Wilder's DMI

[yang24]

Hi all,
I recently study DMI system. I set up a stock screener like this.

ADXR > 30
PDI CROSS ABOVE MDI
ADX TURNS UP

I use 30m chart and standard 14 period setup for DMI.

I thought I understand the indicator right and the screener is just followed Wilder's instruction.

But the result is confusing.
I tested the system for several days I found if I LONG the result of the screener I could only have about 40% win. btw i took profit when ADX above 40 and stop when ADX turn down.

Do I miss anything or DMI is an out-of-date indicator i should just drop it?

It would be very appreciated if someone could share his opinion on this indicator.

Thank you.

 

[Xela]

It would be very appreciated if someone could share his opinion on this indicator.

I'll share mine, then. Five points, really ...

1. I don't actually still use it now, but I think that as indicators go, it's potentially one of the more useful and helpful ones (and isn't "out of date").

2. I think whether the ADX line is rising or falling is more significant than its absolute level.

3. I think that if you're going to use it, it is worth displaying the +DI and -DI lines and getting used to looking at their relative positions and directions, as well as those of the ADX "main line".

4. Overall, I think this indicator's quite good for "broad directional bias": I used to use it to determine when to try to identify long/short entries from the price action.

5. I wouldn't even dream of using it as a specific trade-entry-timing method, per se (and this one's perhaps my main point, here).

Just my perspective ...

(And welcome to the forum!).

 

[Buy1Sell2]

Hi all,
I recently study DMI system. I set up a stock screener like this.

ADXR > 30
PDI CROSS ABOVE MDI
ADX TURNS UP

I use 30m chart and standard 14 period setup for DMI.

I thought I understand the indicator right and the screener is just followed Wilder's instruction.

But the result is confusing.
I tested the system for several days I found if I LONG the result of the screener I could only have about 40% win. btw i took profit when ADX above 40 and stop when ADX turn down.

Do I miss anything or DMI is an out-of-date indicator i should just drop it?

It would be very appreciated if someone could share his opinion on this indicator.

Thank you.

Can be used. Be prepared to get out fairly quickly as once the entry paramaters are met, the move is typically 60 to 70 percent over. You could make better use of it on a larger time frame identifying trend and use smaller time frame for entries and exits in that direction only.

 

[yang24]

I'll share mine, then. Five points, really ...

1. I don't actually still use it now, but I think that as indicators go, it's potentially one of the more useful and helpful ones (and isn't "out of date").

2. I think whether the ADX line is rising or falling is more significant than its absolute level.

3. I think that if you're going to use it, it is worth displaying the +DI and -DI lines and getting used to looking at their relative positions and directions, as well as those of the ADX "main line".

4. Overall, I think this indicator's quite good for "broad directional bias": I used to use it to determine when to try to identify long/short entries from the price action.

5. I wouldn't even dream of using it as a specific trade-entry-timing method, per se (and this one's perhaps my main point, here).

Just my perspective ...

(And welcome to the forum!).

Xela thank you very much.
I said I took profit when ADX line above 40 because I found when its value goes above 40, price tends to go flat (even ADX line keeps increasing )or turn down suddently. Well, since I'm still testing, I could not confirm the conclusion right now.
I start to consider some other indicator for enter. The DMI may not do everything right.

​

 

[Buy1Sell2]

Xela thank you very much.
I said I took profit when ADX line above 40 because I found when its value goes above 40, price tends to go flat (even ADX line keeps increasing )or turn down suddently. Well, since I'm still testing, I could not confirm the conclusion right now.
I start to consider some other indicator for enter. The DMI may not do everything right.

​

Singulary useless.

 

[yang24]

the move is typically 60 to 70 percent over.

Sorry. Do you mean the move of the ADX line is 60 to 70 percent?

 

[Buy1Sell2]

Sorry. Do you mean the move of the ADX line is 60 to 70 percent?

The price move is 60 to 70 percent over once parameters are met for entry.

 

[Xela]

The DMI may not do everything right.

It "does everything right" in the sense that it displays accurately what its instructions tell it to display.

This is true of all indicators.

But it isn't a reliable predictor of anything - and that's also true of all other indicators, too.

If you think that indicators somehow "ought to be predictive", then it may help you to think more precisely about exactly what "predictive" means, in this context. They can be "collectively predictive", arguably. I said more about it here, if that helps. I think it's important to appreciate this, because for many aspiring traders, time spent searching for a magical indicator combination which will tell them exactly when to enter and exit trades can be a huge distraction from the things they really need to learn, and to experience, and to practice, in order to maximise their chances of joining the small minority who find any steady success.

I start to consider some other indicator for enter.

I wouldn't, myself.

Indicators have some value, and some uses, but "telling you when to enter and exit trades" isn't really one of them, in my opinion. Not a good one, anyway.

All the traders I know who are successful and use indicators aren't using them for that purpose: they're using them to help to determine directional bias, and they're using price action parameters to determine their entries and exits. I don't claim that all of these people are necessarily representative of "all successful traders", and there may be some people who disagree with me about it (there usually are), but I do think it's a perspective worth looking at and thinking about, at least.

I must admit, however, that in my opinion (and more importantly in the opinions of some of my colleagues who do use indicators all the time) ADX/DMI can certainly be one of the more useful ones, as a directional bias indicator.

 

[Buy1Sell2]

This is true of all indicators.

But it isn't a reliable predictor of anything - and that's also true of all other indicators, too.




Indicators have some value, and some uses, but "telling you when to enter and exit trades" isn't really one of them, in my opinion. Not a good one, anyway.

This is false yang24. Those who know how to use indicators can use them to reliably predict direction and enter and exit properly. Typically you would use a trend indicator like DMI on larger time frame, a range indicator on medium time frame and an oscillating indicator on the entry/exit time frame. Thus, the DMI is singularly useless. ---Most traders are to lazy to look at more than one time frame and more than one indicator.---This indicator combo is not the edge in trading. Prudent Risk Management is however----Ishmael.

 

[tommcginnis]

Do I miss anything or DMI is an out-of-date indicator i should just drop it?

Wilder's ADX/DMI+/DMI- are about the best single indicator set-up invented. But don't accept whatever lookback parameters were defaulted on your platform: experiment and find what is best for your underlying.

(And how do you know when you've got the best set-up? Because you can put up a piece of paper over the candles, and describe what the candles must be according to what the ADX and DMIs show you.)

From the questions/observations you've posted, I'm sure you will see very plain, and very robust entry/exit indications from those three squiggly lines bopping all 'round on your screen. Do NOT be concerned if you don't see your trigger points reflected anywhere else on the web. Look at your own 1minute or 1hour graphs going back, add up the entry/exit $$$ (take 'em to the next candle, just to be fair!), and develop your *own* protocol.