7. Closing: the Academy's three tracks

Level 7: Linear algebra and portfolio optimization

You have reached the last module of the last level of the Academy.

Before closing, it is worth looking back at the complete map you have covered. The three tracks, theory, technical and mathematics, were not three independent courses that happened to share a topic.

They were three complementary ways of answering the same question from different angles: how do I know whether an idea about markets is real, or only looks real?

What the theory track gave you

The theory track gave you the vocabulary and the intuition, and there you learned things like these:

• What a share is and how a price forms.
• Who takes part in the market and with what incentives.
• How to read financial statements and multiples.
• What volatility and correlation are.
• Why the Sharpe ratio matters more than gross return.
• Which psychological biases make us see patterns where there are none.
• Why a badly built backtest can lie without anyone lying on purpose.

Without that vocabulary, everything that came afterwards would have been manipulating symbols without knowing what they represent.

What the technical track gave you

The technical track gave you the ability to act, and that is where you went from reading to doing:

• Downloading real data.
• Building a screener.
• Coding a momentum, mean reversion or pairs trading strategy.
• Checking on honest historical data whether that strategy would have worked.

The word "honest" is not decoration there. It means without the look-ahead bias that artificially inflates so many backtests circulating on the internet.

Without that ability, the other two tracks are left lame. The theory would have remained pretty ideas with no way of testing them against reality, and the mathematics would have remained formulas with no data to apply them to.

What this track has given you

And this track, mathematics and statistics, gave you the rigor not to deceive yourself:

• Probability and distributions taught you to think in terms of possible scenarios, not certainties.
• Descriptive and inferential statistics gave you the tools to summarize data and to know when an observed difference is genuine and when it is noise.
• Correlation, regression and CAPM formalized with numerical precision relationships between assets that the theory track had only described in words.
• Time series, stationarity and cointegration gave you the right language to talk about how prices evolve over time.
• Hypothesis testing, the p-value and the data snooping problem gave you the discipline not to mistake a pattern that worked by chance for a real strategy.

And this last level, with vectors, matrices, the efficient frontier and the optimization problem, gave you the formal framework for thinking about the risk and return of a complete portfolio, not of an isolated asset.

It also gave you something less flattering and just as useful: understanding why even the mathematically most elegant tool can fail in practice if the data you feed it is not reliable.

The cycle that never ends

What matters is not that you now know scattered formulas from each block. It is that the three tracks combine into a cycle that never ends:

• Theory gives you a reasonable idea to test.
• Mathematics gives you the tools to design an honest test of that idea, and to recognize when the result is noise dressed up as signal.
• The technical side gives you the ability to run that test on real data and to iterate fast.

Anyone who masters only one of the three tracks has an incomplete tool, and it shows immediately:

• Without theory, they do not know which questions are worth asking of the data.
• Without mathematics, they cannot tell a real finding from a statistical coincidence.
• Without the technical part, everything else stays on paper, with no way of checking it.

How to look at your own ideas from now on

If you have completed all three tracks, the reasonable stance from here is not to feel ready to trade real capital with no further questions.

The reasonable stance is exactly the opposite, that is, to use everything you have learned to treat any idea, including your own, with the same skepticism you would apply to a stranger's idea on an internet forum.

Put it into practice when you build a backtest in the quant500.com screener over the S&P 500 universe with honest SEC data. Ask yourself three things:

• How many variants you tried before settling on the one that worked.
• Whether the result holds outside the sample you used to design it.
• Whether the sample size is large enough that the observed difference is not, simply, noise.

And keep this last rule above all the others you have seen in the Academy. The better a result looks at first glance, the more reason there is to suspect overfitting before believing it.

What is deliberately left out

What this Academy deliberately leaves out is as important as what it contains:

• Risk management in real execution, that is, transaction costs, liquidity and slippage.
• Machine learning applied to financial markets.
• The discipline, not mathematical at all but decisive, of staying calm and sticking to a plan when a strategy goes through a bad run that the mathematics itself had already warned you was possible.

Those are the reasonable next steps, not a diploma that closes the subject.

What you take away

It is best to finish without adornment, because a closing that sounded better than reality would serve no purpose. Markets give nothing away for free: they are, for the most part, surprisingly efficient, and any real edge tends to be small, fragile and demanding to maintain.

What you take away from this Academy is not the promise of a winning formula, but something more valuable and more durable: the ability to tell apart, honestly and with real tools, an idea that deserves to be put to the test from one that merely looks like it.

That does not expire, and it is exactly what you need in order to keep learning on your own from today onwards.