The overwhelming empirical evidence that basic chess principles translate into higher Elo ratings

"move your queen twice"

@Aonrathon said ^

This doesn’t show what you want it to, it shows that higher rated players adhere to orthodox strategy, not that orthodox strategy leads to a higher rating.

The data shows that there's a correlation between adhering to orthodox opening principles and a higher rating. Strictly speaking it's not possible to infer cause-effect-relationships from observational data such as this - ideally you'd need an experimental intervention with a control group to show cause/effect.

I've taken enough graduate level econometrics and stats courses (and even taught undergraduate stats back in the day) to understand correlation is not the same as causation. Sometimes real life IS the experiment though, and in this case that is the millions of games that are played and recorded on lichess.org. With the exception of experimental economics where tightly controlled experiments are run in a lab hoping to capture some aspect of real life (which has it's own issues), pretty much all the data econometricians deal with are of this nature. Social and physical sciences often have the luxury of creating data from carefully controlled experiments. Getting back to chess. If I'm learning and someone explains the concept of developing quickly so (among other things) you can connect your rooks to protect each other and attack as a team, and then hands me the results from a large stratified sample showing the table below where:

1. In games played by low level players, one or more players connects their rooks 37% of the time

2. In games played by higher level players, one or more players connects their rooks 82% of the time

3. For the ratings in between it's a steady increase

4. Regardless of rating, if I connect my rooks and my opponent doesn't versus my opponent connects and I don't,
   the win percentage gap is double digits

what's my take away as a player learning the game?

1. I should probably make an effort to connect my rooks more if possible. 2) I wonder how often I'm currently doing it in my own games? i.e. how does my play stack up against players across the Elo spectrum in the regard. What's my grade?

What are everyone else's thoughts on this?

@Rookeroosky said in #33:

... the results from a large stratified sample showing ...

Not from a random sample. If a study shows that right-turns are likely to be better than left-turns, should one base the decision on the study or on the specifics of where one is and where one is trying to go?

"... I would caution you that this and other maxims are not to be blindly followed. In chess, as in life, rules must often be swept aside. ..." - Logical Chess (1957) by Irving Chernev

As a not too strong player, I try to follow all these principles with the possible exception of not bringing out the queen early (only because I play the Scotch with white and often see 4...Nxd4 5.Qxd4).

If I don't follow these principles, usually it is because I miscalculated and ran into trouble early. For instance: I think it is good to castle, but I also think I lose if I fail to castle not because I intentionally delayed castling to pursue some other idea but because of a mistake which compromised my position such that castling became impossible or undesirable. The root problem isn't ignorance of or deviation from general opening principles, but miscalculation of the specific position.

Stratified is not a dirty word. Stratification was used to create a large balanced sample with the goal of having equal sized random samples for each bin. Details are here for anyone truly interested: https://www.elopluschess.com/static/docs/research/en/lichess-stratified-benchmark-methodology.pdf.

I think the people who have seniors draw a clock before deciding whether or not to let them keep driving would disagree with the left turn example above. They would tell them right turns are absolutely better and safer in North America than left turns. If you leave the house once a week to get groceries, it's ok to make 3 right turns to avoid 1 left turn based on studies of accidents involving seniors.

Do not disagree with your quote. Chess principles are absolutely to be broken as needed. If you're facing a Fried Liver attack you don't just keep blindly trying to check off a set of opening principles and ignore the attack. They're still worth knowing and the data speaks volumes that emerging players can benefit from sticking to them more often than not.

ArcticPuffin, do you mind if I share an Opening Discipline panel showing your # of Queen touches before move 10? Last 10 versus lifetime, wins versus losses?

While there is a correlation overall, there are definitely sound and valid openings that move the queen multiple times early (Scandinavian Defense) and also completely legitimate openings that move the f-pawn early as well, such as the Dutch Defense or the Stonewall Attack. There are also a handful of openings where neither side castles.

I think the main difference is that the top players will "violate" the principles intentionally and with a purpose, while lower rated players tend to do (or not do) things randomly without a plan or a reason.

Personally, I'm not a big fan of dogmatically following checklists, such as "capture towards the center" or "castle early" because sometimes you'll realize the other pawn should have captured to open a different file, or you might sometimes wish you spent 2 more moves to castle to the other side. It's more important to always ask yourself why you are making a specific move, than to mindlessly follow a list of principles.

@Panagrellus said ^

This doesn’t show what you want it to, it shows that higher rated players adhere to orthodox strategy, not that orthodox strategy leads to a higher rating.

The data shows that there's a correlation between adhering to orthodox opening principles and a higher rating. Strictly speaking it's not possible to infer cause-effect-relationships from observational data such as this - ideally you'd need an experimental intervention with a control group to show cause/effect.

I understand that. Higher level players tend to be uncreative. The exceptions are exceptions

Beginners often play with pawns. Promote them early. At 6 or 7 horizontals they are lost. Or they lose it immediately after turning into a piece.
Experienced players create a structure of pawns and leave them standing in place. Only when the path is cleared do they begin to move their pawns.
Is it possible to study on what move pawns are lost? At what move is the transformation? How many moves does a new piece live?
Thank you.