Does tempo matter in the QGA?

QGA is potentially great learning ground for handling of the positions with an isolated central pawn. Also, it's very helpfull opening to undrestand before learning the Semi-Slav since essentially Meran variation is kind of refined QGA.

Eric Schiller plays the Caro-Kann and the Tarrasch as black. He claims to like the Panov-type positions, so it may be worth looking into his games if you're interested.

I play the QGA to draw. As in the Petroff, you need to know your theory as black to avoid getting crushed in the mainlines. Unlike in the Petroff, it's more difficult to create an open file and trade everything off. With best play on both sides, you can expect white to have a small but noticeable positional plus. That's probably why Kramnik sticks with the Nimzo/QID against 1. d4.

It is not "that" isolated pawn where both sides have advatage in it, such as the one in Caro-Kann Panov Attack, where the isolated pawn facilitates attacking chances. In QGA the isolated pawn is more a liability, that's why you wont see it often.

So, beside isolated pawn, what else can you learn??!

QGA is very monotonic. For this reason, it is very advantageous imo if you study and prepare your lines using computer. There is no kingside attack, no complex endgame, no complex pawn structure, no stalemate. The position is so "equal" but I doubt you will find a half point because players will just play on until someone make mistake.

I really like playing QGA for a win, but alas, when White refuses to play QGA, I found my position is less preferable. So, I will pick the best theoretical defense (NID/QID) which is rich in positional patterns where understanding them imo is necessary to become a GM level player.

It's different creatures imo. I like to play the White side of Panov Attack and Anti Caro-Kann.

Yes, that is my opinion too. I have tried to understand how to make White's small advantage to become insufficient to win the game. Theoretically I can't see Black has a chance, but White may make mistake, and it is much safer to play (no complicated tactical blows here)

(Addressing the comment insisting that there is little to learn in the QGA) Anything can happen in a chess game; the opening only accounts for the first 10 moves or so of a game. It is simply ignorant to think that you can understand everything about any opening so easily.

When I choose an opening, I actually do not care with the first 10 moves or so. Every openings have typical (possible) middlegame positions, which is very important. Also, some openings have certain typical endgame. Some simply do not have endgame.

I believe that mastering chess at high level is about recognizing patterns (this is the object of my study method). And I'm trying to gather these patterns.

Look at my last Sicilian Dragon game, even tho there are many possibilities in a tactical position, I didn't make even one mistake. To me, the Dragon is refuted, just like what Fischer said, push the h-pawn and sacrifice. We choose the opening because of the middlegame, not the opening itself.

Only by knowing how to handle all the possible patterns, we can play chess at very high level imo. These patterns can be gathered, or simply be experienced while playing the game. The more we see the patterns, the more proficient we are in handling them. And by knowing which side is better/winning in any position (and why and how), chess will be so easy.

It is even better if the patterns (that are most likely to occur) are sought after. This has been done in micro, for example, we used to study how to do a minority attack against typical pawn structure. We study how to use an isolated pawn to help in an attack (and exchange the pawn in the process). But I found that this concept can be augmented even more.

Choosing the NID/QID is one of my way to get in touch with these patterns.

I don't disagree about the patterns. There are "typical" positions, but there are plenty of other positions that can arise from an opening too, particularly when more moves have been played. Remember that most of what happens in a game is determined by the players, not the opening. The king's gambit is intrinsically aggressive, yet it can lead to some fairly quiet positions; this would be because a wise player was not too greedy and played a solid defense as black. Kasparov was a brilliant attacker, yet you don't see him coming out with openings like the danish gambit or king's gambit; he prepares his attack and attacking chances with powerful, incisive play. You get better at chess mostly from playing chess, and no matter what opening you play, that's precisely what you're doing.

Look at this position below. Where is it from? (QGA). How often the same structure occur in QGA game? Is it possible for White to switch the battle to the Kingside where he has 2 Bishops, a Knight and a Queen ready for that? (no) What is the consequence of Black pushing the b-pawn? (He will lose it, in exchange of White center pawn). If White get a pawn majority in the Queenside (2 vs 1) while heavy pieces cannot be exchanged, can he win? (If the Bishop pair is not eliminated I believe yes).

Yes you are right, sometimes our opponent divert. But at high level, this act means choosing inferior position.

I have been playing the KG for quite some time now. Playing it at master level is useless. I play it because I need to master one technical "link/pattern" in playing chess. Knowing all the positional things often will become useless without the tactical skill to complement it.

The Modern Defense (e.g. 1.e4 e5 2.f4 exf4 3.Nf3 d5!) is what you should opt to "refute" (at least draw) the KG without risk. But if you know a-z of KG and tactically proficient, you can 100% refute it through the mainlines (such as 1.e4 e5 2.f4 exf4 3.Nf3 g5!!).

One thing about KG is, if you play against much stronger opponent you can use it to play for a draw. If he refuses, then he will risk himself. See game between IM David Pruess and GM Lilac (Rosentreeter variation, 1-0).

I already get to know some "patterns" and "key" how to play the KG. But damn, it requires strong calculation skill.

Too many assumptions.

I know. That's exactly the point. A simpler position will be an endgame positions, such as Lucena or Philidor hehe.

So you mean the theoretical positions in QGA where white has an isolated pawn (and there are many) are just good for black? Surely this can't be true.

I disagree with most of this: White's kingside attack is one of the most topical themes in QGA; many different type of endings can arise and also there are plenty of different pawn structures possible. The bit about stalemate I don't understand. In what opening exactly is a stalemate an important theme?

I don't understand the line people are talking about ... this is how I think the mainline goes at top levels. It can lead to very complex positions. 

Below is a screenshot of my QGA repertoire. As you can see I only play D20 (when White play an early e4 instead of e3) and D27. But as far as I know, the other lines (D20-D29) in general have similar structure regarding (sterile) e-f-g-h pawns.

From the screenshot you can see there is one D47 (Semi-Slav), that I used to study IQP because there is possibility an IQP arise in D27 (see Alekhine games). But as I said before, it rarely happen because Black should know when to inflict the isolani (cxd4). c5 is required by Black to develop his f-Bishop (through Bxc5). Basically Black has to develop first (Bb7 has been done for example) before thinking about inflicting the isolani (and of course White should not be able to exchange the isolani on d5). But if White wastes some tempo, Black can create the isolani or even protect the gambit (only when White doesn't know what he's doing).

That is an argument for my statement that the QGA is quite monotonic (not rich in "possibility") due to the manouver of Black's c and b pawns, and that the QGA is "drawish" only because it is equal (and not difficult to play), not because other elements such as existence of different color Bishops for example.

I mentioned "stalemate" for two reasons:

1) To strengthen the idea about the nature of drawishness in QGA, in comparison to Petroff for example. In QGA there is no forced Queen exchange, no forced diferent color Bishop, no three-fold repetition, no stalemate, but still many thinks that it is drawish. A game is indeed drawish when it is equal and no complexity can be enforced.

2) That QGA (1.d4 d5 2.c4 dxc4) is static, monotonic, not rich with positional possibilities. It is understandable because with QGA you can for example limit the variation to D20 and D27 with static pawn structure. Compare to 1.d4 Nf6 2.c4 e6 3.Nc3 b4, the Nimzo Indian (or its cousin QID 3.Nf3 b6) where White on the next move can choose Rubinstein or Classical or Kasparov or Shirov or other variations. You can see plenty of different endgame, different pawn structure, different strategy/plan, etc.

thats exactly what I did. Do the variations have names?