QUOTE:
"I am not going to debate the question "is the Petrov boring?". But in my database, with 1300+ games with both players ELO 2500+, about 70% of the games end in a draw. This rate is considerably lower for the King's Gambit." First of all, I completely agree that people cannot debate whether an opening is boring or not since its completely subjective. There's no point in trying since people need to do is shrug and say "well, its boring to me"...end of story.
I decided to check my database (BigBase 2006 with TWIC every week) to see how my results compared to yours. I only looked at games where both players are 2500+ and 1.e4 e5 was played. Here are the results:
Total games: 11885 games 2.f4 scored a standard
54% for white from 136 games (with 33% 1-0 and
42% drawn.)
2.Nf3 Nf6 gives white an impressive
57% from 1743 games (with 25% 1-0 and
65% drawn.)
The first problem is the relatively small sample size in the King's Gambit since its only played about 1% of the time at this level. In fact, in recent years its actually gotten even less popular as there are only a handful of games per year.
Also a significant amount of the Kings Gambit games are from rapid or internet events (about 25%-30%) but, that said, this is still a
very good showing for an unrated opening. (BTW 2...Bc5 appears best statistically and scores 55% for Black with a 2653 performance from a 2556 avg rating; but again, its much too small a sample to get excited about. Just thought I'd share that info).
Now, the Petroff only has about 5%-8% rapid games and is a more substantial sample (since its played about 15% of the time at this level). It's clear that while black is obviously much more likely to get a draw than in almost any other major opening, his winning chances are quite low (10% versus 25% in the Kings Gambit). So, if you simply hate draws then the Kings Gambit is much more likely to be decisive (no surprises), but you should also know that Black is more than twice as likely to win than he is in a Petroff whereas your own chances of winning improve much much less. That whole risk versus reward thing...
However, as Yusupov and several others have pointed out, its almost impossible to get reliable statistics in the Petroff since its the choice of many professionals when playing a pre-arranged draw (where almost no one would play a King's Gambit when they've agreed to a draw). Who knows what the stats would be like if all the "rehearsed" games where removed. So, take all this with a grain of salt, but these are the numbers as they are.
Cheers,
Nietzsche