Is 1.f4 a playable opening?

I have faced 1f4 only four times over the board in normal time controls in decades. Losing two and drawing two. I played it for a season or two, as white, with just over a fifty percent record. Around 45 games. This may be a normal record in club chess when your grading will continually adjust your position in a team and that team's position in a league.
Despite the seeming objective weakness of 1f4, its rarity may make up for this. Who is going to prepare against 1f4? If someone chooses 1f4 and learns the theory especially of the From Gambit, and is prepared to play it for years, it could become a successful ecological playing niche for someone...

From the point of view of practical play, there is much to recommend rarer, mildly sub-par openings IMO.

1) c4 is a better move than 1) f4.
But that doesn't mean they're not comparable.
White refrains from touching his center pawns or his g-knight on move 1.
But c4 and f4 both hit the center fairly hard.
c4 helps the white queen. f4 is more like a middlegame move.
Which is one of the reasons 1) c4 is better and more popular.
------------------------------------------------------------------------------
Neither move seem like good moves to start beginners/novices off with though.
Having said that though - there's a lot of paradox in chess.
If one makes it a point to not play those moves on move 1 with white ...
When playing black you will still find your opponents playing those moves against you!
Including on move 1 !!
----------------------------------------------------------
Suggests another idea that could perhaps be a Top Ten idea when showing openings to beginners/novices.
Try out prioritizing how to play black instead of how to play white.
By improving on how to play black - the student will tend to better grasp the issues for both sides.
That's provided that the student is looking wide - rather than deep.
If looking deep instead (Sicilian Najdorf and the Breyer lines of the R. Lopez) then that doesn't work.
Paradox: Present an overview of the openings or present just one opening?
The second doesn't look realistic. Even one opening has multiple lines.

the best move is always 1. e4 e5 2. Ke2

I was beginning to like 1 )e4 e5 Qh5 the last couple of days.
Hikaru says its playable.
According to an online database though -
after black responds Nc6 Bc4 g6 ... Black wins the game from there more than white does.

Ke2 still better

Qh5 is proven to be playable by Stockfish 16.

Black isn't compelled to play 4) - f5. That looks very arbitrary and unnecessary.

Stockfish 16 notes it as the best move.

I tried it here.
https://www.chess.com/analysis?tab=analysisso far it likes Nf6 better for black's 4th move than f5.
I don't like f5 there at all.

That was my mistake. Sorry

At what point did Stockfish take over, and how much time did it have per move? I'm just asking because I've entered the position after White's 4th move into the CM 9000 (a very much older engine, but as I have it that makes it easy for me to have a peak at it). And it says 4. ... f5 looks inferior to 4. ... Nf6, with the first tilting in favour of White while the second tilting in favour of Black when I play each move and let it analyse for White's 5th move.

But, if I just let it sit at the position to choose a 4th move for Black, although the ChessMaster 9000 finds 4. ... f5, after a short while (not right away), it only appears on the list as the 3rd best move. At the point 4. ... f5 does appear the CM scores the top line for Black as 4. ... Nf6 at -0.23, and the 3rd best move 4. ... f5 at 0.15 (so favouring White). In addition, if I leave it longer, 4. ... f5 drops off the top 4 move list for Black altogether, but from the beginning 4. ... Nf6 remains as the top choice, although the score drops to just -0.13. In fact, over the whole time, the only move it finds in the top 4 that continues to favour Black is 4. ... Nf6, and it finds that at the start and never changes from it, although the score does move towards equality.
I'm not suggesting the CM is better than Stockfish by any stretch of the imagination, but just curious about how much time Stockfish had to evaluate the moves, and at what point it "took over" the game? I've never seen anyone suggest 4. ... f5, but computers think differently than humans, so that's not saying much.

No problemo at all.
There are people here who Never admit a mistake ...

And Angus found exactly what I found.
And more.
But just one look at f5 there and I thought 'that Can't be right! It just can't be.'

Ah, ok! I was doing the analysis and missed this. No shame in making a mistake. If I had to list every mistake I've made in chess, I wouldn't have time to play any chess at all!

Yah, it didn't look right to me either, which is why I wanted to have a look. Computers can come up with weird moves though, so while my guts were disturbed, my brain reminded me that Stockfish sees things I can't even imagine.

Stockfish is very useful.
And generally when it finds positions to be about even or close enough ...
then one could accept that it has found no 'tactical refute' for either side.
But f5 is so bad-looking there that I wouldn't even trust an 'about equal' finding there.
Stockfish isn't always infallible about tactics either.
It has a record of finding positions to be draws that are wins or was it vice versa?
But the point was - it was obvious to humans of most strength levels looking even briefly at the positions that the computer was completely wrong.

Thats my bad I am only rated 850 so I dont have a trained eye, I wouldve played that move myself probably

You could probably adjust that 850 very fast if you start doing tactics puzzles here.
Not openings.
Tactics.
Suggestions to nobody in particular:
Don't be concerned about ratings nor tactics ratings nor tactics rating points. 
Don't concern yourself with the timer either.
I do the tactics puzzles unrated and with no timer.
If I'm not making progess against a puzzle I admit it to myself and go ahead and choose a move on principle and experience.
Instead of trying to crunch all the variations like a computer.
If I get it wrong I get it wrong.
------------------------------------------
Point: Tactics puzzles and basic endings and basic checkmate positions are all 'solved' parts of chess. And I do mean 'solved'.
So as you begin to understand them they begin to become your property.
----------------------------------------
In other words you begin to make progress. Including against openings.
Openings are not 'solved' though. They might never be. Even with supercomputers.
What I'm saying here is a paradox.
I'm saying you can 'adjust' that 850. 'Can'. Not 'should'.

And I'm also suggesting - don't concern yourself about ratings.
As you improve - rating points will follow you around - rather than you chasing them.
But that's a general figurative 'you'.

Nothing wrong with that, it takes time. Just keep basic opening principles in mind, one of which is don't move too many pawns, and another is don't move the same piece more than once, which are probably the two to focus on here.
The thing about 2. Qh5 that makes it tricky early on is that it does a couple things. First, it attacks your e5 pawn, and second it also targets the weak f7 pawn (which is only protected by your King). So, it includes a fork of sorts, although obviously White can't take your f7 pawn just yet. But, if (and when) White brings their Bishop out to c4, now they will have their Queen and their Bishop eyeing that weak spot in the opening position. For both players, the King Bishop Pawn is a weak spot in the line of pawns, because of how only the King defends it. White is trying to exploit that weakness here.
So the Queen is being annoying, threating to become part of a pile up on your weak f7 pawn, and also threatening to just take your e-pawn.
Black deals with that immediate e pawn threat by 2. ... Nc6. That protects the e-pawn and simultaneously develops the Knight to a natural square for it.
Then White plays 3. Bc4, and they are threatening mate on your f7. So that must be dealt with; you need to stop White from being able to play Qxf7. You could try Nh6 to protect f7, or g6 to both attack the Queen and disconnect it from f7, or d6 to disconnect the Bishop from f7. The first puts the Knight on a bad square, and while the Knight does protect f7, the two attacking pieces are still lined up in a scary way. And as soon as White moves their own d pawn, your Knight is going to be attacked by the White dark squared Bishop. If that Knight falls, you're back in a world of hurt. The last option, cutting off the Bishop, isn't great since White can just take your d-pawn with their Bishop, and you're still in the same problem. That really only leaves the middle option, g6, attacking the White Queen and cutting it off from your weak f7 square.
And, you are forcing White to move their Queen a second time (you are now forcing White to violate one of the opening principles!).
White will probably move their Queen to f3, re-establishing that double attack on your weak f7 square, so again threatening mate. But f3 is natural developing square for the White Kingside Knight, so the Queen is in the way of White's development. (This is one of the reasons why this attack isn't really all that great - White is violating lots of opening principles, and Black thwarts White by making moves that are both good developing moves and when ready will try to make White have to violate those principles more).
And we're at the position in question. Sure, f5 does prevent that mate, but it is another pawn move on your part, and a pawn that is normally better off where it can help protect your King if you castle that side. On the other hand, Nf6 let's you capitalise on the tempo you gained when White moved their Queen a 2nd time. It also prevents mate on f7 by cutting off the White Queen's access to it. And the Knight is protected by your own Queen, so it's safe there. It also attacks White's e-pawn, so while you can't take it now, you give White something to think about. You can also choose to fianchetto your Kingside Bishop with Bg7, allowing you to castle quickly (which develops your Bishop, and castling is another developing move). And the final thing to note is that your c6 Knight could jump to d4, where it would attack White's Queen on f3, forcing it to move yet again! Moving the c6 Knight would mean moving the same piece twice, which is not a great idea, and it would also mean forcing White to free up the f3 square for their Knight, but there's no need to help White "get it right", so at this point it might be best to consider Bg7, followed by O-O. If, during that, White puts their Knight on e2, and O-O, meaning you can't just play Ne4 to attack their Queen, not a big loss, but keep that idea in mind (like, if they play Nc3 it might not be such a bad idea).
Anyway, there are some lessons on Chess.com that go through how to deal with these early Queen attacks by White. And, the Nelson bot is a really good training tool to practice what you learn there. Nelson is programmed to play these sorts of early Queen lines, and I promise you, once you get the defence down and get past the first six or seven moves, Nelson just sort of falls over, but if you play it a bunch of times, you will very quickly develop your eye for these positions. Learning how to defend against 2. Qh5 is really just putting those opening principles to practice, and that will help to transform them into a list of "rules" into something more useful, your first steps to positional understanding.

After Qf3 Nf6 I'm wondering if there's still a 'rat' in the position.
Black has weakened his f6 square pretty substantially with both e5 and g6 played.
But maybe white's pieces are just not poised enough to exploit it.
White's d-pawn isn't even moved yet - so white cannot 'leap in' with Bg5 right away.
But can white get that in soon and paralyze three of black's pieces on 'dark squares' ?
There's a bit of a 'whiff' of that there ...