Multi 10:4 Trader EA

[snailbeard]

After studying CSSDiff on several time frames:
It might be possible to form a 'forward trading plan' meaning enable or disable pair XY until it seems more appropriate:

The daily CSS can sometimes be a leading indicator - sometimes price falls while strength increases - entering just on CSS alone could lead to a 300 pip draw down before the rocket upwards of over 600 pips.

However, I have noticed that Sto (5,3,5) seems to work quite well as brake on entering at the wrong time.
Inhibiting the EA until the Daily Stochastic is favourable looks promising.

Also when their is a strong price-CSS divergence on the daily it is a reason to relax RSI thresholds for entries when the daily sto says its ok to go.

From the point of view of manual management of the EA:
Areas of super strong support and trend lines help identify better entries and targets,
and the intermediate levels consistently follow Fibonacci levels:
---------------------------------------------------------------------
Returning to this after further analysis:

The Sto & CSS on D1 can still be late to reflect the end of a range.
On D1 the only consistently helpful information are the strong SR levels, TL and candle patterns.
I picked EurCad for analysis because it is challenging for our EA. The problem with forward testing is that we don't get the full test coverage until we have been running for a very long time, but looking at EurCad over 12 months gives enough challenging price moves to do mental analysis.

[Jonathan2FI]

I like your thinking on find a forward trading plan with HTF CSS and select pairs accordingly - fits with NBs thinking. I also overlay a SR and TL picture and I currently use Evertrend with 3000 bar on h4 which picks up all the key TLs and horizontals without littering the screen.
Shame about can not get the dashboard and ea working reference GUIs. I am looking to make judgements beginning of week then end of day revisions (2200 GMT) and have an ea that applies NB rules 240lwma trend setting, RSI2 trigger pending, pivot entries and or 5lwma x and trade management. This ea fits very well (I know 5lwma x is not used that would be early entry discretionary trade otherwise as per EA) and may even be considered a sledge hammer for a nut (nothing wrong with that if it is reliable ) - should I revert to Steve's scripts or go back to 2.0 ? while dash is made more stable ?Jonathan

[snailbeard]

I like your thinking on find a forward trading plan with HTF CSS and select pairs accordingly - fits with NBs thinking. I also overlay a SR and TL picture and I currently use Evertrend with 3000 bar on h4 which picks up all the key TLs and horizontals without littering the screen.
Shame about can not get the dashboard and ea working reference GUIs. I am looking to make judgements beginning of week then end of day revisions (2200 GMT) and have an ea that applies NB rules 240lwma trend setting, RSI2 trigger pending, pivot entries and or 5lwma x and trade management. This ea fits very well (I know 5lwma x is not used that would be early entry discretionary trade otherwise as per EA) and may even be considered a sledge hammer for a nut (nothing wrong with that if it is reliable ) - should I revert to Steve's scripts or go back to 2.0 ? while dash is made more stable ?Jonathan

If you have several computers then you can try the dash board on different machines. One of those machines is five years old!
Two of my machines with no DLL/dashboard code are still running this morning but it is too soon to be sure that was the issue.
From 3 machines running demos I get 3 sets of pending orders or instant orders
Alpari and Oanda are similar but GP is quite different.
I'm using the pending orders as alternative to the dashboard.

[snailbeard]

Daily Forward trading outlook for USDJPY

I am choosing not to allow the EA to trade USDJPY yet.
The overall trend remains bullish, however, H4 CSSDiff (true) is currently bearish.
USDJPY typically has retracements between 200 & 400 pips, so it might continue via a third leg down, before continuing with the bullish trend,
but confirmation can be confirmed at the next real breakout from current SR band.
In retrospect, we can see channel & resistance areas where we could have disabled further buy orders.
One can also see the possibility of manually trading the retracement on breaking below Fibonacci 161.8 and targeting 38.2 based on H4-CSSDiff and prior knowledge of retracements in the range of 200 to 400 pips.

If I had an option in the EA for re-enabling trading it would be
a breakout value just above 38.2 e.g.: 101.375 combined with H4-CSSDiff moving towards bullish
double pairBuyEnableAtLevel[usdjpy]
and similarly the equivalent for sell orders
Also on option to inhibit trading at SR boundary:
double pairBuyDiasbleAtLevel[usdjpy] - no mores buys above this level
double pairSellDisableAtLevel[usdjpy]
And these would need to be checked and adjusted as necessary.

[snailbeard]

Applying the above logic to USDCHF

Oanda & GP are running a profit on this at the moment, but there is a chance that this will hit Fib 161.8 and reverse to continue down, perhaps to one of the strong support levels around 0.95130, especially as there is a trend line intersection at Fib 0 and CSSDiff -D1 shows weakening.

Interestingly, Alpari's EA killed the trade at 0.96176.
There is no explanation in the log, so I think it is important know what logic resulted in the trade being terminated on Alpari but not on Oanda or GP.

Therefore, I think the EA's target at .98037 is optimistic unless we see signs of strength on H4.

[snailbeard]

USDCHF H4 : CSSDiff is looking bearish
So according to my proposed additions this trade would have been inhibited by CSSDiff-H4.

However, thanks to Andy's profit lock the EA should save some pips if it goes down, avoiding the journey down to the stop-loss, which also happens to be a possible buy entry, if CSS strength returns.

I have sometimes made some nice pips by using NB10.4 stop-losses as entry points

[snailbeard]

Update on crashing issue

Four copies of Empty4 and EA running over 3 PCs:
(All running blind: no dashboard)

Still running

[Meta1]

Hi snailbeard,

do you mean that you don't use mt4gui ? If so, please can you share your version?
I test the Version 2.01b without dll and it works fine for me.

Best regards

[snailbeard]

Update on USDJPY

Yesterday, I said H4 CSSDiff was bearish, this morning it is still bearish.
However, price movement went the opposite way.

Conclusion:

CSSDiff's absolute position on H4 has no useful information.

Nevertheless, a close look at CSSDiff shows that the change of value during ranging might be useful, but generally, it just reflects what is already apparent in the price.

The most useful indicator was Fibo 38.2 which confirmed the breakout from Fibo 23.6.

AlpariUK demo was the only account to catch this move.
Oanda & GP sat on their hands
However, all 3 have a new open buy order from about 101.9

The RSI calculation seems to be too sensitive to give consistent results:
It won't be possible to predict the performance of the EA unless we have trigger which is less sensitive to small variations in data.

In the following attachment USDJPY confirms that the EA on Alpari made good entry:

[MrLong]

I'm just testing the final version, I've added the option to use the DLL or not, I've left the columns for
that purpose.

The use of CSS is the last CSS-1 must be be below 2.50 and current to show us the way above 3.00, into our directon above or below 240.

Also, I'm using without the RSI because I find they don't sync that often, use of the RSI is an option.