Barney,
No, I would NOT generally use just 30 days of data when reoptimizing. That will usually be too short a time frame to identify meaningful trends.
I want to use the longest period of time that I feel is reasonably representative of the current trend.
For example, if you are swing trading, (very short term perspective of a few days to a week or so - trying to grab quick profits from small undulations) try to pick the longest timeframe that seems appropriate and representative of such price action. Then, optimize settings to take advantage of those smaller swings.
If you have a longer term perspective, (usually weeks or months), try to go back 4 or 5 months, or possibly longer, then set parameters to miss the smaller undulations and take advantage of medium or longer term trends. Sometimes, 4 - 5 months is inappropriate, due to a major trend change and a whole new pattern. That should be fairly obvious from eye-balling the chart. Then, you have to use your judgement to select a suitable time frame.
Shorter term trading - Shorter optimization period.
Longer term trading - Longer optimization period.
These are only guidelines. XDEV is a tool, not a black box. Few stocks or markets are identical.
Hope that clarifies it some.
Lou
