training rating and playing rating

Is there any relationship between training rating and classical rating?

No
But it is all the recent rage to make graphs, charts, formulas and predictions that "relationships" exist in chess between ratings and statistics of all kinds. Well, of course some do, but they are meaningless, have no substance. Play around with numbers. It's entertaining, but don't believe a word of it.
Simply because there are too many variables, much in the same way chess games are played differently by all, skill sets, interests, habits etc. cover a wide spectrum.
Example: I have little interest in "training", spend but moments on puzzles. Others take them seriously, do dozens daily, may take a short or long time on solving them. Classical ratings at lichess are 8+ minutes to unlimited time.

Any such relationships found would require a tight control group, playing a 1000 games and doing a 1000 puzzles, all with well defined parameters. This simply does not happen at online chess. Data can not be used that is taken from such a wide array of sources and expect accurate predictions.

@chaparrito77

I found 71,820 Lichess players who have played at least 100 Classical games and attempted at least 100 puzzles.

The full plot of those data points can be found here:

https://imgur.com/a/h24xx

In that sample, the correlation between puzzle rating and classical rating is 0.74. Quite strong, but not perfect.

On average, an increase of 1 point in puzzle rating is associated with an increase of 0.7 point in classical rating.

Of course, this doesn't mean that if you do more puzzles, your classical rating will necessarily improve. The causality can run in the other direction...

Tyvm for the info dudeski.

Have a nice day!

Is there a correlation between the correlation between two variables and the number of posts @mdinnerspace makes refuting it?

@tailuge attempting to refute, attempting.

A correlation can be found between any two events or things. Statistics are a wonderful tool in this respect.
The point is do the results give any useful information.
The data from chess statistics is seen at both ends of a wide spectrum. The numbers are often from very few samples. The "base" used is uncommon.
To then average these numbers, find a median, and make conclusions about individual results is the error, not the "math" involved.
Example: Statistics show that "On average, an increase of 1 point in puzzle rating is associated with an increase of 0.7 point in classical rating." A great stat. An average of all the data. But what does it mean for players? Unless they fit the profile, it has little meaning. The majority of players will fall outside the median by a substantial degree in one way or another.

Don't take it so seriously. I was just teasing.

dudeski_robinson edited14 hours ago #3
@chaparrito77

I found 71,820 Lichess players who have played at least 100 Classical games and attempted at least 100 puzzles.

The full plot of those data points can be found here:

https://imgur.com/a/h24xx

No data is included when the training puzzles were taken. I see many profiles of players who after 1st sign-up, complete dozens+ of puzzles, before playing a single game. Others just the apposite. We see topics all the time wondering about the "training", how to best solve puzzles and the rating system employed.
How about this correlation. The higher your rating, if due diligence is used, the higher your training rating. Will it be higher (practice them often) or lower (not so much.) Will it be the same? Ask the statistics folks.

#8
"Don't take it so seriously. I was just teasing."

I think that we can all agree that this correlation is pretty interestong, but also pretty meaningless.