Robustness – what do these mean? Are they good?
3 replies
nolube
7 years ago #115248
Have had a few robustness results like this, running 200 simulations on the change of indicator parameters setting. I think the logic behind it is that the change of indicator settings do not change the strat much, like whether it breaks the last 200 highs or last 160 or 240 it doesnt make much difference. But does anyone else read anything into these? Are they considered good or unreliable?
tomas262
7 years ago #137837
It seems good in case you base your strategy on indicators. Also it is important how you set the test itself (% value change). You can also use RT to see how far it can go with parameter range until it breaks apart by systematically increasing the % change value
Patrick
7 years ago #137853
It seems good in case you base your strategy on indicators. Also it is important how you set the test itself (% value change). You can also use RT to see how far it can go with parameter range until it breaks apart by systematically increasing the % change value
so RT tests are not working for strategies without indicators? what RT test should we do then?
Rom
7 years ago #137882
Mont Carlo simulation will allow you to change several parameters, like the bar from which “trading” starts, the price of the instrument, etc, It wil run the strategy and display the results. The process is repeated N times (N is determined by you). It is only one of the steps to determine “robustness”. The ultimate test is real time evolution of price series.
Viewing 3 replies - 1 through 3 (of 3 total)