The Loop — Missing Piece of Puzzle
Before I continue; Please be advised this is a personal project so it not just handles my day to day working stuff but handles my side routines too i.e. My readings, Tafseer lectures, Sports and so on.
So, no judging only takeaways the technical learnings :)
On 8 August my new tracking system told me a daily habit of mine had been dead for seven weeks.
It had not. I attend a tafseer lecture every morning before the day starts — sometimes 07:45, sometimes 05:30. I had not missed it once. The notes are on paper, in a notebook, so nothing was ever going to show up in my vault.
The system had measured note-taking and reported it as study.
So, now it has a Journal.md. The most important file in the whole system. More important than the dashboards, the fancy graph views etc.
This post is based on my LLM-Wiki, if you’ve not read that you might miss the folder structure so feel free to skim that as well
The Start
I sat down with Claude Code (Fable) and did a goals interview for the rest of 2026. Books, football, tafseer, writing, investments (if any), even the goal at work i.e. sastaticket.pk. One pattern came out of all of them:
Every goal I have is a consistency goal, not an achievement goal.
Nothing on the list fails because it’s too hard, and nothing fails at the start. In my own words, written down that day:
“I keep it up 1 month, then a 2-week gap, and it repeats.”
That is why todo systems never helped me, you just name a ToDo list app? There’s 9/10 chance that I might have used it already. I have a list of skills that handles my meetings, quarterly checklists, follow-up trackers — all excellent intake. They measure completion. Not one of them measured whether I kept showing up OR How am I doing as a person myself?
So every goal in the new system carries two fields and nothing else:
Goal 4 · Depth — learning
cadence: daily
signal: any new or updated note in 1 - Material/
The Sunday review reports streaks and gaps as numbers. No tone, no encouragement, no adjectives about effort. “tafseer: 51 days since last entry” is a complete sentence.
Football - The Design Lever
The one goal that was already fixed is the one worth copying. I restarted football after a six-year gap and it held — not through willpower, but because there was a club, a place, and a fixed time.
I learned this in Atomic Habits - James Clear, the system for a certain habit exist and despite the 6 years gap, I started it back.
Books show the same thing at smaller scale: I put the book on my working desk, and eventually I read it. System, not intention — exactly what concept - habit formation says and what I’d never applied deliberately.
Reading and tafseer have no club. The system is meant to be the club. Every feature since gets judged against one test: does this add structure, or does it add another thing to remember?
The Wrong About ME
The tafseer misread wasn’t a one-off. Seeded from vault state alone, the system also flagged three personal projects as stalled for 12–14 weeks:
Read as Actually
─────────────────────────────────────────────────────────────
3 side projects "stalled 14 weeks" Shipped, live, parked by choice
Tafseer "dark for 7 weeks" Daily and unbroken, on paper
Both times the vault was silent, and both times silence meant something other than inactivity. Unmaintained is the correct outcome for my lowest-ranked realm while the day job is this demanding — parking or holding something is a decision, not a failure.
The rule that came out of it is now binding on every line the system writes: say “no signal in the vault for N days.” Never “you haven’t done this in N days.” Those are different claims and only the first one is defensible.
Three Fixes in 48 hours
Ten days in, I was travelling for work and every personal cadence went red. I made three fixes in one day, each assuming a different broken mechanism:
Fix Assumed problem Result
──────────────────────────────────────────────────────────────────────────
Football → any physical activity signal too specific 0 sessions
Focus anchor: phone → laptop open anchor didn't travel no line, 2 days
Dump anchor → before closing laptop no fixed evening moment dump empty
Three different anchors, same outcome, inside two days. The obvious next move was a fourth fix — lower the target. That’s the same lever a fourth time.
The mechanism was never the variable.
The available third option was already sitting in the system and had gone unused: stop measuring, on purpose, with a return date. A park needs a reason and a date it comes back — without the second one it isn’t a park, it’s an open decision rotting quietly. Parks now live in one block that both skills read, and they expire on their own date. Nobody has to remember to undo one.
The Journal; Real Game
Here is the part I did not expect. The valuable artifact is not the streak count. It’s Journal.md — append-only, newest first, never edited:
## 2026-08-19 — The personal cadences are parked for the trip.
Three mechanism fixes in 48 hours all came back void.
Decision · Reasoning · Evidence · What was rejected and why
Rejected options get recorded too. Knowing what was ruled out is worth as much as knowing what was chosen.
The system reads that file before advising on any change — so a proposal that contradicts an earlier decision gets challenged with the reasoning, not silently reversed. When I reached for “broaden the signal” a second time, the journal said that exact lever produced zero sessions in 48 hours and was withdrawn.
No habit app has this. They all have charts. None of them have a record of the last four times you tried the same fix.
What I don’t know yet
The founding premise says I collapse around week five. The system is now past week five and still running — and I genuinely can’t tell whether that’s the structure working or just the usual window being longer.
That’s the honest state. The difference is worth knowing, and unlike every previous attempt, this time there’ll be a written record of which one it was.