How deep do you prepare?

So I'm interested in how rare of an opening you will prepare for. Everyone is prepared for white to play 1. e4 or 1. d4 because they happen so often (1.e4 might be played in more than half of your games). On the other hand you have some openings that are rare, like white might not prepare for black to play 1. e4 e5 2. Nf3 d5!, because it happens less than 1% of games. 

What's your cut off for 'to-rare-to-prepare-for'? Is 1% too rare? or do you cut off at 5%?

I have my pet openings that I know some theory for but for the most part I just follow principles and hope I don't get trapped in 8 moves

I always felt it was near impossible to prepare seriously as White against the Dutch and the Modern Defence. Too many different systems, too many different move-orders and move-order tricks, and these openings are not strong enough or frequent enough to justify investing so much time countering them; meanwhile they're not bad enough so even if you prepare a lot and manage to remember everything you're not sure to get a clear stable advantage in all the lines. Of course that means losing badly on occasion against an expert who knows these systems very well, but it is statistically unsignificant (and it's not even sure your preparation will match his experience anyway). So the sensible approach is to study a tricky sideline but since everybody's doing the same these lines eventually lose their bite.

On the other hand it is useful to know some detailed refutations of the Englund, the Vulture, the Clarendon court, some bad Benoni lines and such because you can get free points with just moderate opening homework.

Do not prepare at all. Just play. Analyse your lost games.
If you lose a game as white from 1 e4 e5 2 Nf3 d5, then after analysis you will know how to play that if it ever happens again. Because you have actually played it, you will remember.

I'm prepared for 1. e4 e5 2.Nf3 d5. If I can't refute utter crap like this then what's the point of trying to get an advantage against good openings?

#6
You cannot get an advantage against a good opening.
It is pointless to prepare for 1 e4 e5 2 Nf3 d5. If you depend on preparation to win a won opening position, then how will you ever win a won middle game position?

#7 I know, but this doesnt mean I can't try! What do I have to do? Offer a draw after 1. e4 e5 and 1. e4 c5? Refuting 2... d5 is really simple, I save myself some thinking time by knowing the theory and I make sure to punish it as much as I can, I'm not so confident to think I can come up with a better sequence than the engine

+1

with 1.Nf3 i try to transpose to a 1.d4 position, which most players are happy to do. And those who play the kings indian attack is gonna cry because the king's indian attack is doodoo water.

with 1.c4 I can get away with a symmetrical english, it is afterall the main drawback of the english.

#8
"Offer a draw after 1. e4 e5 and 1. e4 c5" ++ No way, black has many ways to err.

"I save myself some thinking time by knowing the theory" ++ But why would you spend study time to save thinking time against something that does not happen.

"and I make sure to punish it as much as I can"
++ You should be able to punish it without recourse to memory.
Why spend study time on 1 e4 e5 2 Nf3 d5, 1 e4 e5 Nf3 f5, 1 e4 f5, 1 e4 b5 etc.

"I'm not so confident to think I can come up with a better sequence than the engine"
++ It is not about a better sequence, it is about a sequence that wins.

Make an excel file of openings and let it choose your opening repertoir

In this game from a country vs country match (Malaysia vs Canada), my 28th move was a prepared novelty improving on Ljangov's 28. ... Kc6? from the game Sznapik vs Ljangov, Sandomeirz Poland, 1976.

We were still in my home prep when my opponent resigned on move 33.

A Heroic Defense in the Sicilian Najdorf - Kids, don't try this at home! - Chess Forums - Chess.com

I can't say I completely agree with this. Yes, perhaps there is no point to prepare it if you have not seen it before, but once you have seen it, then it doesn't hurt to find a way to play against it so that you know how to play against it next time. 

It is good to prepare to the point in which you understand that White has an advantage. But if you're preparing beyond the point of White having a clear extra pawn, then that may be too excessive.

True principle, but not a very good example. The elephant gambit can be incredibly trick and it is important to know at least some theory if you are planning on playing e4.

I’m prepared to play chess. I don’t prepare for opponents. And they can’t really prepare for me, either, considering that my  favorite responses to e4 are Nf6 and Nc6.

At most, I will review my lines before the tournament 

Offbeat openings are offbeat because they aren’t as good as the mainlines. To beat them, you simply have to play better chess. That’s why I’ve always liked them. You have no one to blame besides yourself if you lose, but winning is all your own effort.

@blueemu that's the best game I've seen on this site, I need to learn to defend like this. I imagine it's even more thrilling to perfectly defend against an onslaught OTB and slowly suck away the bravado of the attack with 28... kc5 then it would be to be on the offense and win. I've never won a game in such a way...

Thanks for the kind words.

My play wasn't perfect, of course... 20. ... Kf8 (pointed out by IM pfren) would have been MUCH stronger than the 20. ... Rad8 that I actually played... but it's a pretty good example of defending under tremendous pressure, yes.

just Staunton tho?