Detecting Fake Price Movements - Convergence Divergence Indicator

Hi RapzyFX

My take from the provided literature is that - yes, if the conditions are met then price at that time is fake.

However, just like any other indicator - doesn't always work. So we need to build a trading system that complements this indicator.

The literature states that the quality of signals provided by CDI are more reliable than RSI. I am not sure how that analysis was done, as CDI and RSI measure different things.

Hi PipsMeUp

This system:
https://www.forexfactory.com/showthread.php?t=635569

Uses the CCFp indicator, which also uses correlation to generate signals. It uses USD across 28 pairs as its counter currency for number crunching.

As you can see from that thread, it is doing really well.

I am not claiming this will turn into something as good, however, it certainly does demonstrate that using correlation analysis in trading has merit to it.

So thank you for your compelling arguments, but I will venture forth and see where this leads us.

(I'm not a big fan of MT4/MT5 so I won't use those CCFp-Diff indicators. Especially because the are compiled .ex4)

Don't get me wrong I'm not telling you can't make money with correlation. I'm telling you that you can't stumble on a paper presenting a wrong methodology write an indicator from that and try to guess a strategy that fits. The paper makes the mistake of calculating the correlation of time series without detrending first. This is a newbie mistake in statistical analysis. It then proposes to fade when the prices decorrelate on the short term. I posted a picture that shows this can lead to a loss. The pdf doesn't explain why this is supposed to be a profit opportunity. IMO that's quite the opposite because your basket is not longer safe to trade (the mean reversion property is lost).

The overall strategy is yet pretty simple: you build a synthetic asset that ranges forever and you swing trade it.

The steps are (theoritically) simple:
- First you try and find a pair of correlated assets.
- Second you regress one onto the other to build a basket. This thread is of interest to you: https://www.forexfactory.com/showthread.php?t=262827.
- Third you check the residuals of the regression (the basket) has a chance to be mean reverting. => https://en.wikipedia.org/wiki/Dickey...%93Fuller_test

This 3rd point needs the assets to have some economical relationship. You can try Gold-Silver, Crude-Brent Oils, DAX30-CAC40, AUD-NZD, DJIA-SP500. There was also good arguments from old_dog and 7bit about this last step: Dickey-Fuller test is weak with fractionnaly integrated series like markets are. 7bit's point was to pretend it almost works yet he knows it's not the case and in case of doubt quite the trade or re-balance.

Note that the regression will usually not be one for one. You're not going to sell 1 EUR and buy 1 AUD to get 1 E/A short. Usually you'll sell 1 unit of EUR and sell 0.7 unit of AUD (just an example I don't know the exact ratio). The ratio changes over time. The re-balancing of the basket is a challange on it own: constantly adjusting costs a lot in spreads and commission but not adjusting often enough breaks the mean reversion and increases your risk.

SELL AUDNZD

1. Open: 1.07627
2. Instrument1: AUDUSD
3. Instrument2: NZDUSD
4. FastPeriod: 12
5. SlowPeriod: 63

Let's see how it turns out.

May have to look at a better exit strategy. It seems like it will take a while to cross over 0.5.

Currently +25 pips in profit.

Either this will result in really good profit (if I wait for it to cross above 0.5) or it will reverse at some point to wipe out profits.

BUY EURGBP (I would wait for price to cross above -0.5 before entering as the downtrend looks strong)

1. Open: 0.85385
2. Instrument1: EURNZD
3. Instrument2: GBPNZD
4. FastPeriod: 12
5. SlowPeriod: 63

Let's see how this turns out.

Locked profit at: 30 pips.

+40 pips now, still running...

Hi Samsha786

What is rules regarding your TP, SL and trailing stop ?

Hi Morne

As per post #1, TP is when the blue line crosses above 0.5. However, I am looking to improve this with a break even / trailing stop.

As for SL, I don't have a hard and fast rule yet - so far I have been using 10 pips.

Closed at 1.07302. About 32 pips.

+18 pips. Locked profit at 10 pips.

Waiting for BLUE line to cross below -0.5 to sell AUDNZD

 

1. Instrument1: AUDUSD
2. Instrument2: NZDUSD
3. FastPeriod: 12
4. SlowPeriod: 63

Let's see how this turns out

THANK YOU

Thread Closed.

i come to slow. Please send me cdi.ex4 plz. i interested it.