Five financial research curses - They cannot be avoided

The above chart is from Marcos Lopez de Prado's insightful powerpoint "Escaping The Sisyphean Trap: How Quants Can Achieve Their Full Potential" serves as a road map for the important issues that will cause failure in financial research. Every researcher has faced all of these five curses. There is no avoiding them. They are not problems to be solved but cure curses to be lived with whenever quantitative analysis is undertaken. 

Overarching the other four curses is primary problem that economics and finance cannot run experiments like hard sciences, so cause and effect relationships cannot be easily structured. Economists are trying very hard to develop controlled experiments but the work almost non-existent in finance. The curse of non-stationarity makes testing even harder. Even if you could determine cause and effect, the intensity of relationships may change significantly through time. With regime changes and unstable conditions, past relationships cannot easily be extrapolated. The rules of the game are always changing, so nothing is stable. 

Finance is a zero-sum game. One man's research is a lifeblood to future profits and cannot be shared. There often is a limit to the amount of alpha that can be extracted, so research is not always public. Studies have shown that alpha declines as soon as working papers and articles are published by academics. There is value with secrecy.

Complexity is high and the amount of variation that is explained by any model is low. Financial asset prices are discounted expectations which are affected by a wide set of factors that change with firm and macro conditions. Pricing in finance is not a linear problem.

Finally, the sample sizes for many phenomena are small. Think of all the business cycle analysis done by economists when in reality we have only seen eight since 1970. If I did analysis with a sample of 8, most would laugh at my ignorance of significance and power testing. 

So how do we solve these curses? We cannot. All we can do is try and mitigate risks and be aware of the problems.