Yes, it's subjective. "??" less so. "#" not at all. But beyond that, you've got a lot of scope for interpretation.
2.f4 fully deserves a "?!" in a world championship match. At your club, such a declaration would be ludicrous.
If I were playing Magnus Carlsen, 1.e4 would probably still get me a "??", because as far as it mattered practically, every continuation would lead to forced mate.
If I win a pawn against a player similar to my rating, or lower rated, I can usually "hold the position". "Holding the position" is another phrase that grandmasters use, and which I translate as: My opponent can't get a winning advantage. Now, I can also use that extra pawn to give back for an even better position. Or, I can trade pieces to a winning endgame, making each simple trade more difficult for my opponent. The extra pawn creates psychological pressure on the opponent, which sometimes leads to errors.
So, I guess I'm saying, yes, being a pawn up, can decide the game. Depending upon the position, losing one pawn could be a "??" even at my level.
If I win a pawn against a player similar to my rating, or lower rated, I can usually "hold the position". "Holding the position" is another phrase that grandmasters use, and which I translate as: My opponent can't get a winning advantage. Now, I can also use that extra pawn to give back for an even better position. Or, I can trade pieces to a winning endgame, making each simple trade more difficult for my opponent. The extra pawn creates psychological pressure on the opponent, which sometimes leads to errors.
So, I guess I'm saying, yes, being a pawn up, can decide the game. Depending upon the position, losing one pawn could be a "??" even at my level.
Now that is quite clever. Giving back the pawn when it improves one's position. I like it.
I see that you like this: ??????????????, which required supersizing my computer screen to see all 14! A really big blunder? 
John Nunn says that at the GM level, even though in theory one pawn should prove decisive, in practice it usually takes two.
It's tricky nailing down a simple definition of the word blunder. I thought it was merely the loss of material. Some authors break down blunders into two categories: positional and tactical.
A little Wiki and a little from Kebu Chess Blog:
In chess, a blunder is a very bad move. What qualifies as a "blunder" rather than a normal mistake is somewhat subjective. A weak move from a novice player might be explained by the player's lack of skill, while the same move from a master might be called a blunder.
The favored explanation for blunders is a “lapse of concentration”. In chess, a lapse of concentration is usually preceded by a heightened emotional state of some kind.
---
Nah...for me, blunders don't stem from some emotional state. It's just being stupid at the wrong time.
(!?) = Interesting, (?!) = Dubious, (?) = Bad, (??) = Blunder
Isn't some of this subjective? Isn't some of this relative? Players will voice strong opinions over certain lines and gambits. What is dubious to one is interesting to another.
I've had players tell me that the King's Gambit is dubious, "Black can always and easily reach equality after 1.e4 e5 2.f4 (?!)". I'm not looking for equality in the KG, especially at my level. I see 2.f4 as a good move, or (!). It creates an imbalance and a problem for Black to solve right away, with excellent winning chances for both sides.
After 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6, the move 4.Ng5 was considered somewhat dubious, as a piece is moved twice in the opening, and 4.d3 still remains the most popular choice. With the advent of strong chess engines and larger databases, both 4.d3 and 4.Ng5 are evaluated as equally good, with 4.Nc3 and 4.0-0 coming in a close 3rd and 4th.
When I hear the word blunder, I think of dopped pieces, i.e., "he just blundered a pawn". Perhaps titled players see dropped pawns as grave errors, or blunders, (??). A loss of one pawn can prove decisive at the higher levels. Do under 2200 players see dropped pawns in the same way, or do they annotate those mistakes as bad (?)
I loathe hanging center pawns in the opening. It makes me mad and it sucks, always! In my private annotations, (??) does not cut it, so (****!) seems to better express my feelings of stupidity.
GM Roman D. will say, "that's just a very bad move". Well, o.k., but when I run that move through a chess engine, the evaluation doesn't swing wildly in favor of the other side, nor does it change by one pawn. To me, the move looks perfectly playable. I understand that Roman is looking at the positional soundness of a move. But for me, if it's not a tactical error, then the move would work at my level, so it's not bad, and I may find it interesting (!?).
Last, what may be interesting to some (!?), may be downright dumb (?!) to others.
What do you think?
??? is also a symbol
Sorry, I fixed it
The way I've come to know these annotations is like this:
! = Good move
!! = Brilliant move
!? = Questionable move, probably good.
?! = Questionable move, probably bad.
? = Bad move
?? = Blunder
??? = WTF?
???? = Are you F'ing kidding me?
????? = You're definitely going to lose
?????? = Do you really want to trade your queen for a pawn?
??????? = Best to take up a different hobby
???????? = No wonder your rating is under 1000
When I annotate my own games, I don't give my opponents move marks, just my own. To me, the symbols break down to something like this:
! -- good move that really shows I understood the position
!! -- good move that finds something devistatingly cool or aesthetic in the position. I don't have many of these
!? -- I had a good idea but it probably wasn't the best in the position and might show a lack of appreciation for positional or tactical aspects of the position
?! -- A bad move played for all the right reasons
? -- a bad move
?? -- I'm throwing away something critical that changes the game from winning to lost, or from even to lost, or from lost to resignable.
In all cases, I try to articulate why I gave the evaluation in natural language, and I focus in on the ?! and !? marks to help me address training deficiencies.
Yeah, most of it is purely subjective, and defenitely depends on the level of players. I've always liked the "Blunder" entry in Oxford companion. A blunder discovered a long time after the event:
yeah, about those, I played my chemistry teacher earlier today, and the worst move I think he made was 2.f3 after 1. d4 Nf6
However at the end of the game he was down 2 whole Rooks, forced to trade his Queen off which favored me, and I Had passed pawn on the sixth. and an active King. His position was full of holes, and the only chance he had of winning I saw, and Defended acurately against.
(!?) = Interesting, (?!) = Dubious, (?) = Bad, (??) = Blunder
Isn't some of this subjective? Isn't some of this relative? Players will voice strong opinions over certain lines and gambits. What is dubious to one is interesting to another.
I've had players tell me that the King's Gambit is dubious, "Black can always and easily reach equality after 1.e4 e5 2.f4 (?!)". I'm not looking for equality in the KG, especially at my level. I see 2.f4 as a good move, or (!). It creates an imbalance and a problem for Black to solve right away, with excellent winning chances for both sides.
After 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6, the move 4.Ng5 was considered somewhat dubious, as a piece is moved twice in the opening, and 4.d3 still remains the most popular choice. With the advent of strong chess engines and larger databases, both 4.d3 and 4.Ng5 are evaluated as equally good, with 4.Nc3 and 4.0-0 coming in a close 3rd and 4th.
When I hear the word blunder, I think of dopped pieces, i.e., "he just blundered a pawn". Perhaps titled players see dropped pawns as grave errors, or blunders, (??). A loss of one pawn can prove decisive at the higher levels. Do under 2200 players see dropped pawns in the same way, or do they annotate those mistakes as bad (?)
I loathe hanging center pawns in the opening. It makes me mad and it sucks, always! In my private annotations, (??) does not cut it, so (****!) seems to better express my feelings of stupidity.
GM Roman D. will say, "that's just a very bad move". Well, o.k., but when I run that move through a chess engine, the evaluation doesn't swing wildly in favor of the other side, nor does it change by one pawn. To me, the move looks perfectly playable. I understand that Roman is looking at the positional soundness of a move. But for me, if it's not a tactical error, then the move would work at my level, so it's not bad, and I may find it interesting (!?).
Last, what may be interesting to some (!?), may be downright dumb (?!) to others.
What do you think?