The AlphaLab blog

With infinite attempts anyone succeeds by chance. Overfitting is a strategy tuned with unlimited attempts to fit the past perfectly.

Studying with the exam answers proves nothing. Out-of-sample data and Walk-Forward are the surprise exam that truly measures a strategy.

The Sharpe Ratio measures how much profit you earn per unit of scare (risk): it rewards the smooth ride, not just the speed.

Test enough strategies and one will look brilliant by pure luck. The Deflated Sharpe discounts that luck and reveals what actually works.

The strategy that won in testing might be worse than average in real life. PBO measures the chance you were fooled.

Reorder your trades thousands of times and if you still win almost always, you are genuinely good. If only the exact order worked, it was luck.

Like a salary that shrinks after taxes, the market charges tolls on every trade that quietly eat your profits.

Drawdown measures the deepest fall before you recover. Learn why it decides whether you can stick to your plan.

Pretty indicators usually show only their best moments. Learn to spot the trick and demand statistics you can verify yourself.

An edge is a tiny advantage, like a casino's. It doesn't print money and it wears out over time. Learn to recognize a real one.

What works in a calm market can sink in panic. Understand why the market changes climate and why your strategy must survive several.

We only see the ones who survived. That illusion inflates expectations in trading. Here is how to fight it.

Torture the data long enough and it will confess to anything. The honest fix: declare your bet before you roll the dice.

Even with an edge, betting too much on one hand wipes you out on a bad streak. Surviving matters more than maximizing.

Splitting eggs across baskets is useless if all the baskets ride the same cart. Diversification only works if your bets move differently.