In this context, I don’t think there’s a significant difference between “looks efficient to people like [you]” vs “is efficient relative to people like [you]”.
But more importantly, the best way for your friend to learn how efficient the market is is by him trying to beat it and failing. He’ll learn more about math and markets that way than if he listens to you and stops trying. I think he’s making the right decision to ignore you. By paper trading, he can do this without risking significant capital.
As for measuring the quality of a strategy after-the-fact, a good tool is Sharpe ratio.
Ok, I will try to nudge him in the direction of analyzing risk mathematically. If he implements the strategy using python, do you think p-values are a good enough tool to analyze whether his proposed strategy is better than luck, or would I need a more complex framework? (If I understand correctly, the strategy he’s using doesn’t involve any parameters, so the risk of overfitting is low.)
That’s a complex question. A p-value is theoretically useful, but so easy to misuse in this context that I’d advise against it.
Quantitative finance is trickier than the physical sciences for a variety of reasons, such as regime change. If you’re interested in this subject, you may enjoy this thing I wrote about the subject. It doesn’t address your question directly, but it may provide some more general information to better understand the mathematical quirks of this field.
In addition, you may enjoy Skin in the Game by Nassim Taleb. (His other books are relevant to this topic too but Skin in the Game is the book to start with.)
In this context, I don’t think there’s a significant difference between “looks efficient to people like [you]” vs “is efficient relative to people like [you]”.
But more importantly, the best way for your friend to learn how efficient the market is is by him trying to beat it and failing. He’ll learn more about math and markets that way than if he listens to you and stops trying. I think he’s making the right decision to ignore you. By paper trading, he can do this without risking significant capital.
As for measuring the quality of a strategy after-the-fact, a good tool is Sharpe ratio.
Ok, I will try to nudge him in the direction of analyzing risk mathematically.
If he implements the strategy using python, do you think p-values are a good enough tool to analyze whether his proposed strategy is better than luck, or would I need a more complex framework? (If I understand correctly, the strategy he’s using doesn’t involve any parameters, so the risk of overfitting is low.)
That’s a complex question. A p-value is theoretically useful, but so easy to misuse in this context that I’d advise against it.
Quantitative finance is trickier than the physical sciences for a variety of reasons, such as regime change. If you’re interested in this subject, you may enjoy this thing I wrote about the subject. It doesn’t address your question directly, but it may provide some more general information to better understand the mathematical quirks of this field.
In addition, you may enjoy Skin in the Game by Nassim Taleb. (His other books are relevant to this topic too but Skin in the Game is the book to start with.)