Premium Blog

Hypothesis tests based on one-sample t-tests and on bootstraps can be misleading, especially when returns are generated by an over-fitted trading strategy designed via machine learning and multiple comparisons.

Statistical hypothesis testing results for the purpose of assessing the significance of trading and investment strategies must be carefully scrutinized because they are in most cases naive or can even be used to mislead the public.

Report on the state of Price Action Lab diversified portfolio, SPY adaptive trend-following, stock market randomness and Hurst exponent. Asset allocation The Price Action Lab diversified portfolio in SPY, TLT, GLD and DBC with weights of 40%, 40%, 10% and 10%, respectively, ...

In this article we analyze the performance of two popular mean-reversion trading systems in momersion regimes, i.e., when there is no clear indication whether a market is driven by momentum or mean-reversion. One conclusion is that one of the systems ...

Optimal leverage for continuous rebalancing of a passive SPY tracking strategy is still above its longer-term mean and far from trend reversal trigger levels.

Data-mining is widely used nowadays for trading algo development

Visual chartists often believe they are doing technical analysis but in most cases what they do has little relation to it. It is not hard to distinguish between a modern technical analyst and a visual chartist.

This is an indicator of market short-term reversals often mentioned in Premium reports. The P-Dow indicator is based on the p-indicator and it works like an oscillator: when its value rises above a threshold that usually means that the probability of a ...

The title could have as well been “Fooled by Statistics”

Although the overnight edge is AAPL is gone, maybe for good, it persists in some other securities. Here is an example with a substantial overnight edge in 2014.