About SQLMaxx
SQL is mostly logic you already use.
The work is learning to write it down precisely.
Why this exists
I tried the standard route first. A free YouTube series, a paid course or two, the practice sites that everyone recommends.
Each one had the same shape. Watch a video. Solve a few problems. Move to the next module. After three weeks of that, I sat down to write a query from scratch and could not start. Finishing a module is not the same as understanding the material, and no part of the system I was using could tell the difference.
What was missing was a model of what I actually knew. A system that tracked each concept separately, watched whether I was getting the hard variants right, and held me at a concept until I was. Spaced review for the things I had locked in, so they did not slip. A clear signal for readiness, instead of a completion bar that lied to me. The tools I tried did none of this. I built one that does.
How learning is supposed to work
The default model in this space is completion. You watch the videos, you click through the exercises, the platform marks the module done, and you move on. The hidden assumption is that finishing the work means you learned it. For most people, on most material, that assumption is wrong. You can finish every module of a SQL course and still freeze on a question your future job would consider basic.
Completion is not understanding.
A different model. The system tracks how confident it should be in your understanding of each concept, separately, and updates that confidence based on how you do on real problems. Easy correct answers nudge it up a little. Hard correct answers nudge it up a lot. One wrong answer is treated as a slip and held flat. A second wrong answer in a row pulls confidence back down. You do not move past a concept until the system is sure. Then, on a schedule tuned to how memory actually decays, the concept comes back for review before you start to forget it.
This is mastery learning with spaced review built in. The research on it is decades old and the results are not subtle. The reason it is rare in tooling is that it is much harder to build than a video player.
SQL is unusually well-suited to this approach. The concepts have a clean dependency structure: filtering before grouping, grouping before window functions, and so on for sixty-six steps. That structure makes per-concept mastery tracking tractable in a way it would not be for, say, a writing course. The model can know what you have locked in, what depends on it, and what to surface next.
What this is not
Four shapes of SQL learning resource exist today. None of them is what SQLMaxx is, and the difference matters.
- The grinders
- High volume of problems and a leaderboard. No model of what you actually know. You finish ninety problems and the system has no answer to the question of what you should do next.
- The problem dumps
- A long list of questions and a query box. No progression, no signal for when you are ready, no defense against forgetting what you already learned.
- The video libraries
- Instruction-heavy, often well-produced, almost no graded practice. Watching someone else solve queries is a different skill from solving them.
- The syntax tutorials
- Clauses introduced one at a time, in isolation, never assembled into the kind of query an actual analyst writes. By the time you finish, you can recognize SQL but cannot yet write it.
What I believe about how SQL is learned
-
The logic is already there
Most of SQL is logic you use to organize your own life. Filtering, grouping, ranking, joining one list to another. The language is notation for what you already do.
-
Practice has to be honest
A system that tells you you are ready when you are not is worse than one that tells you nothing. Readiness is a claim that has to be earned on real problems and real difficulty.
-
Forgetting is the default
Anything you do not review you will lose. The schedule for review can be calculated. There is no reason to leave it to memory or willpower.
-
Hard problems strengthen easy ones
Solving a difficult query well is also evidence that the easier concepts inside it are solid. The system gives you credit for that, instead of asking you to re-prove the basics.
-
The point is real work
The skill you are trying to build is the skill of writing a query against a real schema to answer a real question. Everything else is rehearsal for that.
The receipts
What is actually behind the system, in numbers.
- 66
- Concepts in the curriculum, taught in order, each with its own mastery signal.
- 615
- Practice problems, every one running against a real PostgreSQL engine.
- 0
- Servers between you and the database. The engine runs in your browser.
If you have read this far, the next step is small. The first four concepts are free, no card. Use them long enough to feel whether the system fits how you learn.