Why is 3.e3 considered better than 3.e4 in the QGA?

I mean, 3.e4 takes control of the center, as is said in the books. In addition, white gains space for his pieces and free the DSB. But I´ve noticed the top players plays always 3.e3. Even the engines shows 3.e3 as their first option. Is there a tactical reason for 3e3 being better than 3e4?

3 e3 is not better than 3 e4.

EDIT: 3 e4 is critical, but lots of strong players don't play it, because it's an inordinate amount of work for a secondary defense like the QGA - you have to be prepared for 3...e5, 3...Nf6 (and all the subvariations within those two main lines), as well as the sharp 3...c5 and 3...Nc6. 

3 e3 is probably just as strong, and you only really need to know two lines: 3...e5 4 Bxc4 ed4 5 ed4 Bd6 6 Nf3 Nf6 7 Qe2+!, and the main variation 3...Nf6 4 Bxc4 e6 5 Nf3 c5 6 0-0 a6 etc., when 7 dc5!? is annoying.

Both statements are incorrect.  There are slightly more master games with 3. e4 (3582 vs 3254 in the Explorer). SF9+ shows 3. e4 as the top move (d=31, Cloud). 

 Post number 3, by m_n0, is spot on. I play The Queen's Gambit a lot (played e4, as well as e3), but I prefer e4 for sure. e3 to the QGA is of course the "normal variation", however, this does not make it better. As with any opening sidelines, and deviations, may be stronger than the more common lines - just look at all of the solid deviations in the Semi-Slav for example. e4 just takes more theory.

This is news to me. Avrukh changed his recommendation from 3.e3 to 3.e4 in GM 1B if this is any indication of trends.

People who QG actually still want to hold an = material count. So it really [Censored-me] them off when this happens.

It isn't considered better. Typical opening forum false premise.

I agree. White's main move is actually 3.Nf3 followed later by e3.

3.e3 is one of those moves that look innocuous but contains some bite while 3.e4 is quite aggressive and tacitly offers a true gambit. 

I've gone back and forth on preferring 3.e3. When I first started playing the Queen's Gambit, everyone accepted it and I got great attacks playing 3.e4. But later on in my development, I played 3.e3 more. I was impressed by the history of the opening, learning from LaBourdonnais and Timman. But then 3.e4 became fashionable for a while among elite GMs and I returned to my old stomping grounds. Currently, I tend to play 3.e3 more often than either 3.e4 or 3.Nf3, but all three are excellent. 

Don't confuse fashion with an objective assessment of which is better.

Cerebellum say 0.20+ for 3. e3 vs 0.17 for 3. e4 at depth 50 on brainfish. So both are playable, For people who dont know about brainfish and cerebellum, check this website( brainfish is the strongest chess program in the world, due to opening libary database) 

http://www.zipproth.de/

Thanks guys for sharing your knowledge, @TwoMove, @SmyslovFan, maybe you are right, and 3.e3 is not really considered better, I assumed it as I´ve seen the strongest GM usually play it, but maybe is just a matter of fashion and not about the strongest move.

Strong players don't always choose the openings they believe are objectively strongest. Above all other considerations, pragmatism tends to prevail. 3.e4 is often considered to be the most ambitious try for an advantage, but it demands the most theoretical preparation.

Stockfish 8, 20-ply, evaluates 3 e3 as +0.30 P equivs. and 3 e4 as 0.00, with 3...e5 and 3...c5 as Black's best two replies, though as the moves play out there's not a lot of difference.

Actually, chess.com's Opening Explorer shows that 3 e4 is slightly prefered to 3 e3 among masters, and it had gotten better results.  3 Nf3 is played more often the 3 e3 and 3 e4 put together:

@mecuelgalapieza

I think it's because in the variation where White plays e4 Black can play b5 and save the pawn (3.e4 b5 4.a4 c6) but if White plays e3 it changes the whole thing : 3.e3 b5 4.a4 c6 5.axb5 cxb5 6.Qf3!

When writing this the "insert board option" dosen't work, sorry.

I do believe that e3 insures that the pawn can be retacken while e4 doesn't.

Many a game has been lost while trying to save a pawn.

The second word of the QGA should be a clue to White's plan after 3.e4.

Did you know?

Most QG players WILL get angry if you hold the pawn. They are usually accustomed to holding it, even if it is a "gambit".

they should call it the queen's temporary gambit.