Why the common opening sequence ...Bg4 f3?

I noticed that a lot of book openings show the pair of moves ...Bg4 f3, and I wanted to check why that strange sequence is so standard, since obviously Black will lose a tempo. I have been assuming that Black is attempting to provoke White into exposing his kingside castled king by playing that weakening move, but maybe my assumption has been wrong. Another possibility is that Black ultimately wants to place that bishop at g6, and is just taking an indirect route there. Another possibility is that there exist different reasons that depend on what the bishop is aimed at, such as to attack the queen, versus creating a pin, versus simple development. What do you think?

Below are some examples of what I mean.

As an attack on d1:

As a pin:

As bishop placement:

I guess most of the time it provokes f3 which weakens the dark squares. But it doesn't look like a big deal in some of those positions.

Example 1: Qb3 is the main move, not f3.

Sometimes black sees f3 as a harmful move for white, so he spends a tempo to force it so white would not be gaining a tempo.

OK, thanks everyone. The consensus is definitely what I thought, too: that Black believes f3 weakens White's position enough to be worth a tempo.

Logozar: Yes, I'm aware of that. However, I've also studied that Qb3 line and I dislike it because either White ends up getting ugly doubled b-pawns, or White finds his queen misplaced on the queenside flank while all the action starts happening in the middle. Ultimately I suppose it's a matter of taste. I find the Caro-Kann "slippery," like the French: Black always has a good defense that ultimately embarasses White.

P.S.--I found another ...Bg4 f3 game tonight:

In the caro Kann example black has little choice than to go to g4, he knows f3 will kick him but has no other place to put the bishop, and for white pushing him out the way with f3 is less annoying than beong pinned

Good point. 6...Bd7 would prevent the Black queen from protecting his KB if ...Bd6, and ...Bb7 would aim that bishop at a largely immobile d-pawn.

In the Caro-Kan example Black wants to oppose White's K-bishop, but cannot play Bf5, so has to take the route B-g4-h5-g6. The same often happens in the French Exchnge Variation. In the Steinitz Lopez the idea of f6 is to make a strong point of e5.

Also, when f3 is played, White can no longer play Nf3, which would be the normal development in some of these examples.

The move f2-f3 (f7-f6) has many drawbacks! As well as hampering the knight's development the KB is hampered.  The pawns itself gets in the way, and if the bishop is still at home it can also be hampered by Nfe2. The black squares are weakened (white squares in the case of f7-f6)  and pressure from b6 to f2/g1 and h4 to f2/e1 can be a problem.  The g3 square is weakened.   g2-g3 becomes problematic as g3 and h2 are more vulnerable to sacrifices and the second rank would be very weak with no pawn on f2. Finally the pawn at f3 can itself become a target for a pawn storm , as in the classical King's Indian.  Whether it turns out to be a good move or not depends on whether these drawbacks are exploitable, on the relevance of  the plus points of the move (control of e4 and g4, space at f2 for a Bishop or as luft for the King, blocking an attack on the  d1-h5 diagonal, gaining time by attacking a piece at g4, blocking the possibility of f7-f5-f4-f3 etc.)  and on how well the player of f2-f3 plays afterwards.

There is no general rule.  Everything depends on the specific position.

Yes, I was simply stating another drawback of f3, not suggesting that this was the only problem with it. As Eduard Gufeld once quipped when asked about his opinion on the Sämisch KID, "Ask the gentleman on g1 what he thinks of the move f2-f3."

Yes I realise that you are strong enough to be aware of the many deficiencies of f3, but I thought anyone that needs to ask the original question also needs a comprehensive answer.

Nice quotation!

I strongly disagree that Qb3 is bad - I think it is much stronger than f3 here. The only realistic way he is getting doubled b pawns is if black plays Qb6 and if that happens Qxb6 axb6 Na3 is better for white (though allowing Qxb3 is fine - the pawns are not isolated and white gets the a file).

Qb3 causes black to spend a tempo to defend b7.

Black should defend b7 with Qc8 or Qd7.

Na5 loses tempo to Qa4+ Bd7 Qc2. Fischer crushed Petrosian here.

In the Qd7/Qc8 lines:

After some moves such as: 0-0 Rae1 Ne5 the theory shows black often trading Bg4 or maneuvering it to g6 to trade it with Bd3. White can often make a battery with Qc2 or play Qd1 (After Rae1) and have a kingside pawn storm (sometimes combined with a bishop - after Ne5 Nxe5 Bxe5) such as Re3 h4 Rh3 h5 (sometimes g4-g5, sometimes f4-Rf3)

Yes, I'm starting to see the bigger picture now, I believe: Both ...Bg4 and f3 can have several simultaneous advantages for each player, sometimes almost a necessity for Black in some positions, which is why it is often found as a standard book line (even if not the #1 most popular).

By the way, here is another such opening I found this morning:

Another good observation, thanks! I like to know all the nuances like this, if possible, so thanks for your completeness. I hadn't thought of the drawback of f3 on White's KB.

Another such game I just found:

On Gufeld's quote about the KID Saemisch, he was very much a KID player, and the Saemisch scores better against the KID than the main line, so he may have had another reason to put it down.

I have just received  the book understanding the Queens gambit accepted (great book)  Its recommendation against Bg5 happens to be

I didn't say that 7. Qb3 was bad--after all, it's the #1 most popular move there--only that I didn't like it when I looked at it. Unfortunately, my notes on that variation are in a different file in a different format, so I can't easily find my old documentation. Black got doubled b-pawns in some of those variations, and I didn't like the feel of any of the ensuing games, either as White or as Black, so I stopped playing moves that would lead to such a game.

7. Qb3 {"Nun Attack." #1 pop. w 41%.}
7. Nf3 {"Kieninger Attack." #2 pop. n 36%.}
7. Ne2 {"Czerniak Attack." #3 pop. n 45%.}
7. f3 {"Szekeres Attack." #4 pop. b 48%.}

By the way, here's another ...Bg4 f3 game I just found:

I'm a KID player too, and whilst that may have been the case (the Sämisch scoring better than the main line) back in the days when Gufeld was active, the Sämisch today isn't as promising for White against the KID. There's the 6...c5 gambit (5.f3 O-O 6.Be3 c5! =), which White scores just 52% against, as well as 5...a6 (also 52%) lines as well as other variations which White must prepare for. I checked on Chessbase and White scores 57% with the Sämisch, compared to 58% with the main line. In fact, amongst the ten top-played moves, it scores the 3rd= worst. That's not to say that it's not a perfectly viable opening, but it certainly isn't the theoretically best choice against the KID at the moment.