20dmf rt

[Pascal]

The RT system is already pointing to a short position. The 20DMF will be short if we close below 0


Pascal

13810

[adam ali]

Pascal,

Did the OB/OS level on the 20DMF reach the -70 (officially oversold) level yesterday in RT? From reviewing the chart, it appears that it did which if true might have suggested one to close any shorts, that is once the 20DMF turned and crossed back above the -70 level.

[Pascal]

Pascal,

Did the OB/OS level on the 20DMF reach the -70 (officially oversold) level yesterday in RT? From reviewing the chart, it appears that it did which if true might have suggested one to close any shorts, that is once the 20DMF turned and crossed back above the -70 level.

Yes, it did.

We are however still not in shape to back-test the RT 20DMF. Therefore, such a signal might or might not be correct. Only the future will tell us. Once we will have a RT Model developed, we will probably use some different measure of OS, including a porosity factor to avoid whipsaws.

I believe that whipsaws will be the biggest challenge in using the RT system. We can see that for GDX.
Only RT back-tests using minute data will allow us to better work on these RT models.


Pascal

[adam ali]

Ok - might I then make a suggestion: could you at least alert subscribers in the future when the oversold level is reached RT? I can appreciate the lack of back-tests is problematic but at least we would be aware the level has been tripped and could then decide individually what to do with the information.

Thanks as always for the great work.

[Pascal]

Ok - might I then make a suggestion: could you at least alert subscribers in the future when the oversold level is reached RT? I can appreciate the lack of back-tests is problematic but at least we would be aware the level has been tripped and could then decide individually what to do with the information.

Thanks as always for the great work.

But Adam, you can see the RT as well as me and as any RT subscriber.

Why do you need me to tell you that this EOD indicator has caught - probably by pure chance - some price reversal because it reversed over the OS level. I developed that indicator and I know that it is not precise enough to be used in RT. The little trip below the -70 level might just have been a mathematical inconsistency due to the overnight price gap.

So, in short, I do not want to push this OB/OS indicator that is really not to be used in RT.


Pascal

13812

[adam ali]

I must be misunderstanding something or there's a nuance I'm missing.

-70 is used as the "official" oversold level when it occurs end of day but isn't terribly important when the level occurs RT? Back in December the -70 level was hit three times RT before reversing back above by end of day, leading to the "near miss" short signal reversal - which was described in detail by you and others here.

So this leaves me confused as to why what was of import back then is not particularly important in this instance. I do understand models shouldn't be held to such a high level of precision but at the same time -70 EOD is the level selected at which to make changes in the 20DMF signal.

[Pascal]

I must be misunderstanding something or there's a nuance I'm missing.

-70 is used as the "official" oversold level when it occurs end of day but isn't terribly important when the level occurs RT? Back in December the -70 level was hit three times RT before reversing back above by end of day, leading to the "near miss" short signal reversal - which was described in detail by you and others here.

So this leaves me confused as to why what was of import back then is not particularly important in this instance. I do understand models shouldn't be held to such a high level of precision but at the same time -70 EOD is the level selected at which to make changes in the 20DMF signal.

Back in December, we closed at -69.89, which is very near -70. This is clearly a "computation error" away from a signal.

This time, on an EOD base, the lowest that we went was -62.53. This is much more than a "computation error."

So yes, there is a significant mathematical difference.
I however would not contest if you use these RT signals. They have not been back-tested, but in the future a back-test might prove that you were correct.


Pascal

[adam ali]

Ok - I do understand your points and appreciate the further elucidation.