Time series momentum of a price series means that past returns are positively correlated with future returns. It follows that we can just calculate the correlation coefficient of the returns together with its p-value (represents the probability for the null hypothesis of no correlation).
One feature of computing the correlation coefficient is that we have to pick a specific time lag for the returns. Sometimes the most positive correlations are between returns of different lags. For example, 5 day returns might show negative correlations, while the correlation between the past 20 day returns with the future 40 da returns might be very positive. We should find the optimal pair of past and future periods that gives the highest positive correlation an use that as our look back and holding period.
Cross sectional momentum refers to the positive correlation of a price series past and future relative returns, in relation to that of other price series in a portfolio.
Futures exhibit time series momentum mainly because of the persistence of the sign of roll returns.
If you are able to find an instrument (an ETF or another future) that cointegrates or correlates with the spot price or return of a commodity, you can extract the roll return of the commodity future by shorting that instrument during backwardation or buying that instrument during contango.
Portfolios of futures or stocks often exhibit cross sectional momentum: a simple ranking algorithm based on returns would work.
Profitable strategies on news sentient momentum show that the slow diffusion of news is a cause for stock price momentum.
The contagion of forced asset sales and purchases among mutual funds contributes to stock price momentum.
Momentum models thrive on “black swan” events and the positive kurtosis of the returns distribution curve.
Is your goal the maximization of your net worth over the long term? If so,
consider using the half-Kelly optimal leverage.
Are your strategy returns fat-tailed? You may want to use Monte Carlo simulations to optimize the growth rate instead of relying on Kelly’s formula.
Keeping data-snooping bias in mind, sometimes you can just directly optimize the leverage based on your backtest returns’ compounded growth rate.
Do you want to ensure that your drawdown will not exceed a preset maximum, yet enjoy the highest possible growth rate? Use constant proportion portfolio insurance.
Stop loss will usually lower the backtest performance of mean-reverting
strategies because of survivorship bias, but it can prevent black swan events.
Stop loss for mean-reverting strategies should be set so that they are never triggered in backtests.
Stop loss for momentum strategies forms a natural and logical part of such strategies.
Do you want to avoid risky periods? You can consider one of these
possible leading indicators of risk: VIX, TED spread, HYG, ONN/OFF, MXN.
Be careful of data-snooping bias when testing the efficacy of leading risk indicators.
Increasingly negative order flow of a risky asset can be a short-term leading risk indicator.