Repainting: what can you do against it???

Hi,
what can you do against repainting of indicators? When you do a visual mode backtest of an empty EA and you drop the indicator on the chart, you really see how it performs (repainting or not),
opposed to simply putting it on a live chart (too slow to evaluate repainting). So if you have seen that it does repaint, what can you do against that?
Is there a general rule that you need to follow? Or do you have to go through the code, step by step?
As an example I have attached a Hodrick-Prescott filter, an econometric method to split timeseries in a trend part and a cycle part. I like it a lot,
but only it repaints... How do I clean up this code so I get a "realistic" view of where it will have been on the past bars?

thx for the feedback guys, this forum is great!

i cant imagine what 'nobs' number you are using but you must really like it
you'd have to add a delay on the whole thing, so it is repainted/calculated/reported at a specified interval.

what can you do against repainting of indicators?
you can rewrite it or write a new one so they wouldn't repaint

Or you can do this.

Kenz987,
thank you for the code. Strange thing however, you only start to see the difference when doing the empty-EA-visual-mode-backtest thing: you don't see any difference when you have just dropped them on the chart. Any pointers on how to see that right away without having to do the empty EA backtest? (this should be the way all indicators are shown in the first place I think, to avoid the many false Eureka-moments..)

kd3726,
nobs is not that is important, it's the lambda that is the smoothing parameter. The reason why I really like it is that Hodrick-Prescott is the optimal smoothing of a timeseries, the reason that I hate is that it is not a causal filter: it uses data to the right of the current bar to compute the current bar HP-value, hence, not really usefull in real-time trading, unless you can find some kind of transform...