Do amateur players benefit from playing white? (Statistical analysis)

tldr: On average, playing white gives amateurs a 0.035 bump in points scored (where a draw is 0.5).

We often hear GMs discuss the advantage of playing as white, but I wasn't sure if that applied to amateurs as well. To assess this, I downloaded all games played on Lichess in April 2017. When only considering Rated Classical games, this gives us a sample of about 2.8 million games.

Question 1: Does white have an edge in amateur games?

Answer 1: Yes.

Black wins (White loses) 46.0% of its games
Black draws (White draws) 4.3% of its games
Black loses (White wins) 49.7% of its games

On average, being white increases by .035 the number of "points" scored (with a draw equal to 0.5). I obtained that estimate using a simple regression model.

Question 2: Does the "white advantage" increase with ratings? Do better players draw more benefit from opening as white?

Answer 2: I ran a bunch of regression models to test this, but didn't find a result. So no, it doesn't look like the white edge increases with ratings.

Hey that's pretty cool. I've wondered this as well.

I always thought Black benefited more but maybe that's just me.

I recall some study about over the board play and the result was that the white advantage diminishes with increasing rating.
Modern top grandmaster games e.g. Carlsen seem to indicate that the white player does not even try to gain an advantage and is happy with an even middle game with chances for both sides to go wrong.

@tpr

That's interesting. My intuition for amateurs was that any opening advantage would be "diluted" over the course of the game by tactical and positional mistakes.

I guess there may be countervailing forces, where amateurs get caught in less opening traps as white, since they have more control.

It’s hard to believe that people under 1200 See any white effect at all.

What would be interesting is a study of GM games specifically in the last 2/3 years of game results.
From watching the elite tournaments, it seems Black is winning as many games as White, even had a few more wins in a couple of tournaments. Elite players have stated White's opening move advantage is not as big as once shown, as Defenses have made greater improvements relative to new theory for any White advantage.

Anyway, win percentages favored White by 15% for the top GM's in days past. Recently, it's apparent to be quite closer.

I see 2 possible reasons for the result.
1) Opening traps as you mention. White has more chances to overrun a weak opponent. Think of legal mate, Ponziani defence, Petroff Russian Defence, Two Knights Defence Fried Liver Attack and so on.
2) Top grandmasters have much more defensive skills, preventing attacks, simplifying to endgames they know they can hold to a draw etc.

In computer tournaments black has almost no chance to win.

Houdini scored some wins with black over Kodomo in the 2017 TCEC superfinal.
Alpha Zero also beat Stockfish a few times playing black.
If two equally matched opponents play perfectly, then the outcome is a draw. Only black has more ways to go wrong than white.