Hello All,
I'm still alive and just reporting in from the Bunker on my latest thoughts that may interest some of you....
Attached are a couple of charts showing a Kalman filter I have just managed to get working. I started this quite a while ago but could never get it working properly.
As my reader(s) will know, I have an aerospace background and have designed several guided missile autopilots in my time using Kalman filters. But for some strange reason, I could never get them to work satisfactorily for the markets.
But then I happenned to be reading the EURUSD thread in the Interactive forum and noticed a few posts by Hindustani (for example http://www.forexfactory.com/showthre...09#post6190809).
(Why is Old Dog reading the Interactive threads you may ask -- he is a life-long swing and position trader on the Daily and Weekly charts! Well.... I have a confession to make. Now I am retired and have far too much time on my hands, I am in serious danger of getting sucked in to the scalping world all the dashing young blades on this forum are so fond of. )
Hindustani's "CCI" indicator looks remarkably like my Kalman and I suspect that there is a lot more to his indicator than just CCI. In fact, he says in one post (http://www.forexfactory.com/showthre...29#post6190929) that he is using an external Matlab setup to do the calculations.
So this got me thinking, and I reckon I might have found the bug in my approach!
This thing is a scalpers dream -- the attached charts from today are quite typical in all respects.
Probably the most important thing to realise about the Kalman filter is that it is NOT A FILTER! It is an estimator for signals buried in lots of noise. Not a bad description of the market on low time frames!! (See my thread here http://www.forexfactory.com/showthread.php?t=336715 on this).
There is absolutely no question of "repainting" or any of those silly issues that plague lots of MT4 indicators and the price-position estimates are quite statistically robust (p < 0.05). I am taking a very simple approach basing the current estimate on the current price pivot, and the uncertainty (noise) on the range (hi - lo) of the bar. But simple seems to work just fine...
I made some very good pips today and am in danger of becoming a day-trading addict if I don't exercise some strict self-discipline!
Like Hindustani, I can't post the indicator as it is too complex. I have a rather convoluted system on my terminal that exports tick data from MT4 via the DDE server into Matlab. Matlab then does all the serious crunching that would be far too difficult in MQ4 language, and then returns the results to MetaTrader.
I would hate for any serious programmer to see my code - I would be far too embarrassed!
But some very fine programmers on the Platform Tech forum have made very similar stuff available if anyone wants to have a go.
I am just offering this out of interest.
Kind regards and good trading to all,
Old Dog
.
I'm still alive and just reporting in from the Bunker on my latest thoughts that may interest some of you....
Attached are a couple of charts showing a Kalman filter I have just managed to get working. I started this quite a while ago but could never get it working properly.
As my reader(s) will know, I have an aerospace background and have designed several guided missile autopilots in my time using Kalman filters. But for some strange reason, I could never get them to work satisfactorily for the markets.
But then I happenned to be reading the EURUSD thread in the Interactive forum and noticed a few posts by Hindustani (for example http://www.forexfactory.com/showthre...09#post6190809).
(Why is Old Dog reading the Interactive threads you may ask -- he is a life-long swing and position trader on the Daily and Weekly charts! Well.... I have a confession to make. Now I am retired and have far too much time on my hands, I am in serious danger of getting sucked in to the scalping world all the dashing young blades on this forum are so fond of. )
Hindustani's "CCI" indicator looks remarkably like my Kalman and I suspect that there is a lot more to his indicator than just CCI. In fact, he says in one post (http://www.forexfactory.com/showthre...29#post6190929) that he is using an external Matlab setup to do the calculations.
So this got me thinking, and I reckon I might have found the bug in my approach!
This thing is a scalpers dream -- the attached charts from today are quite typical in all respects.
Probably the most important thing to realise about the Kalman filter is that it is NOT A FILTER! It is an estimator for signals buried in lots of noise. Not a bad description of the market on low time frames!! (See my thread here http://www.forexfactory.com/showthread.php?t=336715 on this).
There is absolutely no question of "repainting" or any of those silly issues that plague lots of MT4 indicators and the price-position estimates are quite statistically robust (p < 0.05). I am taking a very simple approach basing the current estimate on the current price pivot, and the uncertainty (noise) on the range (hi - lo) of the bar. But simple seems to work just fine...
I made some very good pips today and am in danger of becoming a day-trading addict if I don't exercise some strict self-discipline!
Like Hindustani, I can't post the indicator as it is too complex. I have a rather convoluted system on my terminal that exports tick data from MT4 via the DDE server into Matlab. Matlab then does all the serious crunching that would be far too difficult in MQ4 language, and then returns the results to MetaTrader.
I would hate for any serious programmer to see my code - I would be far too embarrassed!
But some very fine programmers on the Platform Tech forum have made very similar stuff available if anyone wants to have a go.
I am just offering this out of interest.
Kind regards and good trading to all,
Old Dog
.