Supernova GBP/JPY Mini Trend Catcher

Quoting Pipsqueaker

Disliked

I am trading the 15m on GBP/JPY, AUD/JPY and EUR/JPY and picked up +20 last night on the AUD/JPY, and +143 on the EUR/JPY

Here is screen shot from live account and check out the trades. I am going to bump up the risk% after this week.

Due to the lot size I have been able to let it run over night and wake up to pips in the morning. I turn it on during asian as I get some good drift pips and manually close just before US session when I wake up to let the dogs out.

Ignored

Quoting rmyers

Disliked

Hi Matt and Dee,

Could you 2 please check you set files to make sure I selected Lag exit = true? I don't have the set files on this machine, so I can't check.

If I have, could you please check your charts to see if I the exit conditions are being met, but I am not handling it? I'm just worried that I have a bug since I don't see your trades closing.

Rick

Ignored

Quoting UFO Engineer

Disliked

Hi MY LORD GNAT !

This will be very nice ...

UFO

Ignored

Quoting wolfeman

Disliked

good stuff there Matt, will you keep us posted?? ECN + MT4, every G/J traders dream!! (And that's enough abbreviations for one night!) :

Ignored

Quoting orion70

Disliked

Thanks for the info.

I will look into the brokers you mentioned. What is particularly galling about ibfx is how they manipulate the spreads when there is a lot of volatility. I can understand why they would do that during news times as there is not time to "lay-off" trades to ECN's but their stop-hunting seems rather predatory.

When they stack the odds against you, it is hard to scalp for a few pips here and there. That's why I trade longer tf's and aim for 40-50 pips per trade.

Ignored

Quoting fxtradera

Disliked

My test statement (sntest7) is not updating at this point. I have this odd screen problem again that is impacting at least the MT4 screen displays (EAs still appear to be functioning ok) and I suspect that may be the reason here. I will have to reboot to resolve but I have so many MT4 sessions and EAs running right now that I will wait until a quiet market time later today to minimise any interrupts.

Anyway, a couple of other points. It's very early days and market conditions now are very different to much of last week. I've had some other demo accounts running in the background longer, and while they made good gains last week, they have been clobbered this week. This animal, at least on the 1hr time frame, works very well when it latches on to prolonged trends. Right now, there's a lot of chop and while it catches some of the gains, it is also slower at turning around when the market goes the other direction. As Rick has pointed out, upcoming testing of MM and locking in profits later should help there.

One week will help to iron out any bugs in entries/exits but don't immediately discount your indicators because of poor results - they may have worked really well last week or before. Personally, I'll be scrolling back at least a month to see how my personal set up looks. Later...

Ignored

Quoting rmyers

Disliked

Hi FxTradera,

You are absolutely right about not discarding indicators because of poor results. Isolating the indicators was the best way to find if each was working properly. As to performance of the individual indicators, that was secondary. One week on 3 pairs and 1 TF is a very limited test, but we needed to debug and gather a little baseline data. Next week we will open the throttle and observe the individual indicators on most TF's. That will not cause us to eliminate indicators either, but will allow us to assess the reactivity over different TF's. This could tell us, for instance, if using a M15 exit TF with an M30 entry looks promising. The week after, we will add in MM, ATR based lot sizing, etc. This will show the effect of these features on the pofitability and safety of the app. Finally, we will make it SuperNova again by adding in the indicator combinations for entry and exit.

Also, as I mentioned before, I have been testing a very reactive QQE that catches fast moving markets quickly. Should be interesting to see ho it works when I put it in SuperNova

Rick

Ignored

Quoting jasius

Disliked

I went ahead and downloaded SUperNova EA v1[1].4.zip from December 11, 2007. I saved all of the indicators and replaced EA with newest version SuperNova EA (Glenn) v1.7 Beta 6.

When I want to test it I want to specify entry and exit rules, but I am not able to change options from False to True. If I change them by clicking on EA and re-open them, I can see that those options are still in their default positions. Thus I can't specify the entry rules as dicussed in December 4, 2007 package You Better Read me.txt file.

So my questions are:
1. is using SuperNova EA (Glenn) v1.7 Beta 6 with indicators and templates from December 11, 2007 files OK?
2. How to change entry and exit rules in EA so the values are saved after I change them

thanks everybody, keep up the good job

Jonas

BTW, still can't figure out if these EAs work right... Backtesting in Metatrader over the past few months gave me slight decrease in assets, whereas I keep on seeing all the testers getting wonderful profits (Pipsqeek for example). Is it because of the old indicators and non-working entry-exit conditions?..

Ignored

Quoting Gnat

Disliked

Hallo Rick!

Took some time as this computer did not have the snagit so far. Here are the three snapshots and the setting file capture. There is something wrong - my opinion - on the EURCHF trade. Although it is still running nicely.

Have a look and let me know what you think!

Greetz!
Matt

Ignored

Quoting fxtradera

Disliked

Hi Rick

Comprehensive reply. As usual. Great!

I'll just throw this one out there to ask if anyone else has noticed or if you think it's worth the development time. QQE/CCI entry is often late. I'm toying on a seperate account with a QQE entry and CCI exit with a bigger period - this can keep you in when temporary retracements occur. However, if the CCI never gets over the 100 on a long in the first place and then goes south, then the exit wouldn't work and it might require additional coding to get out at, say, 0. Make sense? Worth it, possibly?

Just thinking on the fly here, I'll also play with the 100 level for CCI and change that value a bit to see if that can negate what I suggested above.

Or you want to hold any further code developments aside from bug fixes until after week 4? As I say, just throwing ideas out there...

Ignored

Quoting rmyers

Disliked

Hi FxTradera,

Those are good ideas. I think that one of the reason's Supernova has been successful is the use of multiple confirming entry/exit points. CCI works well with QQE, but the current rules we use makes it limited as a single exit criteria since a false signal can cause it not to work. I wonder what the efect would be of defining a channel for the positive and negative thresholds? We would still use the standar exit criteria, but let's say we were going short by taking the signl at the lower border of th channel. Our exit criteria would either be our normal exit or a cross above the upper part of the entry channel which would indicate a failed trend.

What do you think?

Rick

Ignored

Quoting rmyers

Disliked

Hi FxTradera,

Those are good ideas. I think that one of the reason's Supernova has been successful is the use of multiple confirming entry/exit points. CCI works well with QQE, but the current rules we use makes it limited as a single exit criteria since a false signal can cause it not to work. I wonder what the efect would be of defining a channel for the positive and negative thresholds? We would still use the standar exit criteria, but let's say we were going short by taking the signl at the lower border of th channel. Our exit criteria would either be our normal exit or a cross above the upper part of the entry channel which would indicate a failed trend.

What do you think?

Rick

Ignored

Quoting SimonJ

Disliked

Hey everyone!

I did backtest QQE to QQE only for GBP/JPY 4h for september 2007.
Because of delays of 1 candle in the buy and sell signals it does pretty much reversed of what we want it to do.
If these lines would be straight between the QQE signals this backtest would have done profit. I know alright that it waits for the candle to close so the qqe signal is confirmed. BUT! When we have QQE to QQE, don't we want to place a new order immediately after the previous one was closed? But it waits for this candle to close too before it places an order!

EDIT: OH shit I noticed after I wrote this that it is worse than I thought. Uploaded new pic!

http://i25.tinypic.com/2dwb2no.jpg

I have also done exit on CCI only and tried from 0 to 100 with 10 interval and 50 treshold was the best. This was on a backtest that was done same period as above but 1H TF. Did over 30% profit. I had AllowReEntry to true. Without that it was not in profit.
I will continue to do CCI exit backtests because I think CCI seems to be very good for exit.

Ignored

Quoting cashfinder

Disliked

As I discover for QQE entries (I am not sure other entry signal applies)

Case 1) if you have position before, it will close the position next bar to the exit signal. And the new entry will then be one bar later. So theres a one bar delay in your graph.

Case 2) if no position on hand , when QQE signal come, position will open on the next bar, so NO delay happens.

Conclusion: is it possible to make it stop and reverse in the same bar? if it can be done, the delay problem should be solved.

As for exit signals, I do not find delay, it closed at the next bar after the exit signal appears.

Ignored

Quoting orion70

Disliked

I'm not real sure how you envision using channels but this is what I use for live trading.

http://i5.photobucket.com/albums/y17...0/channels.jpg

Ignored