1.g3 vs 1.Nf3

What do you mean by „good“? If you’re an engine it doesn’t really matter.

@Sarg0n said in #9:
> I use Nf3 for or all sorts of d4 openings, it is just a transposition. I don’t have always a g3 follow-up in mind.
>
> Like 1.Nf3 Nf6 2.d4 d5.
>
> I don’t like c4 and g3 because they allow 1. ...e5!
but you allow 1.c5 so you will have to either play sicilian or play KIA or reti via 2.g3

I agree that 1.Nf3 looks slightly more useful (because 1.g3 allows a move like 1... e5, which is not necessarily better than 1... d5 but gives Black more choice). Still, 1.g3 is no mistake. If Black wants to play d5 (or Nf6) anyway, it often makes no difference.

The initial position is sorted by popularity and Nf3 is played more often than g3.

Compare the databases to see which is more popular after the move is done.
After 1. g3 ... the most popular percentage is
Master database: 37%
Lichess database: 29%

After 1. Nf3 ... the most popular percentage is
Master database: 47%
Lichess database: 32%

Well Nf3 is still in the lead compared to g3.
A popular move only means many people played that move, but does not prove that the majority was right to play it.

Now whats the WDL average
g3 ... 36% ... 39% ... 25%
Nf3 ... 34% ... 44% ... 23%

Nf3 will give more draws.
g3 will give more wins for white.

Preventing another player from winning should be a good thing.
White's e4, d4, Nf3, c4 affects Black's win ratio a bit more than the other initial moves.
If white plays anything else, the percentages show that black's odds of winning increase.

I like where the question is setting itself up to be a comparative discussion of ideas back to the board itself. well, done, so we can confront our words and concepts. not lines of moves. (those are fine, but raw data to explain and throw at each other as examples or counter examples in support of the more important and I would say abstract ideas that might help us reduce the complexity of chess from a bunch of equally distinct strings of moves to board based reasonings chunks, chess is complex enough, so we might still get a lifetime expertise based social competition layer, even with a more consistent and efficacious theory of chess (not opening theory thing, but probably explaining it). Theory toward understanding the forces in play above the specifices of the placements). Grandiose wording. but I could argue each and every aspect of what I said. with good faith.

Like in physics, we do not name all the exact water molecules placement in a every clouds, and give that different name if one of those molecule is not same placement. (I know rigged analogy, but wait). While I would think we could predict shape evolution to some tolerance level, that is not even what physics was even about, but maybe predict invariant summaries of variables that can predict the dynamics of their existence as a set of clouds or its average pressure, temperature, and other determining macroscopic variables to monitor that would allow the reduced complexity models to predict the ensemble behavior of those variables later. Well, I wish I was talking about mathematical objects here, for I might have twisted some physics to make a story.. but it might be robust to benevolent criticism from a physicist....

I did not read the thread, just commenting on the level of the op. question. That is my kind of question driven chess. wrong or not. it is what I like to ponder, while playing in my slow games.

NB: i suggest reader to add their own punctuation and take breaths while reading. do not gasp.

@MAGNus_204 said in #12:
> but you allow 1.c5 so you will have to either play sicilian or play KIA or reti via 2.g3

What about 2.e3!? Pretty trendy I would say.

Most games (concerning our topic) are reaching the position after 1.Nf3 d5 2.g3 Nf6 3.Bg2 anyway, so in this case it is not relevant whether you have started with Nf3 or with g3. And 1.g3 e5 is no problem for White either, 2.c4 is perfectly fine.
1... c5 cannot be avoided anyway.