It will open a long on the open of bar after the close of volume above the threshold.
Rick
Supernova GBP/JPY Mini Trend Catcher
- #4,683
- Jan 29, 2008 2:17pm Jan 29, 2008 2:17pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Hi Matt,
I think I have it figured out. For this example I am looking at the 3rd image in your post. If you have a short signal, the Lag logic in the EA says exit when the second Lag crosses back up from below 0.15 to above 0.15, signal the exit. But because the trend was not well established, Lag2 never made it low enough to come back up and give the signal. Three things about this:
1. This is why QQE is so good with the other indicators. It would have gotten you out with a small loss.
2. Lag is really reactive on small TF's, but the values we use as defaults do not seem reactive enough for the 1 hour. It might be interesting to modify the lag 2 indicator to see if you can find a good 1 hr signal. Then look backwards and see if the value applies equally well on many trades.
3. If Lag2 crosses above the 0.85 mark, then you also have no trend and we should pobably make that the alternative exit criteria rather than modifying the value on my above point. That would have gotten you out even earlier than QQE
So Matt, could you please look back and see if my point 3 applies to other false trends? If it does, I will generate a new beta with that test and we can try it out this week.
Rick
Crazy day today, so I am late once more. I will look into the LaG issue and will report later on. I will try the alternative settings from FX_Ninja and also look into the past for other occassions to get point 3. sorted out.
Be back in a few timespans!
Ciao
Matt
- #4,687
- Jan 29, 2008 4:20pm Jan 29, 2008 4:20pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
3. If Lag2 crosses above the 0.85 mark, then you also have no trend and we should pobably make that the alternative exit criteria rather than modifying the value on my above point. That would have gotten you out even earlier than QQE
So Matt, could you please look back and see if my point 3 applies to other false trends? If it does, I will generate a new beta with that test and we can try it out this week.
Rick
Hallo Rick!
It seems that this is really worth a try, as otherwise trades would be going on for a very long time, although the objective observer would be screaming to stop it!
If that is possible to add to the code it would truly add value to the Laguerre issue. I have skimmed through a lot of pairs on the search for these scenarios and have found that the less volatile the pair, the longer one stays in with this combo - even against all odds!
Will go to bed now as tomorrow will be another kids morning for me!
G'night!
Matt
- #4,689
- Jan 29, 2008 5:30pm Jan 29, 2008 5:30pm
Hallo Rick!
It seems that this is really worth a try, as otherwise trades would be going on for a very long time, although the objective observer would be screaming to stop it!
If that is possible to add to the code it would truly add value to the Laguerre issue. I have skimmed through a lot of pairs on the search for these scenarios and have found that the less volatile the pair, the longer one stays in with this combo - even against all odds!
Will go to bed now as tomorrow will be another kids morning for me!
G'night!
Matt
Yes, that is what I noticed. Lag seems to do a good job of highlighting broken trends. Of course, Supernova's strength is the use of multiple indi's to achieve exits that baffle individual indis. We are not SuperNova'ed right now and it is annoying to watch our EA when we have taken away all its strengths. It will be a lot more fun to watch it when we make it SuperNova again
Please let me know what your research finds. The changes we are talking about seem to be very useful and that is one of the reasons for this week of ind isolation.
Rick
Hey Ufo!
Do we have another nice pikkupakkasta day in Finland again? Hope you are doing fine. Seems the link is broken up there, at least I cannot get the file!
Hope to see it later then!
See yu!
Matts
http://www.freespaces.com/freemind/i...ntasy/gnat.jpg
- #4,694
- Jan 30, 2008 2:03am Jan 30, 2008 2:03am
From what I have researched this is a result of multiple orders flooding the broker server from your PC it will lock you out of everything except hitting SL/TP for anywhere from seconds to minutes. This is something we need to address as if the EA sends an order on multiple pairs and get this lock out it won't try again until the close of the Bar up to your maximum cross setting.
So I have had it fill 3 bars later. I think we need to add code that will continue to attempt to fill during that bar. What do you think Rick???
Yes, that is on my list. On other EA's I have written, I set a flag on failure and retry on each new tick. I have not retrofitted that code to this EA, but it will greatly improve EA slippage (broker slippage is bad enough
).Rick
- #4,696
- Jan 30, 2008 3:17am Jan 30, 2008 3:17am
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Hallo Rick! 
I am still in all three initial trades and will maybe drop put of one or two during today and will post the pictures then. I had some intermediate snapshots ready, but did not want to post hypothetical mumurings here. In sideways markets the LaG is definitely not such a smashhit, but it was initially added to aid the very small TFs as on Imran's system.
You are right about your statement that there should be no frustration about the dradown of this testing session. First of all we are not using Supernova in an intelligable way and are just testing single functions. Then we are using a pair that has so far never qualified for Supernova usage with an EA - the EURCHF - and it is not even doing bad on that. Furthermore the only setting I would have given a try initiall is the one of Carol, probably followed by the dot-to-dot from Rick.
I see the whole thing moving amazingly smooth for so many people involved directly and on the stages, so that we may come to conclusions within a few weeks here. Thanks Rick for adopting us for your EA coding as you may have turned elsewhere by chance and we would have never made it this far. And also special thanks to Glenn, who never let go the EA idea and was stubborn enough to bring it so far!!
I actually would need more sleep now, but I opt for the second best option - coffee! I made a big one! Please help yourself!
Ciao & Good morning from Old Europe
Matt
I am still in all three initial trades and will maybe drop put of one or two during today and will post the pictures then. I had some intermediate snapshots ready, but did not want to post hypothetical mumurings here. In sideways markets the LaG is definitely not such a smashhit, but it was initially added to aid the very small TFs as on Imran's system.
You are right about your statement that there should be no frustration about the dradown of this testing session. First of all we are not using Supernova in an intelligable way and are just testing single functions. Then we are using a pair that has so far never qualified for Supernova usage with an EA - the EURCHF - and it is not even doing bad on that. Furthermore the only setting I would have given a try initiall is the one of Carol, probably followed by the dot-to-dot from Rick.
I see the whole thing moving amazingly smooth for so many people involved directly and on the stages, so that we may come to conclusions within a few weeks here. Thanks Rick for adopting us for your EA coding as you may have turned elsewhere by chance and we would have never made it this far. And also special thanks to Glenn, who never let go the EA idea and was stubborn enough to bring it so far!!
I actually would need more sleep now, but I opt for the second best option - coffee! I made a big one! Please help yourself!
Ciao & Good morning from Old Europe
Matt
- #4,697
- Jan 30, 2008 3:20am Jan 30, 2008 3:20am
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Hallo Ufo!!
Wow - another fine art from your smithy there. Beautiful monsters you have there, ROTFL
- are you insinnuating anything by that? LOL
Thanks for that and whereever these guys are dragging me, I hope there is a bed!!
Ciao!!
Matts
Wow - another fine art from your smithy there. Beautiful monsters you have there, ROTFL
- are you insinnuating anything by that? LOLThanks for that and whereever these guys are dragging me, I hope there is a bed!!
Ciao!!
Matts
- #4,700
- Edited 4:05am Jan 30, 2008 4:03am | Edited 4:05am
I wonder if the QQE could be coded to exit at the moment it crosses, rather than waiting for the candle close. That could save a lot of pips; then, when the candle does close we would have a fresh signal for the next trade. Anyone try this?
I know the QQE will sometimes cross, then uncross while a candle is open, but from what I have noticed that is just before it makes it true "cross". Might be worth looking into.
It might be a strech for ME to suggest a change to the heart of our system; but I know I have many times looked at a closed trade thinking "damn, if only it got out here, and opened a new one here, that woulda been perfect"
Thanks rmyers that you can do it.
Without this "qqe delay" and the other I mentioned earlier it would make much more pips