Fractals, ZZ & the Pissing Dog | Page 16 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

ECN && STP && Scalping && Hedging - Broker Review - Oct 2009 24 replies

Alert on Fractals (MT) 2 replies

Zigzag & Fractals 10 replies

EA wanted for sound alert depending on fractals condition 2 replies

Fractals Edge Fx 3 replies

Attachments: Fractals, ZZ & the Pissing Dog
Exit Attachments

Fractals, ZZ & the Pissing Dog

  • #301
  • Edited 10:51am Jan 23, 2014 10:39am | Edited 10:51am
Quoting gg53
{quote} killerno & candero, It's obvious that your algo isn't processing the "error" prediction as it should. Also, part of that process is to insert that error zz leg into the database and recalculate. And therefor I don't know how many such data items you failed to insert into the database during 2013. Maybe those are the reasons why you are getting such results - or maybe still missing explanation about what to do in error predictions that occur before the predicted swing point. So - I think that there is too much varibles missing from your side...
Hi GG,

thanks for the answer. The error prediction means if a confirmed ZZ appear more then 3 bars from my signal, thean I gonna calculate from this ZZ point. Yes probably we are missing the Handling Prediction errors 3 doors explanation and after we get in the longer legs better result. Unfortunately everyone has the same issue who I spoke with (emonts,burnsss,Carl..) for longer legs we are getting too early signals.
For next bar predicitons do you accept only the high probability prediction, I mean e.g. over 90% and also the total length probability should be over 50% or even over 70-80%..?
Thanks
  • #302
  • Jan 23, 2014 11:07am Jan 23, 2014 11:07am
Quoting gg53
{quote} killerno & candero, It's obvious that your algo isn't processing the "error" prediction as it should. Also, part of that process is to insert that error zz leg into the database and recalculate. And therefor I don't know how many such data items you failed to insert into the database during 2013. Maybe those are the reasons why you are getting such results - or maybe still missing explanation about what to do in error predictions that occur before the predicted swing point. So - I think that there is too much varibles missing from your side...
I think if our/everybodies algos where doing well, this thread would be quite lonely, place of frightening emptiness with no living souls around

If I take your results out of comparison, I'm expecting the results to differ between ourselves, because it is normal as we operate with a similar concept with different tools, knowledge and skill set. Furthermore, we can't forget the different uptake and understanding levels, so to speak, of everyone. The deviation can't be excessively large, though.

How can you say how many legs my algo has "failed to insert"? I don't use databases, I use arrays, fortunately I can work this way as speed is not an issue for me, other things are. Every ended leg gets inserted into my arrays, it learns as market goes, but in a walk-forward fashion, not knowing what lies ahead. I've used different lengths of data, at one point 100k more bars than you, it didn't change the behaviour. For me it proved your concept has a very strong backbone, but in a need to iron out some details on my end. I'm not sure how fundamental those details are, but work goes on.

G, don't take my "malfunction" comment personally, it wasn't meant to be that way, sorry. I respect you very much, and I've learnt a lot from you for which I'm grateful. I take my hat off for what you are doing here, I realise how much time, effort and energy goes into maintaining this thread. I'm not trying to spark a feud with wrongly turned out remarks.

C.
  • #303
  • Jan 23, 2014 4:27pm Jan 23, 2014 4:27pm
Another light bulb moment. The failure legs for ZigZag is shown by the indicator I posted that kept failed zigzag points. If we include the zigzag size as the pip distance from start to finish AND the fib retracement level from prior fractal then we are almost to the point of 1, 2, 3 paterns or Elliot waves. By checking the largest zigzag legs such patterns might emerge more clearly.

I also just realized I might be creating a basic neural net with my data structure. I am now looking into that possibility as well.

I have also modified my fractal indicator to include the form of the final 2 candles having lower highs for down and higher lows for up as well as higher highs or lower lows for the first 2 bars.
It then outputs the count of the fractal types. Next is to check the next bar after the fractal to see how often the bar after the fractal reverses or continues in the same direction. Maybe even check the 2nd and 3rd bar.

I plan on going the same for the ZigZag patterns.

Robert
  • #304
  • Jan 23, 2014 11:11pm Jan 23, 2014 11:11pm
Quoting MrPip
Another light bulb moment. The failure legs for ZigZag is shown by the indicator I posted that kept failed zigzag points. If we include the zigzag size as the pip distance from start to finish AND the fib retracement level from prior fractal then we are almost to the point of 1, 2, 3 paterns or Elliot waves. By checking the largest zigzag legs such patterns might emerge more clearly. I also just realized I might be creating a basic neural net with my data structure. I am now looking into that possibility as well. I have also modified my fractal...
now about the pip distance is what my mind has as i deal with manual insertion of data.am not yet where a lot are not even closer but that pip thing still ringing in mind as i do the candle counts on zigzag
  • #305
  • Jan 24, 2014 9:08am Jan 24, 2014 9:08am
Quoting killerno
{quote} Hi candero, For me the red "circle" ZZ is valid 5 bars after it when the low fractal after it confirmed.(you can see in the picture). I don't know what we are missing. regards
killerno,

Going back to that example: ZZ position is confirmed only 5 bars later.
Following that confirmation point I can't find in my DB a valid pattern for swing signal that contains only 5 down bars.

G.
  • #306
  • Jan 24, 2014 9:19am Jan 24, 2014 9:19am
Quoting gg53
{quote} killerno, Going back to that example: ZZ position is confirmed only 5 bars later. Following that confirmation point I can't find in my DB a valid pattern for swing signal that contains only 5 down bars. G.
hi GG,

yes it is confirmed in the 5th bar after it. So it means you calculate from this point where it get confirmed? I tought you calcualte from the red circle point, hence that was the confirmed ZZ end point.
thanks

Attached Image (click to enlarge)
Click to Enlarge Name: error2.png Size: 21 KB
  • #307
  • Jan 24, 2014 11:44am Jan 24, 2014 11:44am
Quoting killerno
{quote} hi GG, yes it is confirmed in the 5th bar after it. So it means you calculate from this point where it get confirmed? I tought you calcualte from the red circle point, hence that was the confirmed ZZ end point. thanks {image}
killerno,

No, I start calculation from the confirmed ZZ in cases of "error" predictions.
What I meant was that I don't have a swing pattern with "01" and that length and a "next bar" with high probability against longer pattern at that point.

G.
  • #308
  • Jan 24, 2014 12:56pm Jan 24, 2014 12:56pm
Hi Carl,

At old thread, page #95, post #1898 : you wrote this.

Quoting crodzilla
{quote} burnssss, Yes. SET1 - Bar DNA and counts are filtered by SET2. Ensure that SET1 counts are also part of SET3. Carl
What do you mean with the "counts" ? Is it the COUNT(*) result in database (amount of rows that appear with specified pattern) ?
Example:

SET1 : [1 0 1 0] = 178 appearance
SET2 : [1 0 1 0] = 33 appearance
SET3 will be [1 0 1 0] with 178 appearance (NOT 33 appearance)

SET1 : [1 0 1 0 0] = 68 appearance
SET2 : not found any pattern of [1 0 1 0 0] = 0 appearance
SET3 : None. So, it will be 0 appearance, pattern [1 0 1 0 0] is not valid.

Is that right ?

Thanks
-smallcat
  • #309
  • Jan 24, 2014 1:22pm Jan 24, 2014 1:22pm
Ok, I think it's time for practical application.
The question on hand is something like this:

Can you roughly double your account within a month?

Not using Martingale and other "crazy" things - just logical things but with totaly different aproach.
Forgetting all that you have been taught about Risk/Reward and Money-Management.
Using those ideas - but re-thinking their logic and stop quoting other "Guru's" numbers related to them.

The only things that you have is:
1. Basic Forex market knowledge.
2. Statistical knowledge.
3. ...and a little "Edge"

What is the best Risk/Reward?
What is the best Money Management rules?
What to trade?
Don't bother to quote the ordinary answers - I think I know them already...

Is this challenge, under those conditions, possible and logical?
I'm going to try it, publicly, using my little "edge", Renko charts and "weird" statistical rules - starting next week.
Your help, advice and insights on the subject are more than welcome.

G.
  • #310
  • Jan 24, 2014 3:17pm Jan 24, 2014 3:17pm
Quoting gg53
Ok, I think it's time for practical application. The question on hand is something like this: Can you roughly double your account within a month? Not using Martingale and other "crazy" things - just logical things but with totaly different aproach. Forgetting all that you have been taught about Risk/Reward and Money-Management. Using those ideas - but re-thinking their logic and stop quoting other "Guru's" numbers related to them. The only things that you have is: 1. Basic Forex market knowledge. 2. Statistical knowledge. 3. ...and a little "Edge"...
Can you roughly double an account ?

Trade-explorer is set up and waiting to start on next week open.
Strategy:
Little "edge" applied on Renko charts, statistical tools and "Out-of-the-Box" thinking on Risk/Reward & Money-Management.

Wish me luck...

G.
  • #311
  • Jan 24, 2014 3:18pm Jan 24, 2014 3:18pm
Quoting gg53
Ok, I think it's time for practical application. The question on hand is something like this: Can you roughly double your account within a month? Not using Martingale and other "crazy" things - just logical things but with totaly different aproach. Forgetting all that you have been taught about Risk/Reward and Money-Management. Using those ideas - but re-thinking their logic and stop quoting other "Guru's" numbers related to them. The only things that you have is: 1. Basic Forex market knowledge. 2. Statistical knowledge. 3. ...and a little "Edge"...
G, I was wondering how long it would take you to get to the implementation stage!

I would trade this strategy the way it was designed in its indicator form - as a pure reversal system. This obviates risk/reward. Simply determine how much you want to bet.

To determine bet size I would calculate the maximum historical system drawdown (in pips) and convert drawdown to dollars in terms of your initial deposit. For instance, if your initial deposit is $1000 and the system max drawdown is 500 pips, then in terms of your initial deposit your drawdown at 1x leverage is (500 pips) / 10000 * ($1000 initial balance) = $50 or 5% drawdown on account.

Determine the maximum percentage drawdown on the account you are willing to endure. For instance, if you are willing to accept 30% maximum account drawdown, then trade the system at 6 times leverage (or less).

For ease of implementation, update trade size weekly (trade size = equity * leverage). If your system is doing really well mid-week, update (increase) trade size 2-3 times per week. This extra increase in trade size can add 20-30% increase in returns to your account. Don't ever decrease trade size mid-week to avoid account drag (http://www.forexfactory.com/showthread.php?p=7095962#post7095962).

Best of luck!
  • #312
  • Jan 24, 2014 3:25pm Jan 24, 2014 3:25pm
Quoting gg53
{quote} Can you roughly double an account ? Trade-explorer is set up and waiting to start on next week open. Strategy: Little "edge" applied on Renko charts, statistical tools and "Out-of-the-Box" thinking on Risk/Reward & Money-Management. Wish me luck... G.
You probably don't need much luck. But good luck
  • #313
  • Jan 24, 2014 3:40pm Jan 24, 2014 3:40pm
Quoting FXEZ
{quote} G, I was wondering how long it would take you to get to the implementation stage! I would trade this strategy the way it was designed in its indicator form - as a pure reversal system. This obviates risk/reward. Simply determine how much you want to bet. To determine bet size I would calculate the maximum historical system drawdown (in pips) and convert drawdown to dollars in terms of your initial deposit. For instance, if your initial deposit is $1000 and the system max drawdown is 500 pips, then in terms of your initial deposit your drawdown...
FXEZ,

You mentioned average DrawDown - based on what strategy?
I don't think that implementing that indicator on Renko charts can be called "reversal system".

Account size is $10,000 and needs to be doubled within roughly a month.
That's ~20 trading days, ~$500 a day profit.
Calculations and statistics start from that point.

G.
  • #314
  • Jan 24, 2014 4:04pm Jan 24, 2014 4:04pm
Quoting FXEZ
{quote} G, I was wondering how long it would take you to get to the implementation stage! I would trade this strategy the way it was designed in its indicator form - as a pure reversal system. This obviates risk/reward. Simply determine how much you want to bet. To determine bet size I would calculate the maximum historical system drawdown (in pips) and convert drawdown to dollars in terms of your initial deposit. For instance, if your initial deposit is $1000 and the system max drawdown is 500 pips, then in terms of your initial deposit your drawdown...
FXEZ,

Your account drag (http://www.forexfactory.com/showthread.php?p=7095962#post7095962) and money management article is one of the BEST I've ever read!
Thanks.

G.
  • #315
  • Jan 24, 2014 5:12pm Jan 24, 2014 5:12pm
Quoting gg53
{quote} Can you roughly double an account ? Trade-explorer is set up and waiting to start on next week open. Strategy: Little "edge" applied on Renko charts, statistical tools and "Out-of-the-Box" thinking on Risk/Reward & Money-Management. Wish me luck... G.
GG, cant wait to see it happen as i believe it is possible. Best of luck
  • #316
  • Jan 24, 2014 8:11pm Jan 24, 2014 8:11pm
Quoting gg53
... Can you roughly double your account within a month? ...
Wow per month. That is not just 'a little edge' anymore. When I saw your screenshots back in the old thread, I remember thinking doubling per year. I slapped myself out of that split second daydream almost immediately.
  • #317
  • Edited Jan 25, 2014 1:27am Jan 24, 2014 10:42pm | Edited Jan 25, 2014 1:27am
Quoting gg53
{quote} Can you roughly double an account ? Trade-explorer is set up and waiting to start on next week open. Strategy: Little "edge" applied on Renko charts, statistical tools and "Out-of-the-Box" thinking on Risk/Reward & Money-Management. Wish me luck... G.

G,

I wish you Good Luck.

Let success be yours.
  • #318
  • Jan 25, 2014 12:00am Jan 25, 2014 12:00am
Quoting gg53
Ok, I think it's time for practical application. The question on hand is something like this: Can you roughly double your account within a month? Not using Martingale and other "crazy" things - just logical things but with totaly different aproach. Forgetting all that you have been taught about Risk/Reward and Money-Management. Using those ideas - but re-thinking their logic and stop quoting other "Guru's" numbers related to them. The only things that you have is: 1. Basic Forex market knowledge. 2. Statistical knowledge. 3. ...and a little "Edge"...
Hi GG53,


Some questions more of us will have.
What we will do with what we did up to now. It was only a training, a warming up ? We still go on with the "old" system and in parallel we will do more or less the same based on Renko bars. Can you tell us some more about your plans/roadmap.

And of course always curious to new systems.

Wish us all luck.

BR.
Huub
It is not important how often you fall down, but how often you get up again
  • #319
  • Jan 25, 2014 12:53am Jan 25, 2014 12:53am
gg53

Thanks for starting a TE. It really makes you legit. Now matter what happens going forward. Everyone can see it. That alone brings honor to what you are trying to do.

Serious people look at whether or not you are willing to expose your trading in real time, with a neutral arbiter.
If you are good at something, never do it for free--Joker
  • #320
  • Jan 25, 2014 2:36am Jan 25, 2014 2:36am
Quoting emonts
{quote} Hi GG53, Some questions more of us will have. What we will do with what we did up to now. It was only a training, a warming up ? We still go on with the "old" system and in parallel we will do more or less the same based on Renko bars. Can you tell us some more about your plans/roadmap. And of course always curious to new systems. Wish us all luck. BR. Huub
emonts,

I'm still here, not going anywhere...
I said from day 1 that I'm also working and testing Renko/Range bars.
It's exactly the same system and algorithm - but I think it will bring better yield with some innovative R/R & M/M.

G.
0 traders viewing now