After making my engines evaluate the starting position for more than 10 hours each, here is what I got:


Fritz 12:

Depth = 25/65

Time = 10 hours and 30 minutes

  • 1.f3 : -0.90
  • 1.g4 : -0.77
  • 1.Nh3 : -0.56
  • 1.Na3 : -0.52
  • 1.h4 : -0.47
  • 1.b4 : -0.47
  • 1.a4 : -0.42

All 13 other moves are -0.25 or more.


Houdini 3 Pro x64

Depth = 29/76

Time = 11 hours and 45 minutes

  • 1.g4 : -0.51
  • 1.f3 : -0.50
  • 1.Nh3 : -0.38
  • 1.Na3 : -0.37
  • 1.h4 : -0.25
  • 1.b4 : -0.20
  • 1.a4 : -0.17

All 13 other moves are -0.12 or more.


One thing seems very probable: the worst possible first move is either 1.f3 or 1.g4.

So between these two moves: which is the number #1 worst possible move?

  • 1
    duplicate - chess.stackexchange.com/questions/132/…
    – Wes
    Jul 3 '14 at 23:35
  • 3
    @Wes: Not really since my question is specific to only 1.g4 and 1.f3, while that other question discussed all moves. And there is absolutely no mention of 1.g4 on that other question.
    – Fate
    Jul 3 '14 at 23:41
  • 3
    That... doesn't mean he's a proponent. It's more of an eff-you. 1.g4 prepares to develop a piece to a good square. The other two moves don't.
    – Cleveland
    Jul 4 '14 at 2:07
  • 3
    Petrosian, if you really want a comparison of the relative merits of 1.f3 and 1.g4, it would probably be best to limit the wording of your question to just that, and to discard from your post all talk about these being the two worst possible, as that is quite a controversial and distracting assertion to make.
    – ETD
    Jul 4 '14 at 13:23
  • 2
    1.g4 (the Grob) is pretty goofy but is playable and has some established theory. 1.f3 is just plain bad.
    – dfan
    Jul 12 '14 at 15:36

What is the worst possible first move for White: 1.f3 or 1.g4?

f3 doesn't contribute to the piece development ( it actually hinders it by taking away the best square for the Ng1 ), it doesn't fight for the center ( it only strengthens the e4 square that White can claim anyway with 1.Nc3 or 1.e4 ) and above all it exposes the king.

g4 vacates g2 for the Bf1 ( fianchettoed bishop is usually good ), it helps in the fight for the center indirectly ( you can weaken Black's control of d5 and e4 if Black plays ...Nf6 at some point, by simply playing g5 ) and pawn at g4 can be useful in kingside attack if Black castles short. It does weaken f4 and h4, so it is somewhat dubious, but it is playable ( although I doubt White can reach full equality...).

Therefore, when we compare the above facts, we can see that 1.f3 is clearly the worst first move, based on the opening theory ( since only one move is played it is to early to make final conclusions, since White may transpose into decent position with his next moves ).

  • It may be noted, that king safety is an especially important factor in "human" play. An engine may enjoy the extra piece in a FriedLiverAttack - but a human is just absolutely screwed. Also: Never play f[3,6]! Neverrrr! Apr 23 '19 at 14:01
  • I may agree with the answer, but I don't with the reasoning. We cannot say that a movei s good or bad just based on words, but rather by how the opponent can exploit it. I can also claim that the weakness of the h4 square is devastating and therefore g4 is worse, since f3 does not cause anything like that. For that, we need actual analysis!
    – David
    Apr 24 '19 at 9:50

I have probably the largest high quality database in private hands on the planet, including human and engine games. Tens of millions of games. A proper assessment of a move's merit has to include both a deep evaluation of the strongest engine available (right now that's Stockfish) combined with comprehensive empirical data.

Empirically, f3 is definitely the worst. Nh3 and g4 challenge it but cannot match it in terms of chess malpractice. As for the defenders of the Grob (g4), sorry, it's a bad move, no question, and this opinion is double-reinforced by how often it has been seen, which is about 4-5x more often than f3. Its principal advantage is that it doesn't lose material right away by force.


It's terribly wrong to dump an engine output as an opening analysis! Engines are weaker in strategy comparing to humans. An opening analysis should be done by a computer-aided human, not by an engine itself.

  • 1
    Why the downvote? Pointing out that it isn't valid to use engines in this way is helpful.
    – Tony Ennis
    Dec 27 '15 at 6:03
  • 2
    I think 1.a4 and 1.h4 can't possibly the worse moves.
    – SmallChess
    Dec 27 '15 at 8:02
  • @Student T Why? I would really like to hear your reason.
    – ferit
    Dec 27 '15 at 10:33
  • @TonyEnnis Nonsense :)
    – ferit
    Dec 27 '15 at 15:17
  • He is actually right. If he wasn't, there would be no opening-related content, you would just show the position to Stockfish and learn his answer. Indeed, first engine choice are not always mainline theory
    – David
    Apr 24 '19 at 9:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.