Is the Sicilian Defense good?

@15

"what does that mean -- to say e5 is better from a theoretical perspective?"
++ There are less pitfalls for black after 1 e4 e5 than after 1 e4 c5.
This is logical: 1...e5 occupies a central square e5, controls a central square d4, opens a diagonal to develop Bf8. On the contrary 1...c5 only controls central square d4 and opens a diagonal to develop Qd8. On the other hand with 1...c5 black gives white a lead in development in exchange for a better pawn structure. That is more difficult for both.

"With best play, it's ultimately a draw either way"
++ Yes, but after 1 e4 c5 black has less margin than after 1 e4 e5.
Essentially 1 e4 c5 increases the probability of white making an error,
but at the cost of increasing even more the probability of black making an error.
In other words: it is easier for white to draw after 1 e4 e5 than after 1 e4 c5.
It is more difficult for black to draw after 1 e4 c5 than after 1 e4 e5.
That is why we see more 1 e4 e5 in top level matches: you win a match if you win one game and draw all other games.
That is also why we see more 1 e4 c5 in tournaments: if you win one game and draw all other games, then you do not win the tournament but you end up in the middle of the table.
To win a tournament a top grandmaster must win with black against lesser opponents.

"assuming absolutely best play -- then 1. ... e5 can't be better (or worse)."
++ That is right. Also 1...e6, 1...c6 and presumably a few other moves are enough to draw,
but require much more precise play to do so:
less legal moves to hold the draw, more legal moves that lose.

The thing is, what you’re describing is a practical advantage, not a theoretical advantage. If the Sicilian theoretically equalizes — and it does - then it’s theoretically equal.

And it turns out, even practically speaking, c5 is at least as good as e5. We know this from stats.

Both e5 and c5 are sound AND reasonably practical. All that hot air is just trying to get around those two facts, which are pretty much indisputable. Maybe one appeals to you more than the other. Fine, but don’t try to turn that into some universal statement.

@23

Both 1 e4 e5 and 1 e4 c5 draw so game-theoretically they are equivalent indeed.
However, after 1 e4 e5 less of the legal moves lose and more draw,
while in the Sicilian more of the legal moves lose and less draw.
That is also a theoretical advantage with some practical consequences.

Except those “practical” consequences aren’t really very practical if, in practice, the Sicilian scores (slightly) better than e5.* In practice, people figure out how to find those good moves.

*this is true at both the amateur and master levels. In fact, the lichess database shows Black scores absolutely equal with the Sicilian but six points worse with e5

@25

"In practice, people figure out how to find those good moves."
++ It is very difficult, like walking on the edge of a razor. An example: black and white find all the good moves, but in the end black is too tired and errs in a theoretically drawn position.
https://www.chessgames.com/perl/chessgame?gid=1993385

Master game stats:

Lichess game stats:

So why does c5 score better if e5 is practically better? How do you explain that away?

@28

The lower level results are irrelevant. The results of the Sicilian are better because the stronger players play it. Those games are mostly speed chess. The draw rate is ridiculously low.

The master results are somewhat relevant. If you do not want to lose as black, then 1...e5.
If you need to win as black then 1...c5, but at greater risk of losing.

Even the master results are not dependable: otherwise 1...g6 would emerge preferable.

I would agree insofar as I don’t think the differences among these are big enough to dictate what someone should or shouldn’t play. You should play what you’re comfortable with.

But they do refute the old canard that the Sicilian is practically more difficult.

@31

"I don’t think the differences among these are big enough to dictate what someone should or shouldn’t play." ++ If you filter out blitz and rapid games and filter out lower level master games keeping only > 2700, then differences will show clearer.

"You should play what you’re comfortable with."
++ Most grandmasters are uncomfortable with losing.
Top grandmasters are uncomfortable with too many draws.

"they do refute the old canard that the Sicilian is practically more difficult"
++ That still stays: Sicilian is more difficult, but for both sides.

Here's the thing being missed here - tygxc literally just said it but didn't make this point.

If you play the sicilian as black, your opponent is more likely weaker than you are and you're going for a win. That is part of why the results for the sicilian are so good there.

For a fair way to do it when there are people changing their openings based on their opponent's rating, you need to look at the average rating of the player vs the average performance of them playing that opening. The average vs performance takes into account the rating difference between players, so it's not just like the smith morra or Nakamura's b3 with a huge score for white because people only dare play them against fish.

Sicilian still slightly scrapes it in terms of performance at the lower levels or even 2500, but it's only like a few elo points difference. It shouldn't be cause for concern or for upheaval of your whole repertoire. Playing the Sicilian you're not suddenly going to have equal chances as black as if you were white as those statistics would imply. At the Super GM level I believe e5 very slightly wins out in performance again.

I assume this is due to the sheer amount of options white has against e5 that black has to be prepared for and know.

It's an interesting hypothesis, but I'd want to see some evidence for it.

As far as I know the Lichess algorithm doesn't have a bias in favor of matching a weaker White player with a stronger Black player. And I don't know of any evidence that many people switch their openings depending on whehter they're playing up or down (especially since most pairings are between players pretty close in rating -- are you really going to change your approach if you're playing someone 50 points weaker vs 50 points stronger??). Lichess's statistics are based on a *lot* of data, so you'd have to show a pretty significant bias to make this explanation stick.

Sounds to me like a way of trying to come up with a rationale for a position that some people seem to feel strongly but that the data doesn't actually support.

@35

"the data doesn't actually support"
++ You cannot draw conclusions about an opening from data about speed chess.
For what it is worth: 42% draws 1 e4 c5 and 49% draws 1 e4 e5 by masters support that Sicilian games contain more errors and thus that the Sicilian is harder to play for both sides.

Right. So the Sicilian is harder to play — except that it scores well in a colossal number of games at a website comparable to this one. And it scores well in master games. But *somewhere* there’s a place where it’s too hard. Is that your point?

If it were truly too hard, blitz is exactly where you’d see this, since defense requires more time and precision than attack.

@37

"blitz is exactly where you’d see this"
++ No. In blitz 1 e4 Na6 or 1 e4 h6 can score well.
In your table 1 e4 g6 would be superior by your logic.

We see the difficulty of the Sicilian in classical games.
I gave an example game @26.
Two other example games:
https://www.chessgames.com/perl/chessgame?gid=1048762

https://www.chessgames.com/perl/chessgame?gid=1044020

We no longer see that in ICCF correspondence: all draws 1...e5 as well as 1...c5.

You seem to really want to believe that the Sicilian is a bad practical weapon for black, to the point where you’re posting a sample of two as a counterweight to a data set of 300 million games. It’s clear I won’t convince you.

@39

"the Sicilian is a bad practical weapon for black"
++ No, it is a good practical weapon for black to play for a win but at greater risk of losing.
It is harder to play for both sides than 1 e4 e5.
I presented 3 games to illustrate how hard it is to play.

If you're seeking to 'not lose' then .... e5 is likely best. If you're seeking imbalance in a position where a win more 'likely', then .... c5 is probably better. I like the Caro Kann personally, but I'm a club player so I'm not as much into imbalance. It helps though.

No

To play the Sicilian as Black, you need to have very strong tactical awareness. You also have to be willing to memorize moves, and LOTS of them. The better you get, the more you will need to memorize.

I'd only recommend the Sicilian to someone who keeps drawing people. And if there's anyone that has the ability to keep drawing people, it's the strongest players in the world.

You're right: the Sicilian (as Black) is good. But you have to know what you're doing, and I mean really, really well. You have to know when to stop/start attacking and when to start defending. I'd recommend something much calmer, such as the Caro-Kann, because it's a lot easier to not blunder and because you don't need to be familiar with tons of variations.