transpose into any of half a dozen more "standard" Black defences.

A good book on that opening is Tango! by Richard Palliser.

Palliser usually does a pretty solid job.


BTW I also have "Play 1.d4!" by Palliser and I'm alway curious to see what authors recommend against each other, especially if it's a recommendation aginst their own recommendation

Well against the Tango he advocates 1. d4 Nf6, 2. c4 Nc6, 3. Nc3 e5, 4. d5 Ne7 5. h4!?
Fun stuff.

QFT

I agree Palliser is an A+ Author