I trade with 200 tick charts...which can be noisy. However, I plot an adaptive moving average with a period of 50....i.e. 10,000 trades. It works well, but I need to make a trend determination.
Given 50 bars, what should the delta be to determine the trend ?
Avg[0] - Avg[10] where 10 would be based on the period.....so 50 = 10, 100 = 20.....
these are only examples.
If the offset is too small, any slight change in the trend will be detected.
This of course would be the pullbacks....which absolutely KILL ME.
So I want an offset that would ignore the PB's.
Should I backtest this, or do one of you have a rule-of-thumb for this determination ?

syswizard 

I trade with 200 tick charts...which can be noisy. However, I plot an adaptive moving average with a period of 50....i.e. 10,000 trades. It works well, but I need to make a trend determination. Given 50 bars, what should the delta be to determine the trend ? Avg[0] - Avg[10] where 10 would be based on the period.....so 50 = 10, 100 = 20..... these are only examples. If the offset is too small, any slight change in the trend will be detected. This of course would be the pullbacks....which absolutely KILL ME. So I want an offset that would ignore the PB's. Should I backtest this, or do one of you have a rule-of-thumb for this determination ?

you are right 200t is very noisy, so much that it will chop up whatever trend indicator you use, switch to fib indicator look for reversal points instead

cory 

you are right 200t is very noisy, so much that it will chop up whatever trend indicator you use, switch to fib indicator look for reversal points instead

Not big on fibs....in fact, fib charts give me a huge headache.
200t is easily tamed with a 50 period moving average.....just that I need to determine trend from the average.
Using least squares regression for trend is overkill IMHO....just need to look-back a certain period.