Time window option to more accurately cluster trades for Monte Carlo Simulations

Hello guys,

this is a request for an additional feature in the Monte Carlo Simulator.

At present all trades for the data sample period are jumbled up in random. What we know is that trades can be correlated and if the strategy you are testing is across a portfolio then some trades entered at similar times will be correlated with one another. At present the Monte Carlo simulator will jumble up all these trades which adds diversification to the portfolio and therefore gives a narrower distribution of expected futures profits and losses.

Solution:

an option of either jumbling up all trades as done presently or cut the trades into time segments and then jumble these up.

For instance, you could have a setting that allows all trades taken within a one week time-frame to remain together and then jumble up all of these one weeks time segments to create the new simulated return distributions. Therefore if my portfolio took trades in silver, gold, and platinum in the same week and i held them all for the next 5 days, present if these were all big wins or losses, the Monte Calo Simulator jumbles them up, distibuting them through the time series and so creating a smoother equity curve, rather than more accuratly clumping them together.

the present solution will by its very nature add diversification to a system portfolio and so give a smaller return distribution, both to the up and downside.