A delicate secret | Page 28 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

The Secret: How to break through your psychology 8 replies

Why "x" a secret? 39 replies

The Secret of Reduced Market Spreads, by Joe Ross 2 replies

Winning secret 2 replies

Twoblink and "The Secret Sauce" 79 replies

Attachments: A delicate secret
Exit Attachments

A delicate secret

  • #541
  • Jun 13, 2014 4:33pm Jun 13, 2014 4:33pm
Quoting handy148
Carl - thanks for version 6. A question - if you you went back say 10 years on some pairs and still found levels that hadn't been returned to wouldn't that rather disprove the theory to some extent?
Disprove what?
I think we already know we don't know what EURUSDD is doing (at least I don't). And he's not going to tell us.
So, we can't disprove his theory.

I think we are trying to find a way to do something similar.
Even if its close, we should be able to strategize a way using various trading tactics and money management to find something of our own.

Carl
  • #542
  • Edited Jun 14, 2014 9:41am Jun 13, 2014 5:06pm | Edited Jun 14, 2014 9:41am
Version 7:
(I hope these stats are useful)

- Added: Level cleared - Cumulative BULLISH/BEARISH pips potential.
- Added: Level cleared - Cumulative BULLISH/BEARISH average per level potential.

Hopefully someone can do some interesting analysis this weekend.
I will be out of pocket this weekend.

Good Luck to all, have a great weekend and Happy Fathers Day!

Carl

- Removed buggy indicator
  • #543
  • Jun 13, 2014 5:24pm Jun 13, 2014 5:24pm
Sorry, tiny but significant bug found...

Please update to Version 7a.

Thanks.

Carl
Attached File(s)
File Type: mq4 !crodzilla-Price action V7a.mq4   9 KB | 923 downloads
  • #544
  • Jun 13, 2014 5:47pm Jun 13, 2014 5:47pm
THANK YOU CARL!!!!!
cheers
hannele
  • #545
  • Jun 13, 2014 6:00pm Jun 13, 2014 6:00pm
Quoting crodzilla
{quote} Disprove what? I think we already know we don't know what EURUSDD is doing (at least I don't). And he's not going to tell us. So, we can't disprove his theory. I think we are trying to find a way to do something similar. Even if its close, we should be able to strategize a way using various trading tactics and money management to find something of our own. Carl
eurusdd's theory was that it should return to that level before leaving the screen... which is a lot less than 10 years. Basically for eurusdd it was supposed to return to the level in about 24-30 time periods 87% of the time.

Basic probability, the closer the price is to the mean between the highest high ever and the lowest low ever the more likely it is that the price will return to that level.
  • #546
  • Edited 2:07am Jun 14, 2014 12:14am | Edited 2:07am
Quoting Thoughts
{quote} If we think of the levels as providing signals then only one strategy is clear to me at the moment but I suspect its a 'dangerous' one. (i.e. placing entry and exit pending orders at x and y points of the 'signal' candle; with the SL above/beneath the nearest pronounced support or resistance zone). However I surmise the R:R ratios and win-rate would be terrible but I could be very wrong since I've no tests done. If we think of the levels as profit targets then we can use them as a rough measure for R:R ratios with signals generated from...
thank you for reply, thoughts!

yes, the biggest problem of this analysis is we can only place the target level. nothing suggests about entry level.
because of this, we may be in the situation of huge drawdown for a long time.
maybe grid trading, though i have never seen profitable grid system and i dont want to live with drawdown...
(i have never traded real forex and so i have no trading strategy currently)
maybe hedging (place buy and sell at the same time), i have to be very careful about money management... not sure
edit: or we may be enter the opposite direction to correct the levels.

still thinking

ps: i got some signals lately. so i placed buy/sell order by intuition (no indicator, no price action, just intuition) with tp target.
both were ok, but im sure i will never do such trades with real account
Attached Image (click to enlarge)
Click to Enlarge Name: euu.png Size: 26 KB

australia lost...
  • #547
  • Jun 14, 2014 7:34am Jun 14, 2014 7:34am
Quoting Eurusdd
In order to cater for day-traders: most of us day-traders anyway, I have a plan to reach target within a day. A modification across all time frames can have target reached within a day. The following pictures show that.
I'm not sure if there is an edge with this method but I'm curious to figure out the logic and concepts behind these levels.

Has anybody noticed that the logic for the white and yellow arrows seems to have changed significantly after this post (post 83)? I can mostly recreate his earlier levels.

I dont think anybody is using the levels he is showing in his earlier charts.
Attached Image (click to enlarge)
Click to Enlarge Name: eurusddaily.png Size: 28 KB
  • #548
  • Edited 8:57am Jun 14, 2014 8:25am | Edited 8:57am
alearner,
Yes, things have drifted a bit. Carl's created a good indicator which was extremely generous on his part. Would you share what you have? My first impression was that this is a very simple approach and is nearly analysis free, except for the deviation issue.
Swamper
  • #549
  • Edited 1:09pm Jun 14, 2014 8:44am | Edited 1:09pm
Quoting crodzilla
Sorry, tiny but significant bug found... Please update to Version 7a. Thanks. Carl {file}

Thanks a lot for all your work, Carl !
Great and nice from you to share it.

I think that many of us (me for sure) will cooperate with analysis and comments, seeing all the effort you put in this climbing development.

  • #550
  • Jun 14, 2014 3:38pm Jun 14, 2014 3:38pm
I’m having a look at Carl’s indi, and while I look at it I am writing some thoughts to share, hoping they may be of any help for our further common thinking processes.
I post a H1 Eur/Usd chart as a reference, but I am looking at several TFs.

1) Drawdown and Trend

Reading: “ Avg 44.8 pips per level potential “ (Bullish) I think:
“ok, that could mean that I could have an edge entering a trade at around 40 pips distance from the level”

Then I read: “Largest open Bullish Deviation 515,1 pips” and I think:
“ ok, my enter for potential 40 pips profit could see a DD of 500 pips…so I should perhaps consider a MM based on adding long entries on my position until possibly 500 pips... ? "

Same for Bearish, with other numbers, of course.

As Carl said, I also think that we could / should add to our Edge what most Traders always say: “go with the trend.”

Meaning somehow that until the main Trend clearly changes we should consider only or mainly the levels created in its direction, as our objective expectation is price making new HH or LL, dipending on Trend's direction.

Ergo: entering a trade in direction of the main trend when price moved away from the last created level is more or less trading retracements, which is a famous and used way of trading.

Our Edge would be that we base on statistics our Entry and Exit points.

2) Time Factor

Another interesting Value is the “ Average Bars to Clear “.

I have these numbers with my broker:

M5 – Bars checked 4777 – average to clear 40 bars – Highest Nr of bars to clear 612
M15 – Bars 2957 – Average 20 – Highest 246
M30 – Bars 2500 – Average 38 – Highest 1288
H1 – Bars 2581 – Average 37 – Highest 832
H4 – Bars 6912 – Average 64 – Highest 3221
D1 – Bars 2057 – Average 33 – Highest 493

Of course I think I will have to check much more bars of history, but for the moment (even perhaps including H4 which has the highest Average) I think:
" the average is good, it is for sure within a screen with a reasonable scale."

Anyway the Nightmare of having to wait 832 bars on H1 before the levels gets cleared is still there..

Alone on these numbers of average there would be a lot to study and they could probably bring some edge too.

However the perhaps most interesting number I found is the Sum of :

“High Levels Cleared” + “Levels not cleared ABOVE”
Compared to:
“Low Levels Cleared” + “Levels not cleared BELOW”

Compare it yourself on your different TF and I wonder if you will also find out that most times they are either EXACTLY the SAME Number or very near to it….

Does this Hide any Concept I am missing ?

Is there perhaps really something in the REASON why those levels are cleared ?

Is it just a case that the number of Levels is more or less the same ?

Does this only mean that at the end of the day, the Whole Market actually only moves in a certain Range at all ?

Sorry for the long post, I usually jump over myself when I see something like this ...
Attached Image (click to enlarge)
Click to Enlarge Name: H1.jpg Size: 161 KB
  • #551
  • Jun 14, 2014 3:59pm Jun 14, 2014 3:59pm
This is just another thought.

I am comparing Carl’s and TomCat’s Indis.

Somehow both of them consider that price could be retesting the previous High or Low before proceeding over.

TomCat’s Indi seems to me drawing levels at the close of the first reversal candle after any swing, if the signal candle is confirmed by the next one closing in the same direction.

Crodzilla’s instead uses the High and Low of some candles, more or less with the same concept (in fact many signal candles are the same for both) but I have not 100% clear how it work.

Perhaps only if there is a certain nr of pips in the signal candle ?

TomCat’s strategy seems to generate more signals, even if some are so small moves that for profiting from them my MM should consider entering a trade already with Price moving away for just a few pips.

I wonder if we could get the same statistics also on TomCat’s Indi for further studying.

Anyway if we can have a clear Idea of Trend Bias (and also in Ranging Markets) results look for both very interesting.

At least there is something we can say for sure:
These indi do not repaint.
When the signal candle is closed, price remains there where it was on the chart
Attached Image (click to enlarge)
Click to Enlarge Name: Confronto.jpg Size: 186 KB
  • #552
  • Jun 14, 2014 5:38pm Jun 14, 2014 5:38pm
Grin's Axiom: A horizontal level drawn on a chart will be either reached or not. :nerd:
  • #553
  • Jun 14, 2014 6:34pm Jun 14, 2014 6:34pm
Hey Carl, nice to see you in another thread. Thank you for your contributions and indicator


@Whenever/All:
I'll give my 2 cents in here and say that I'm not a fan of using averages when it comes to a stat modeled 'system' for the elementary flaw that extremes pull the numbers too high. If there is an edge, my belief is that you'll see a skewed distribution(1 tailed) rather than a bell curve(2 tailed); At least, this is my experience with the statistics that I've done (not related to this project). Averages are ok, until you put in the SD on there and the SD turns out to be just as big as the average! Not a good sign in my opinion. There are more reliable methods of performing analysis on such data.

I'd love to take a look at more of it myself in excel, but my mt4 skills are about as good as my cooking skills-bad. Something something filewrite()..
"Mrrraggglhlhghghlgh"
  • #554
  • Jun 14, 2014 7:07pm Jun 14, 2014 7:07pm
Quoting LiquidGenius
Hey Carl, nice to see you in another thread. Thank you for your contributions and indicator @Whenever/All: I'll give my 2 cents in here and say that I'm not a fan of using averages when it comes to a stat modeled 'system' for the elementary flaw that extremes pull the numbers too high. If there is an edge, my belief is that you'll see a skewed distribution(1 tailed) rather than a bell curve(2 tailed); At least, this is my experience with the statistics that I've done (not related to this project). Averages are ok, until you put in the SD on there...
Thanks, LG. I don't really have a lot of time, but I am intrigued by this for some reason.

LG, what data would you need?
I had thought about exporting all historical levels, when they cleared, max deviation, etc. to file to help someone with matlab or similar.
But, I don't know anyone that knows how to use matlab (i do, but he's not interested).

Carl

PS. Yea, couldn't resist... Thanks for looking Whenever. Interesting stuff. Write as much as you'd like...!
  • #555
  • Jun 14, 2014 8:11pm Jun 14, 2014 8:11pm
Hi All, Hi Carl,
I have reread some first pages of that thread, and i've probabbly missed the point...
On one of EURUSD image chart , there is an add . "pointzero.com", so I went back to seach and visit that page...
There is a page called PZ AB_CD . That could be a simple retracement.
There is an interesting page here too: http://www.mql5.com/en/market/product/3918
Google "1_2_3 retracement" too and you"ll be closer than we though of EURUSD point of view.
Just guessing.

Anyway have a good WE.
Tomcat98
  • #556
  • Edited Jun 15, 2014 2:02am Jun 14, 2014 8:28pm | Edited Jun 15, 2014 2:02am
Quoting swamper
alearner, Yes, things have drifted a bit. Carl's created a good indicator which was extremely generous on his part. Would you share what you have? My first impression was that this is a very simple approach and is nearly analysis free, except for the deviation issue. Swamper
In initial posts I think white arrow is whenever you have a higher high or a lower low.

Yellow arrow (down) may not follow immediately after a white arrow (down). Similarly for Yellow arrow up.
Yellow arrow (down) only appears if their if a lower high than prev candle.
Yellow arrow (up) only appears if their if a higher low than prev candle.

Levels are drawn from Yellow arrows.

Like any trade its all about percentages.
  • #557
  • Jun 14, 2014 8:33pm Jun 14, 2014 8:33pm
Quoting crodzilla
{quote} Thanks, LG. I don't really have a lot of time, but I am intrigued by this for some reason. LG, what data would you need? I had thought about exporting all historical levels, when they cleared, max deviation, etc. to file to help someone with matlab or similar. But, I don't know anyone that knows how to use matlab (i do, but he's not interested). Carl PS. Yea, couldn't resist... Thanks for looking Whenever. Interesting stuff. Write as much as you'd like...!
Same, I guess these EURUSDD based threads must be addicting =p

For basic analysis what you suggested is perfect. Just a compiled list of:
Time to clear
Max drawdown(and thus max pip potential)
OHLC of the 2 bars (signal and hit bar)
(csv format)

However other than that, there are a LOT of potential things worth looking at, that I think can just be done by watching the actual indicator before exporting the data elsewhere, such as where the level is appear relative to the day's min/max, it's current short term momentum, OHLC of the bars at the max DD points and things of that nature.
"Mrrraggglhlhghghlgh"
  • #558
  • Jun 15, 2014 12:50am Jun 15, 2014 12:50am

  1. http://assets.forexfactory.net/nfs/c...ar299948_1.gif Eurusdd (http://www.forexfactory.com/eurusdd)
  2. Joined Oct 2012 | Status: Member | 1,388 Posts (http://www.forexfactory.com/search.php?do=process&provider=Member&searchuser=299948)

omega appears first on 4hr, 10pips from price. then within 30mins it should appear sequentially on the rest. once it appears on 1min, price is ready to go. by that time, price should have closed right above/below a certain level. this is what appears are consolidation to many. I know what they r doing. Most of the time. for a buy, 1hr must close above the number I gave
Two options: UP or DOWN!!! That should be EASY.

  • #559
  • Jun 15, 2014 2:37am Jun 15, 2014 2:37am
OH DEAR ME,

Ive stumbled upon this thread by fait or misfortune .

THIS VERY QUEST DROVE ME TO THE BRINK OF INSANITY.
I left it behind some time ago.

my verdict then; personally I could not find the mathematical solution . i believe there is one.

lets make some sense of what we are actually trying to do.

lets not allow FALSE ENLIGHTENMENT TO BLUR OUR VISION . i say this because there are many.

THE SECRET WITHIN THE CANDLE. there are many little secrecies within the price , they show firstly on the micro levels , the 1 min time frame is the the best visual of this.As time frameS moves up they become less obvious . There are many ways in which the observer can find these .
The false enlightenment at this stage is having random points on your charts.


using a 5 minute time frame to illustrate .the average signals generated for the hours between London open and new York close is 2,5.to compare on a 1 minute time frame the average is 5.this average is calculated using a 10 pip minimum distance from signal level.

ill confirm eurusd ,s claim .the majority of the time price will revisit the signal candle in this 5 min example within 12 hours.i would have to refer to my past records to give the average true % value.this value is important when the time comes for somebody to calculate the ratio.the grail itself.

again with out my logs, the approximate % returns are . based on 6 months .

10 to 20 pips from signal level 45 %

10 to 50 pips ....................... 80 %

10 to 100 pips...................... 93 %

10 to 500 p1ps.................... 98 %

now these are calculated excluding the time factor.

if every signal is taken (which is a requirement if we intend to break the maths)my method was to trade each signal at 10 pips,2nd entry at 20 pips,3 rd entry at 50 pips 4 entry at 100 pips.the way I worked my entry was to maximise on the sweeping of pips.

The false enlightenment .

you can have a period of days weeks or months were you are collecting thousands of pips,but as a rule the runaways which are overpowered by the next time frame up and so forth will reduce the sum back to 0.

now you can trade this implementing discretionary tools and do fine.but my objective is to find the way in which we can skewer the maths .

any ideas ?
  • #560
  • Jun 15, 2014 3:57am Jun 15, 2014 3:57am
Hello all,

The only difference between the Price Action indy (logic identical to Carl's indy) and Tomcat's indy is the rules.

If you look at the original on TSD they say the indy is based on an article in Currency Trader magazine.

I downloaded it and had a look.

The author based the rule set on the results of a huge search using the Kantu parameter-less trade generator.

This is quite a simple idea (at least the basic search is -- the genetic optimisation is a bit more complicated )

So I rolled my own in Matlab and am currently ploughing through a few zillion possible rules looking for the best ones for our purposes.

Thus far I have found a large number of rule sets that give very similar performance to either the Price Action indy or Tomcat's indy.

Bottom line is this -- If the program ever finishes and finds anything useful, it will only take 30 seconds to rip out the current rule sets from the indy's and insert the new rules.

Then we will have some new stats to play with.

Of course, the rules may be essentially arbitrary, in which case we are on a fool's errand and I am wasting a lot of good CPU cycles .

I'll report back in due course.

Best regards,

Old Dog

.
0 traders viewing now