The determination of good from bad investments using quantitative measures.
"Meta-overfitting" may be endemic in finance as well as other fields of research.
We find evidence that hypothetical performance data can be significantly over-optimistic compared to subsequent realised performance.
Without an appropriate understanding of data, it can be too easy to discover fool’s gold.
We caution against drawing conclusions from short time series and emphasise the importance of acknowledging the uncertainty on performance estimates.
Publication bias could have a part to play in the disappointing performance of popular equity indicators.
A unique experiment allowed us to demonstrate the effectiveness of our research process in a relatively short time.
October has been the most volatile month for stocks on average over the past 87 years. Is this due to chance?
A Winton research project with UC Berkeley used a technique for adding noise to datasets that makes it impossible to identify individuals, while preserving broader statistical patterns.
Charts are widely used in finance. As well as representing information efficiently, they can reveal relationships or anomalies in data. But their effectiveness or otherwise is a function of the choices made during chart construction.
We assess which methods are most pertinent for different trading strategies, and explain why a robust research process is vital to Winton’s approach.