04-24-2018, 06:31 PM
Hi
thats intresting. Dynamics in the good old analog fasion.
I found that:
since the indicators live of there own averageing process, it becoms probelmatic to change the averaging-candle-ammount.
THIS WOULD REQUIRE a proper reset and then a preset of the new averaging length bevor we could use the new length value.
i have written my own indicator that does exactly that:
if the trend would change very fast, it will toss the old averaging cell and wait for the new trend to settle, BEFORE she initialize the new averaging-length with the newest datapoint.
...only one datapoint for 15 vars...then i stuff them all with this newest datapoint to beginn with.
i have found that all the indicators work the same way, where they live from the averaging process.
this makes indicating a trend change very slow compared to the realtime signal.
(not always true: guess the expo function would make a change dynamicly heavier because of the exponential nature.)
Since my idea works perfectly well for the macro trend (support and resistance) i will now go ahead and write my own
RSI-indicator, this is only a easy formula that deals with the candle.data.
But then i implement this reset/set function into it.
To go around the digital problem i use always % calculation. In this way i then have the pure analog state of a value and with multiplication it makes a user.setting.ammount.
example:
a slow indicator (MA) would indicate the trend angle by pos/neg number.
this can be added to the RSI thresholds. BUT wich ammount do we set?
Solution just add a number in the user settings that then can be used as multiplication factor.
RSI.threshold_hi = MAfactor * user.setttings.value
value = 0 //the modulation amount is compleatly shut of. love it
value = -3 //neg/pos/none... nice
i tried with RSI but the results were not beneficial over 5 month backtest.....
thought the angle of the MA must be in fase with the trend but we have no optical graph that could show whats going on.
...number analysis
would be intresting to hear your results.
thats intresting. Dynamics in the good old analog fasion.
I found that:
since the indicators live of there own averageing process, it becoms probelmatic to change the averaging-candle-ammount.
THIS WOULD REQUIRE a proper reset and then a preset of the new averaging length bevor we could use the new length value.
i have written my own indicator that does exactly that:
if the trend would change very fast, it will toss the old averaging cell and wait for the new trend to settle, BEFORE she initialize the new averaging-length with the newest datapoint.
...only one datapoint for 15 vars...then i stuff them all with this newest datapoint to beginn with.
i have found that all the indicators work the same way, where they live from the averaging process.
this makes indicating a trend change very slow compared to the realtime signal.
(not always true: guess the expo function would make a change dynamicly heavier because of the exponential nature.)
Since my idea works perfectly well for the macro trend (support and resistance) i will now go ahead and write my own
RSI-indicator, this is only a easy formula that deals with the candle.data.
But then i implement this reset/set function into it.
To go around the digital problem i use always % calculation. In this way i then have the pure analog state of a value and with multiplication it makes a user.setting.ammount.
example:
a slow indicator (MA) would indicate the trend angle by pos/neg number.
this can be added to the RSI thresholds. BUT wich ammount do we set?
Solution just add a number in the user settings that then can be used as multiplication factor.
RSI.threshold_hi = MAfactor * user.setttings.value
value = 0 //the modulation amount is compleatly shut of. love it
value = -3 //neg/pos/none... nice
i tried with RSI but the results were not beneficial over 5 month backtest.....
thought the angle of the MA must be in fase with the trend but we have no optical graph that could show whats going on.
...number analysis
would be intresting to hear your results.