0-0 in the French?

the_real_greco

As a regular viewer of the CCC and a middling (at best!) player, I'm somewhat in the dark as to what strong human games look like. So I have a very general question about castling in French-Advance games.

In computer chess,  the very strongest engines that have given up playing 0-0 as black. SF-Booot was a good example why:

Fun game, right? Booot gets crushed- crushed!- on the kingside. The best engines stay well away from the kingside:

So, my question: What do humans usually do as black? Is the kingside practically defensible? Do they play 0-0-0? Do they drift their king into space, like Antifish did? Obviously it depends on openings, but... ?

Thanks!

robertomil81mil

I don't know 😕

the_real_greco

No one appears to know. Which is unfortunate.

PawnstormPossie

You can look at a database (2700 ELO min and C05 Tarrasch Closed-C07 Tarrasch Open) to get an idea how they play these lines.

You can also look at the Advance by fltering for C02.

Another thing to consider is move order. If you search for positions after 7 moves or so, you can get more results which could include other openings.

dpnorman

Firstly there's a difference between the Advance and the Tarrasch. Secondly as usual in chess it's not much to do with generalities and more about specific considerations in any exact position whether 0-0 is a strong move. 

blueemu

Neither of those games was a French Defense Advance variation.

llamonade

Huge attacks on the king are normal in many variations of the French. You can't draw conclusions from just 2 games.

Plus engines are bad at long term considerations (like an attack that hasn't started yet).

Humans likely wouldn't play 9...0-0 in the first game because it just looks bad. Engines don't understand that though.

the_real_greco

OK, OK. Fine everybody, I don't know my French variations. But the key feature here was that white played e5 and the the d- and e- files locked closed. And hey- maybe I was asking about the Tarrasch and didn't know it!

The point here is that AntiFish and Stockfish are (from what I've seen) the only engines that don't castle to their doom. And even for them, it takes crazy-strong hardware. Those that DO castle kingside get crushed, crushed, crushed, basically no matter when they decide to do it.

I'm just curious as to whether humans have decided the same thing, or if defending the kingside can be done (in the Tarrasch, I guess).

Also- if you think there is any aspect of a game where engines are 'bad', you haven't been paying attention.

Tebow2Baker
llamonade wrote:

Huge attacks on the king are normal in many variations of the French. You can't draw conclusions from just 2 games.

Plus engines are bad at long term considerations (like an attack that hasn't started yet).

Humans likely wouldn't play 9...0-0 in the first game because it just looks bad. Engines don't understand that though.

 

llamonade
the_real_greco wrote:

if you think there is any aspect of a game where engines are 'bad', you haven't been paying attention.

Actually it's the opposite. People who think engines are good at everything basically know nothing about it tongue.png

But sure, very soon (especially with NN engines) they will be better than humans at all aspects.

llamonade
the_real_greco wrote:

I'm just curious as to whether humans have decided the same thing, or if defending the kingside can be done (in the Tarrasch, I guess).

In chessbase livebook, only 6% of human games feature castling on move 9.

Humans know to lay the groundwork for counter chances before giving their opponent a huge target.

 

drmrboss

I don't  caste early in Caro Kann (similar to french, but Q side Bishop is out). In both opening,  "the king is relatively safe in center and there are a lot of options to do in queenside). In this game, I castled O-O-O when his pawn storm is coming to my king side. (3 mins blitz)

 

darkunorthodox88

in the french (as well as many french like lines in 1.b6), you have to be very careful when castling 0-0. the closed pawn structure is such that the kingside is one huge target to both the minor pieces as well as a pawn storm with the f-pawn. Black should postpone where to put his king as long as possible,unless he somehow traded a lot of white's attacking pieces.

In some of these lines, the engine may give 0-0 an ok eval while giving 0-0-0 an eval sometimes as bad as close +1.00, but engines unlike humans often fail comprehend that the queenside is much safer.