The stats are from that other web site’s database. They actually keep track of what players play and let’s you check what moves are played and how winning they are by rating. The site has been around since 2010, has about 32 million games, and most of the games have been played in the last few years. Feel free to go and check for yourself.
May you please suggest the site's name?
So I did a deeper dive into the statistics and started breaking down win/loss percentages by rating.
KG 1600 rated: 53% win for white, 44% win for black.
QG 1600 rated: 53% win for white, 43% win for black.
KG 2000 rated: 49% win for white, 47% win for black.
QG 2000 rated: 51% win for white, 43% win for black.
KG: 2500 rated: 50% win for white, 44% win for black.
QG 2500 rated: 48% win for white, 42% win for black.
KG Master games: 30% win for white, 36% win for black.
QG Master games: 33% win for white, 19% win for black.
So obviously, amongst masters, the QG does significantly better for white. What I found really strange was that 2500 rated players did better with the KG as white than 2000 players. This is completely counterintuitive. The narrative is that better players handle the KG better than lower rated ones, but this is completely not the case. My guess is that if you're 2500 and playing the KG, you *really* know your theory.
Anyway, the entire narrative that QG is better than the KG is based upon the way masters play, and the *unsound* inference that this somehow applies to the way mortals play chess. The only reasonable conclusion, based upon the evidence provided by a large database of games, is that you can play either and will do essentially equally well, assuming equal understanding of the positions that arise.