The continuity wall
Four walls so far, each about a moment: your list could not stay fresh, your screen could not see the market, your backtest could not reach 2008, your cockpit could not refresh. This one is about time, and it is the thesis of the whole track.
A shipped tool is a running cost, not a finished thing
The other four walls arrived when you asked for something. This one arrives when you stop asking — because a deployed tool keeps consuming whether or not you are there.
Three costs, and only one is money:
- Data. Every day it runs, at the prices you measured: 1 call per live symbol, 5 per screen, 10 per fundamentals request.
- Attention. About an hour a month, from the previous lesson.
- Decisions. Every advisory, every platform notice, every changed field is a small decision that only you can make.
Do the arithmetic once, properly
Interactive use: ___ calls/day (course 4's refresh budget)
Alerting: ___ calls/day (symbols x checks)
Monitoring: 0 (health costs no quota)
------------------------------------------------
Total: ___ calls/day
Your allowance: ___ calls/day
For the worked cockpit from course 4 the interactive line alone was 1,404 a day at five-minute refreshes, and the alert loop was another 936. Against a documented free allowance of 20 calls a day, the honest bottom line is that a continuously running version of this tool does not fit in the free tier — not degraded, not slower. It stops.
Fill in your own numbers. They will be different, and the exercise is the point.
The sentence the whole track was building toward
You now own something that works, that you understand, that you can defend, and that you would show to someone. Keeping it running is what the subscription is for.
That is not marketing, and it is worth saying why: you can now check it. You measured every price yourself. You know the four walls because each one broke something you built. You can compute your own daily requirement and compare it to a published number.
This is the difference between being sold something and pricing something. If your usage fits the free tier — and for a tool you open twice a day, it does — then it fits, and you should not pay for anything. The wall matters when the tool becomes something you rely on, and only you can tell when that happened. Live pricing is on the pricing page and deliberately not quoted here, because a number written into a lesson goes stale and then lies to whoever reads it next.
Retirement is a real option
Not every tool should be kept alive. A backtester used for one question and answered is finished, and that is a success rather than a failure.
What is not fine is the third state: still deployed, quietly broken, still spending. If you stop maintaining a tool, turn it off deliberately — that is a decision, and the slow accident is not.
The finance behind it
Sticking with a thing after the excitement has gone is its own skill: Why does surviving a drawdown change what you do next?
Try it now
Fill in the arithmetic with your own measured numbers and write the total in your README's limitations section. Then decide, explicitly: keep it running, run it manually, or retire it. Whichever you choose, having chosen it is the outcome this course wanted.