Best way to train openings for beginner?

Hi everyone,

I've been exploring a way to learn chess openings. I've developed a script that generates a PGN containing the most commonly played lines against a specific white move (from lichess db). The idea is to use this PGN as an input for some sites where you can train lines.

I wanted to get your opinions on this approach. Do you think it could be a useful tool in your own studies?

Also, if anyone is interested in getting these PGNs, please let me know! I'd be happy to share. I'd also like to know which specific openings you'd be most interested in.

Looking forward to hearing your thoughts!

What sites? Wouldn't they already have this?
I would be interested if you had this for specific persons, this could help.

@P7formula said in #2:

What sites? Wouldn't they already have this?
I would be interested if you had this for specific persons, this could help.

For example I found this site:
https://chesstempo.com/
Where you can import a PGN and repeat the lines in puzzle form. I'm not aware if lichess offers something similar but I think it's similar to some chessable exercises.

It's an interesting approach but fundamentally it has flaws that originate from the fact that following database lines has the following flaws:

1. Most commonly played line isn't always the best line. So if you don't use match results then you could play into a losing line.
2. Match results aren't always accurate since it's not unusual for players to resign in winning positions or lose a game for reasons completely unrelated to the opening.
3. Some main lines that statistically have a high win % for one side are abandoned when a refutation appears so database will only contain games from before the refutation was known. Databases don't give evaluations and even if you have one if you don't understand it knowing it is pointless.

It also has a couple of flaws in terms of learning:

1. Most effective way to learn an opening is by learning ideas and plans. Learning long variations without knowing why they're played is likely to make it harder to remember and also means you'll lack plans when opponents diverge from theory.
2. It's quite likely that following a database won't get you a structured repertoire. For example depending on whether you want to allow and avoid transpositions opening choices can vary drastically as for example 1.Nf3 c5 isn't viable if you want to avoid a Sicilian. In addition most players will want similar plans so that for example if playing against the Sicilian you'd want a similar system for 2...d6 and 2...Nc6 when there's a high chance the two lines transpose into each other.

Generally my advice for a beginner learning an opening is get a book or video you trust and you find easy to understand. Then absorb the main ideas and try it in blitz games. If there are lines where you regularly struggle then study them in a bit more depth and learn exact moves.

"It's an interesting approach but fundamentally it has flaws that originate from the fact that following database lines has the following flaws:

Most commonly played line isn't always the best line. So if you don't use match results then you could play into a losing line..."

Let's say I created this for Vienna game as white. In this case I'm only taking the moves for black from the database. The following white move is actually generated by stockfish/saved cloud analysis. So from white's perspective there shouldn't be any problems with landing in a bad spot. I also planned to maybe add some annotations to the generated PNGs for some typical blunders that people make.

I prefer books, but this can be a nice additional tool, just don't think beginners could make the most of it without further explanations, maybe more useful for intermediate players.

I've checked into the mentioned website and agree that your plan could work there, you could go through all the lines very fast and even remember the general movements. Yes, it is a good idea , but; if you had individual persons that could really help, perhaps you could be payed for it?

Pick very limited solid openings not buried in theory.

1st 6months focus on top 3 defenses against your chosen white opening. Then familiarize with the rest. Especially the clown gambits every youtuber is promoting. Be aware of them so you have correct response.

Use that database at/near your level to prepare for most common moves throughout opening into early middle game. There can be major differences at 1000 compared to masters database.

You are more likely to see convoluted Fried Liver, smith morra, or Alapin than someone willing to play open sicilian.

You are more likely to see Albin or Englund than a Meran or Tartakower against d4,c4. Dutch is very popular at lower levels and practically nobody plays Grunfeld, Benoni, or Queen's Indian.

So divide your study time accordingly.

@Zukerpie said in #5:

Let's say I created this for Vienna game as white. In this case I'm only taking the moves for black from the database. The following white move is actually generated by stockfish/saved cloud analysis. So from white's perspective there shouldn't be any problems with landing in a bad spot. I also planned to maybe add some annotations to the generated PNGs for some typical blunders that people make.

Well the Vienna isn't an opening I'd recommend you to spend too much time learning and it also illustrates some of the problems with your approach. The reason is that black has some responses that pretty much lead to a forced draw.

I'll also note that in the position after 1.e4 e5 2.Nc3 Nf6 the main line in the Vienna goes 3.f4 but unfortunately stockfish gives that an eval -0.2 of while 3.Nf3 is around +0.2 at which point you're transposing into the Four Knights Variation. Interestingly databases say 3.f4 has a 46.5%/22.7%/30.8% win/draw/lose rate while 3.Nf3 is 30.5%/31.3%/38.2% despite being objectively better. However if you then follow the main line a bit further to get to 1.e4 e5 2.Nc3 Nf6 3.f4 d5 4.fxe5 Nxe4 5.Nf3 Bc5 win/draw/lose rate drops to 29.7%/28.5%/41.8% so a position on move 3 where white wins 15.7% more games than he loses has flipped to one where he loses 12.1% more games than he wins despite white playing the best moves.

It can definitely have use as a training tool but I'm not sure it's a learning tool. If you can input the right lines (e.g. from an opening book) then it's a good training method. However I'm unsure if it's possible to extract best lines directly from a database even with engine assistance.

Play over many COMPLETE Games ... So you learn Everything or at least a little of Something each time' @Zukerpie