My ULTRA Aggressive System against the Caro-Kann

Seems like good, how much your rating and your opponent?

Great how much your rating?

Isn't that thing shelved?

Of course white does not have to blunder with 7.Nxe4? as GM Williams did, but 6...a5! looks like a very strong novelty- actually white has to be very careful to avoid a very bad position.

So? he is a IM, you are proving nothing

so what you said its bad without even looking at the video

The bayonet attack is 4 Nc3....not 4 g4.

The 4 g4 attack....similar.

Get cracking lol....if you dont take the advance variation seriously you will be butt-hurt quickly.

thanks nighteyes!

I saw the videa and the point he gives is that he wins becouse of the psychological factor and not becouse of a better position

no thats not true only a few positions are "poker chess" in the rest its equal

Well, as I said, I chose 3. ....c5 because stockfish choose it. I dont trust GM books! e.g I have seen a lot of opening position on Benoni and Budapest whereas those books might say black is ok, but it is not in TCEC computer chess.

GM books or human games are based on statistics, and evaluation of current position.

Computer evaluation are based on perfection(depending on amount of calculation) and evaluation of unseen position at depth 30-50 whereas most cant understand how those evaluations come from.

Following computer evaluation is like riding a mad horse. However, if you know how this mad horse will behave, then it is fun and enjoyable. These are my tips how to follow the computer evaluations.

1. You must already be good at tactics, computers know almost nothing except tactics.

2. An evaluation of +1.50 doesnt mean stockfish see this position as +1.50, stockfish see the end  position of his principal variation at certain depth.

3. Current stockfish has contempt of 0.21( so he will always do overestimate position with +0.21, that means, rather than seeing 0.00 in equal position, current stockfish will evaluate as +0.21).

4. To know how stockfish evaluate current position , 

use the following command

"eval position fen (your fen position)

3. In this diagram stockfish static evaluation is seen on the right side of table. However the evaluation of his position +0.53 means the position is based on depth 26 from his principal variation of the top line. To know  where this position is coming from, stop the stockfish, follow the game according to position which end as "the position end as Rab1 and 0-0). and then evaluate the position again what stockfish think.

4. As stockfish evaluation come from perfection (based on calculated moves), +1.50 doesnt means you are likely to win.

There is a possibility that stockfish see 1 position as +1.50 but other 99 positions can be a loss. So in practical play , you may end as 99% losing by following the engine line in case you missed a single tactics.

5. Know Fail low and Fail high

When stockfish is changing the principal evaluation on his main root, there is Fail low or Fail high, stockfish need to calculate the  excluded lines again from the first search, so give enough time to stockfish to evaluate again.

6. Always give enough time yourself to evaluate most of the possible positions. As I mentioned above, +1.50 means it is trust able only when you already analysed another 99 positions. So using computer evaluation may take you about 100 times more time consuming than human evaluation.

7. Have your own chess knowledge, 

Stockfish may misevaluate +2.00 in different color bishop ending or fortress positions, and you must know by your own that the position is equal.  

In conclusion,  Human evaluations based on the current position is more reliable than computer evaluation based on the extreme depth of a position that you never know. (P.S, however, I have been following computer programming for more than 5 years and I always enjoy with my guinea pig, stockfish)

Actually Leela evaluation is much more trust-able than stockfish , provided that they both are same strength. (however Leela is still 200-300 elo below stockfish)

Leela prefer 4. dxc5 based on her own statistical experiences .That means 4.dxc5 has more chance of winning in her experience (she already had +20 million trained games) 

OK Phoenyx....

Isn't 3...c5 a gambit?

Yes, although Black gets the pawn back in most lines, but faces a few positional challenges.

@Alltheusernamestaken challange me to a game in the Qh5 line (2+1 - 10+0) if you gut the guts all i know is what is shown in the video (but at least i watched it 1 time)

Hahaha if I have the guts? Lets go

challenged @Alltheusernamestaken to 3+0 rated and he went offline

I accepted @testaaaaa challenge and this happened:

EASY WIN