The notion that the dragon (and especially the accellerated dragon) are "inferior" is absurd.

Yes, you will lose some games very quickly if you play the dragon (or the Najdorf, or most open sicilians). 19 moves is pretty fast, but it's true, when you play the yugoslav and you lose, the loss will often be ugly, brutish, and short.

One mistake and you're facing a quick middlegame mate.

The nice thing is that, one mistake from white and HE'S getting mated, too - so it's not one-sided. That's why the dragon isn't "inferior."

It is extremely high risk. Nobody's getting mated on move 19 of a closed Ruy. And high-risk openings aren't for everyone.

But they're not "inferior." You get rewards for that high risk.

I am not talking about risks, but about results. The Dragon and accelerated Dragon are inferior, because they don't show good results in grandmaster play.
Simply speaking, Black does not score well! There are other and better sicilian variations where Black scores much better!

Nobi, as Ronaldinho already pointed out, GM's may score more losses with the dragon than most openings, but they also score more wins. It's an opening with a lower drawing percentage than others. So are you looking at statistics just based on percentage of losses or what? Because the dragon may have a slightly lower overall percentage than some other openings in GM play (I'm not the type to look at stats, so I honestly don't know), but it should be relatively close to other openings that are popular at that level.

More importantly, whether an opening is the "best" at grandmaster level has absolutely nothing to do with whether or not it's worth using at the amateur level, or even for low level masters. At low-mid level of play, the person who knows an opening better will get an advantage, regardless of the objective quality of the opening. Once you hit around 1600+, that's no longer true, and you do have to play openings that are at least slightly sound. But even then, any opening that's played by grandmasters even occasionally should be good enough for those of us who aren't playing for the world championship.

In other words, the very fact that there are grandmaster level statistics for the dragon is all the proof I need that it's a sound opening that's good enough for any player here on CF.

--Fromper

While I disagree somewhat with the 'anything is playable at an amature level' in principle (and I compleatly admit this is a matter of taste, but it is mine.) I think there are some serious problems with the recent trend towards statistics particularly in regard to opening preperation in chess. My first complaint is as a chess player. I think hanging your hat only on the current statistically favorable openings crushes creativity by stearing everyone of every playing style towards only the 'best' openings, and that goes way beyond saying people shoulden't play openings that have been either refuted or are compleatly impossible to respect at the grandmaster level. My second complaint is that of a statician, are these 'statistics' weighted properly, do they take into account controls for many variables such as - changes in opening theory, what lines and subvariations are played, opponents, situations, time controls, players who consistently play lines as opposed to someone trying a surprise variation? At the end, my point is one must ALWAYS be very careful when using statistical analysis. It is undoubtedly an important and powerful tool, but it is also an extreemly complex one, I know plenty of Masters students in statistics who struggle with its challanges, so don't take the numbers that pop up on chess base or are printed in a magazine article as gospel. Most openings ARE playable, and I am confident that players who play the opening they can feel passionate about, rather than just what the numbers say will do better. Also great inovators tended to make the openings better than anyone believed they could be. My old coach though not a world champion contender was a strong GM and brilliant with the Alekhine, perhaps not the best statistically but great for him.

Keres,Smyslov, Alekhine, and maybe Fischer, studied particular openings.
What I know Keres and Smyslov did is when they lost or drew a game against a particular player they went back to the drawing board. They went over the games and looked for improvements. It was a matter of getting it right. Reinfeld in fact wrote a book on the subject specifically about Keres and Smyslov.
Keres in the Dover edition indicated especially against a game against Reshevky that he prepared the opening.
Barden would keep his notes on cards. Note that Alekhine indicated never to memorize an Opening. Never make a move without knowing why.
What I would suggest is forget about an opening repertoire. Learn to play the game. Capablanca did not know Openings. Reshevsky as a just under teenager defeated Griffith quite rapidly. Griffith was the editor of the predecessor of MCO. Learn the principles of chess, establishing a plan in the Middle Game, and the rules of the end game.
I once played a Golden Knight Postal Game (USCF) against a player. Nothing he played was in a Chess Opening Manual. He was the top Postal Player for years.

First of all: Statistics prove nothing. Do you know what the best scoring move for White is? 1 g4.

Second of all: Even if statistics did prove anything, after you get rid of all of the badly played dragons (the a6 lines, and many of the old dragon games, which were played before they knew how to play the dragon correctly), you are left with decent statistics. Look at Carlsen. Look at Tiviakov. There are plenty of other examples out there. The Dragon is a fine opening.

I think Fromper's point is a good one.