What percentage of opening repertoire moves should be memorized...

...before considering adding some more moves to it?

Update 5th August 2023:

I started off with separate "repertoires" for each opening either that I was interested in learning or that I had prepared for a specific opponent. Often it involved a lot of mainline theory that rarely came up in online games as opponent's would play sidelines or sub-optimal moves that came with threats which I didn't know how to refute. I would learn a few openings in great depth and know exactly what to play on move 12, but not even know what to do on move 3 if the opponent played some unusual moves.

My rating is all over the place. I think I am under-rated at Rapid here and more often than not I have at least a slight advantage. I don't care about my blitz / bullet ratings. Over the board, my rating is higher and I play against OTB opponents up to the 1800s. I get comprehensively crushed (in all stages of the game) by 1800s OTB but I still play to compete and aim for a positive result not just to have fun.

So openings in OTB games are much more theoretical battles than online live games where I see a bit of every strange thing imaginable. Then there are the daily games where opening books are allowed and often the games recreate a Super GM game up until move 15 or so. That experience is useful for learning "state of the art" openings but it is completely unrealistic to memorize moves to that level.

Recently I have created a complete opening repertoire for white and another complete repertoire for black using Chess Book.

Briefly the process is:

• Choose you rating level
• Find the biggest gaps and select from a list of recommended moves which are based on your level
• Continue the process until 100% complete

There is a bit of description on theory, for example at the end of a line it will show a page with some typical ideas (like bring the bishop to d6 and castle Queenside). But it is mainly about only learning positions that actually come up in real games at the given level.

My repertoire avoids the most complex openings such as the Najdorf.

Whenever I drill the openings I find positions where my chosen move is very unnatural and think "I would never play that there" and then look at the different options and what there evaluations and win-rates and often add an alternative move which I find more natural if it is playable. Surprisingly if you are told there are two moves in a position it can be easier to remember both of them than it is to remember one move in a position. This is because I know that there is one natural looking move and one unnatural one.

But adding an alternative move means there is a whole lot of subsequent moves to add for all those different lines. And every time I practice openings I add more moves than I actually memorize in the session.

The last time I got a message saying:

"Practice complete!
Your stats from this session:
You practiced 26 moves
You played the correct move 10 times (38%)
You have 1424 moves due for review now!"

At this point I realized the ridiculousness of this and logged off.

I’d say it depends how much is in your repertoire. But I think you should always try to memorise 100% of your repertoire otherwise there isn’t really a point of having extra moves in a database or whatever 👍

As much as you can truly know without guessing. And you know for a fact the reason why each move is played. Then memorize one move at a time after that till you understand the whole line.

Even Super GMs do not have 100% of their full repertoire completely memorized. You frequently hear players admit after a game that they reached a point in the game with an uncommon position and they could no longer remember their prep.

Depends on the theoretical standing of your openings - in some openings, like the mainline 6.Bg5 Najdorf, you are very badly prepared if you only remember up to move 15, whereas in something like the Gurgenidze Caro-Kann, having some theory helps but it is more important to understand what is happening in the structure than memorising exact lines. 
The goal is always to know more than your opponent, so generally you should try to find an interesting idea and follow it until you've gotten most of your pieces developed and have castled.

Depends on your level and what openings your play. If you play the Caro-Kann (and if I'm not mistaken you do play it) then you don't need to know a ton of theory as the positions tend to be no so sharp. But if you play the Anti-Moscow Gambit in the Semi-Slav which is extremely sharp and one inaccuracy could lose the game, then you need to know a lot of theory if you don't want to be dead lost by move 20. As for level, at 1200-1500 a lot of people don't know much theory (Even at 2000-2100 a lot of people don't seem to know much theory), so you can get by just fine without knowing to much. I don't know what your theory knowledge is like, but as long as you're not getting blown away in the opening and reaching playable positions then you should be fine and will only need to memorize more when you get to a point when you're getting bad positions out of the opening. Sorry if this was a bit scatter brained, but hopefully this helps!

I started off with separate "repertoires" for each opening either that I was interested in learning or that I had prepared for a specific opponent. Often it involved a lot of mainline theory that rarely came up in online games as opponent's would play sidelines or sub-optimal moves that came with threats which I didn't know how to refute. I would learn a few openings in great depth and know exactly what to play on move 12, but not even know what to do on move 3 if the opponent played some unusual moves.

My rating is all over the place. I think I am under-rated at Rapid here and more often than not I have at least a slight advantage. I don't care about my blitz / bullet ratings. Over the board, my rating is higher and I play against OTB opponents up to the 1800s. I get comprehensively crushed (in all stages of the game) by 1800s OTB but I still play to compete and aim for a positive result not just to have fun.

So openings in OTB games are much more theoretical battles than online live games where I see a bit of every strange thing imaginable. Then there are the daily games where opening books are allowed and often the games recreate a Super GM game up until move 15 or so. That experience is useful for learning "state of the art" openings but it is completely unrealistic to memorize moves to that level.

Recently I have created a complete opening repertoire for white and another complete repertoire for black using Chess Book.

Briefly the process is:

• Choose you rating level
• Find the biggest gaps and select from a list of recommended moves which are based on your level
• Continue the process until 100% complete

There is a bit of description on theory, for example at the end of a line it will show a page with some typical ideas (like bring the bishop to d6 and castle Queenside). But it is mainly about only learning positions that actually come up in real games at the given level.

Never used Chessbook before so I'm not sure if you're able to, but if you'd like, you can send me your opening files and I can try to go over them and see what should be the must know theory and what you can wait till later to memorize. You said that you get crushed by 1800+ players in all stages of the game (so I assume that means you're getting bad positions from the opening), so I'm curious what they play against you, like is it a specific opening a lot of them play that you're uncomfortable against, or is it just that they know more?

I just play a Sicilian game for the firts time in years. I was aiming at the O´Kelly + Kan to avoid all heavy theory. I was surprised in move two by 2.f4. And there are many more 2nd moves like 2.c3 etc. It is the same in every opening/defence complex. I think it is a waste of time to learn more than 4-5 moves but cover most variations. Add 3-4 normal follow up moves and a few word on where to be active on the board and which side to castle normally. Also, chose openings/defences7variations that is not too comprehensive. Fort Knox in the French is a good short cut; there are others.

What these people are saying is not good. Study your openings to the last! 21 moves deep, you will roll them.

You have not memorized your openings 21 moves deep, so why are you recommending this? Do you understanding how much work it is to learn all lines all the way up to move 21?