Two weeks of logged practice changes the picture more than two weeks of daily play.
Two weeks of daily play produces about fourteen puzzles and a vague sense of progress. Two weeks of logged practice produces something much more useful: a record of what actually breaks.
The log does not need to be elaborate. Four columns are enough: the starting word, the move count, the point where the plan failed, and the one word that would have kept it alive. After twenty entries the pattern is usually obvious, and it is almost never the pattern the player expected. Most people believe their weakness is vocabulary. Most logs show it is the first ten seconds.
The second finding is usually about pacing. Logged practice shows how often the first move is made without a pause, and how much difference the pause makes when it happens. That is a concrete improvement available immediately, with no new knowledge required.
The third finding is about difficulty selection. Practice sessions tend to drift toward either too-easy boards, which feel productive, or too-hard ones, which feel meaningful. The log shows which, because the failure points cluster in a way that makes the difficulty visible.
The fortnight is the right length. A week is too short for a pattern; a month is long enough that the log stops being opened. Two weeks is a complete experiment and a small enough commitment to finish. The daily ladder keeps running alongside, unchanged, as the control. The daily puzzle is still the test; the log is about what happens between tests.
The log also makes a good baseline for later. Two weeks of entries is a snapshot of a specific player, and the same player can produce another snapshot in a month. Comparing two fortnights is far more informative than comparing two individual solves, because it removes the daily noise.