I plan six hours of deep work a day and land about two, is six a fantasy for everyone or just for me

Partly disagreeing, because the ceiling is task dependent and lumping it into one number hides that. Writing something new from scratch, three hours is my absolute limit and the fourth is worthless. Debugging or reviewing code, I can go most of a day if the problem is interesting, because the feedback loop keeps pulling me back in and it costs far less to hold in mind. So the honest version is that your ceiling for the hardest category of your work is around three, and the rest of the day can still be real work as long as you match the task to what is left.

Three to four hours is the number that keeps turning up, both in the deliberate practice research people quote and in every honest time log I have ever seen. Two is not a personal failure, it is the low end of normal for genuinely hard cognitive work, and six sustained across months is something almost nobody does no matter what their productivity newsletter says. The trap in your setup is that the calendar block is six hours, so every day reads as a sixty percent failure even on a good day. Block three, protect them viciously, and let the afternoon be admin without pretending otherwise, and you will probably get more done because you stop spending the last two hours feeling behind.

I tracked mine properly for a year with a timer I had to start and stop by hand, which is annoying enough that you do not cheat. Yearly average came out just under three hours a day of genuinely focused work on working days, with a best week at about four and a half and a worst week near one. The pattern that mattered was not daily heroics, it was that the bad weeks were nearly always weeks with a context switch, meaning a new client or a big meeting mid week. Removing one context switch bought me more focused hours than any technique I tried.

Three good hours a day, every day, for a year, beats anyone's best week. Block three, bill honestly, stop measuring yourself against a number nobody hits.

Stop counting hours and count units of output, because hours are an input and you are optimising the wrong side of the equation. I switched to a rule of one meaningful thing shipped per day, defined the night before, and my hours did not change at all while my output roughly doubled. Most of what the extra hours were buying was polish on things that did not need it.