We told one player the same thing 57 times

After every session you play, Eloist sends a report. We were proud of how fresh they looked: across 5,361 sends the subject line was unique 54% of the time, because subjects are built from your own games. “2 of this sitting’s losses were on the clock”. and your chess never repeats exactly.

Then we counted something less flattering. Underneath the fresh sentences, the report could only diagnose five things: openings, clock, tactics, endgame, tilt. One player received the openings lesson 57 times. Eighty-one people got the same lesson five times or more. Every letter was worded differently. Every letter said the same thing.

Fresh words are not fresh findings

The failure has a shape we now look for everywhere: a generator with good variety in front of a supply with none. Rotation, phrasing, subject lines. All of that machinery was working perfectly, and it was varnish on five planks. No amount of better wording fixes a vocabulary of five.

So the fix was never ‘write better emails’. It was: give the diagnosis more to say. The engine already measures which piece plays your worst move, when in the game you collapse, how your clock behaves move by move, how this month compares to your own last year. None of that reached the lessons. Now it does. And a report that has nothing genuinely new to say about you says so, plainly, instead of re-dressing last week’s finding.

If you have ever felt a product repeating itself at you in different words. This is probably why. Count the supply, not the phrasings. We keep that sentence pinned above the desk now.