Against 1.e4, Alekhine's or Nimzovich Defence?

Ok @thrillerfan , I just looked in the opening explorer, and found out white does score well in the modern (because I only play ...dxe) and the exchange, but in the four pawns attack, white doesn't do as well after 5...g6. (I apologize in advance for the theme, if it hurts the eyes)

I was actually quite surprised that black did as bad as it did after 4.Nf3; My preferred variation against the Modern is the 4...c6, and although black doesn't score well (it was around 30% wins, I think), I actually do pretty well. But at my level, it's not the opening, I guess.

I do not trust the database here.  Too much garbage.

If you look at chesstempo.com, narrow it to strictly games where both players are over 2400, the 4 pawns attack (position after 5.f4) scores over 59 percent and your sideline 5 ..g6 it still scores 55 percent for White.  So maybe you found 1 line of the 4 pawns that barely meets up with the standard average.  But you also cannot compare the isolated line that scores best for Black and compare it to the entirety of another opening.

But yes, the Alekhine is not very good at all compared to the big 4.

The big difference between the Alekhine and say, the Sicilian, is that the Anti-Sicilians might score 51 percent for White and the Open Sicilian might score 53 percent.

With the Alekhine, the variance is much higher.  It might still score 54 or 55 percent overall, but that could come from a 48 percent score for White without d4 and 60 percent with.

The numbers are hypothetical as I have not looked across legit databases and filtering out weak games, but the concept is what I am trying to get across.

They don't score well because the more critical a line gets, the more precision is needed from black to get a decent position and to equalize. so if you keep going let's say 10 more moves while choosing the most popular option from black's side, you will get closer to the standard average.

the 2...d6 lines played like a Pirc ( Schuyler's dark knight system suggestion) are pretty respectable. You avoid the austrian attack type lines and play it often resembles a KID but with the c-pawn on c2.

I dont think any line in the book's chapter on it, is worse than maybe a 0.4 . the most annoying line for me has been when lower rated player chicken out and play an exchange variation when i eventually play e5, and the game turns pretty boring and its hard to squeeze a win.

Well said, database numbers dont necessarily reflect how good or bad a position is , but can also reflect a hard line to play or one where one specific trick has claimed lots of victims. Sometimes even someone playing their own opening doesnt remember with exact precision the exact line of an obscure but testing variation in move 12 for example.