So, people,
the thread is enriching with still new members! Great. I liked very much also the last contribute by phillip.
Now: the point is serious and I just try to give my piece of cheese, given that I’m really curious about the deepening theoretical information.
I tried to apply a simple normalization function to the indicator. I just divide the MACDC by its range, and set the margins [I]on average[/I] to -5.0 // 5.0. This automatically and without the use of an “arbitrary correction constant” like Clevel was.
“On average” because when the indicator is ranging for long periods on one side (bullish or bearish, i.e >0 or <0, respectively), it can happen that the indicator falls below -5 or above 5, but these are really “outlier periods”, as sometimes Pipy cited.
The inconvenient with this math. transform. is that the graphs, for currency pairs giving normally very high absolute values of the MACDC, is that the graph is quite compressed. At a first glance the 0.5 entry point is quite good, anyway. In some periods you never touch the 3% (TP level, if I understand correctly). But, always if I well understand, the TP is not such a problem, because it can be also fixed based on the ATR (which will be correspondingly high in high volatility times).
Pipy: are we completely screwing up the system?
Now, with this normalized indicator, I can start thinking about an EA.
(Sometimes brk or others will explain me how to insert pictures better than that!)
mmm, good work Fabio. Im just not sure the solution to the problem is by transforming it within in its own model ie changing the histogram within itself…
Id much rather like some kind of solution/explanation that unifies the fact that all these pairs have different D/P ranges with the whole 0.5 to 3 thing
Although Id love to understand this from first principles Im coming to the conclusion that perhaps the best way would be for Pipy just to show us his art and post his trades as they come. He’s been awfully patient and given us a lot of explanations…the best way now is to see the system at work
With the successful trades we’ll be able to work backwards and hopefully help Pipy to refine the explanations for people like myself that dont have the experience that he does
Let’s see, chaudmo.
First, Pipy has already posted quite a lot of trades here.
Second, the nice thing of a normalized indicator is that it is amenable to automation. If I manage to write the EA (it’s taking some time, and I’m employing all the time spared from job & family), it’ll be possible running some back-test-based optimization for instance on the entry level (0.5, or else?).
All the game here is to understand how, via this indicator, we can catch the point where prices are coming into the main gaussian region from the outliers region, if I got the message from Pipy. And I am quite confident that this can be achieved with some math… Of course a formal demonstration of the relationship between price distribution and the MACDC would be useful… But likely, this is a sort of empirical rather than formal law…
Another issue that can be discussed is whether the normalizing transformation applied is adequate, or if it’s better using another parameter, like the average of the MACDC over x periods, or the range of the last two (three…) swings.
Pipy’s eye is for sure the best entry point. My problem is twofold:
I have no such good eye
I cannot keep my eye there for long times.
I have the impression that Pipy is getting in when he notices that the MACDC level is reaching a certain threshold that is significant based on the current swinging range of the indicator…
I’m just trying boiling it down in some mathematical term.
Personally I dont think a success of a system is in its mathematical formalization. Iam a (finance) mathematician myself but I have learned the best systems rely on judgement in th end. Circumstances always change…then what do u have…only formulas.
I myself am quite happy if Pippy takes us along with him and his experience. Some mathematical formalization is necessary but lets see how the system works. The last thing I will do is start making up assumptions and making my own indicator without Pippy. 4 or 5 trades is not enough…
come on Fabio man, pull yourself together!!! (is a joke!! my little kid keeps on saying to me, he pulls my leg until it clicks and it really hurts, oh my…and then he say…pull yourself together daddy!!)
…GAUSSIAN regions??? LOL lol come on fabio, u dont need to look clever here…all these people are clever on this forum, in experience or knowledge…do u think many people on this forum know this gaussian regions!? I myself studied it for 3 years…I provide quant modelling for our traders and i could start talking about all this here…but these forex traders here could blow my balls off trading…so lets concentrate on trading, thats what we’re here for. i try to be humble in this company. SOME mathematical codification is necessary of course but lets get this conversation away from degree level theoretical maths!! the least complicated indicators are the best
Pipy come back and get me away from this talk of gaussian regions man, i dont want to turn into a robot!! I learn from ur expert eye and whatever else u have. Lets start making money with the basic knowledge we have rather than cannibalize your system…the success of the system will provide us with more theory…
For my part, I ask Pippy, and i try to keep basic without talking rocket science
do u say to us this…that
u tried 0.5 to 3 on your charts to backtest…but that work with equities not with forex. It worked only when the scale was 3 to -3…Then you tried scales with 10 to -10
For that range you found that only 3/10 of the range was the optimum range for best profit …
Then you tried scales of 20 to -20. Well the best range to make profits was then 3/20 or 0.15% of the range. In this way, we could formalize the scales as y= 3/x?? Is always 3/x but as the numbers get bigger its harder to “see” the bars at the start of the trend?
Maybe this is all wrong, but in any case I wait for pippy and take my lead from him
phil
oh fabio come back, you are obviously very intelligent , i didnt see i was being so mean i apologise. I just dont agree with you thats all Iam sure im wrong in my views but no one has to agree with them at all
I just wait for pippy rather than conjecture thats all. We should all work together for sure, everyone is very informed here. I think Iam impatient thats all. was not meaning to start a match for slanging by any means!
hey guys ill be able to get to this by next week i was driving and smoked a deer at 80mph totally destroyed my car luckily i got another one at my parents house in florida cause im a pack rat and dont do trade ins :)… so ill be back next week cya
k guys sorry for being away alot of stuff going on… ok i guess ill start here…
i guess its ea’able but im not good enough programmer for all of that… i barely can write indicators
So I think I will try to automate a bit this system, as far as possible, given that is really a sort of “complete science” (complex, therefore) of trading. I am pretty sure not to incorporate all the features (the “advanced” pipy’s lessons) and not to be able to reproduce the kind of flexibility that is granting an experienced guy like pipy the optimal profits out of the system.
you may want to try to manually trade this way i know it wouldnt be automated but i feel a machine only hinders a trader… many people sing their benefits but there really are to many things to understand at once trading nothing is really black and white. well with the advanced stuff but i guess the easy stuff would work i dunno tho.
I hope this way also to clarify some aspects of the method that, after re-reading the thread are not yet very clear to me for two reasons:
[ol]
[li]I’m not a “Yankee” :D, and therefore I had some hard time trying to interpret some not very “plain English jargon”.
[/li]> [li]There are still some theoretically unclear issues, like the longly debated (last time by chaudmo) scaling problem. This demonstrates that pipy is really first of all an “artist”, and that for him reading graphs and indicators is much easier than for most of us.
[/li]>
[/ol]
dont worry im not a yankee either im from NC
ill show scaling again in a few posts. with a pic and explanation.
There are several things to be better formalized in math or algorithm, trying to learn from arts. But perhaps this will also help a general clarification of some aspects.
What do you think about the feasibility?
For instance: the first thing that can be discussed about the scaling problem is that, adjusting Clevel for getting a -5 / 5 scale is really quite arbitrary, because it does not depend just on currency pair and timeframe chosen, but also period of time and lenght of the time (lenght of the X-axis) considered.
im not much of a programmer so i cant really program it to be self manipulating and whatnot i jujust programmed it so i understand it the 5/5 thing isnt arbitrary its there for a reason to scale the graph tho its not always 5/5 … thats why modifying clevel works best. higher the clevel the tighter your trade basically. the tighter the trade the more accuracy but less movement. the lower the clevel is wider the movement … less accurate.
because the .5 to 3 always stays the same…
explaining the .5 to 3 - random numbers that i like to use you can use 1 to 5 if you want as long as the data is scaled for it…
but these numbers are ALWAYS the same then you can change it on the fly cause then ur just trying to curve fit to the chart and that doesnt work.
Examples: in my hands, on a H4 chart of EURUSD, with indicator’s (MACDC) parameters (36,78,27), a Clevel of 5.0 is needed to keep the -5.0 < MACDC < 5.0 if I consider the period Oct 23 / Nov 21 (today). If I then change the x-axis range to include just the last week (anyway more than 30 bars), the same Clevel parameter gives as a result a Y-scale ranging from -0.25 to 1.04, and therefore should be changed. By keeping Clevel = 5.0 for EURUSD/H4, MACDC would have never reached levels < -3.0 or > 3.0 at least from April to middle August 2008.
modify the clevel will fix this… change it so all the big movements contain these points.
Of course a skilled “live trader” like pipy or you other guys can easily “adjust” to the varying market conditions, and even skip to consider the values of the indicator (actually you “follow the shape”, also visually). An EA (and a hobby trader relying on it because cannot follow the market real-time) cannot…
yup easily adjustable just make sure not to try to fit it to the curve.
Is it possible in your view fixing a standard? Or can we find a math formalism able to substitute this -5/5 interval (and the related 0.5% - 3% trading range) with something independent from scales?
Just an input. Sorry if I will not show up very regularly, due to many other duties…
FS
the 5 / 5 variable is based on your risk basically and personal settings not really a standard. you can do it tho if youd like itd just be your standard.
this whole window upper and lower thing is personally changable its not rigid.
you modify clevel to your risk basically the higher the window the shorter more accurate but less profitable your trades are this is risk premium basically
sweet masters i only got a B.S im trying to go back tho when i have free time
First of all Pipy great work…I thought at first you are a finance professor! Ive worked a lot on MACD and my final dissertation was on Black scholes pricing with MACD and its clear you have a large body of knowledge.
i did alot with that as well as this isnt the only market ive fudged around with.
Like many academics, you have the ideas clear in your mind but, if you permit me in saying, perhaps would help everyone if you explained them with more clarity. (I say this as a someone who had to change his thesis n times due to bad explanations - apparently good maths was not enough :=((( ) For instance your discussion on probabilities I liked a lot, but seriously chris I have to cut and paste the words into sentences myself to even figure the thread of the discussion!
yeah i know i suck i need to sit down and write on paper and erase and rewrite and erase and rewrite so its clear haha when i was in school i would get really hammered and write all my papers drunk then wake up in the morning and fix them but id have the majority of the ideas down that may be why i do it just habit.
You’re doing a awesome job in giving us all this system, im sure i speak for all when I say that there is much potential in this system and I hope you persevere because people like you giving this knowledge can change even peoples lives so thank you so much. (this sounds strong, but seriously thanks!)
I have read this thread maybe 5 times and the scaling issue I still cannot fully grasp…
The price action c ratio is a good development…i think this helps to reduce the white noise of the market in shorter timeframes. Maybe the confusion is due to the % issue
yup thats what its basic job is to eliminate that noise in anytime frame… yeah its not a % i messed up i think early when i listed it like that its % in ES with a different formula in fX its different its just solid numbers.
Obviously to 0.5 to 3 is an empirical range, but for equities…But, all the great research you have done shows that the 0.5 to 3 is not applicable to forex. This is where people (ME!) are confused I think. You keep on saying 0.5 to 3 but actually you are scaling this. You cant multiply by 100, say that its X%…then multiply by something else and say its still a percentage unless the c vlaue is different for different currency pairs
.5% to 3% in equities is entrance at different levels with 2% being the best and 3% not giving any signals. takes out white noise in equity markets. you exit on price target or cross in equities. not hitting 3%… this wont work in fx i mean it might but then you might as well do regular macd swingline trading.
FX market doesnt move the same your seeing small incrament movements that dont really show up as a full % so it has to be scaled if your looking at numbers i dont really look at the numbers anymore tho just direction and price action…
but your basically raising clevel to make your window wider your max rises higher and your min goes lower or more neg but your points… .5 to 3 stay the same and its not a % this causes your trade to tighten and make the macd move safer … its just an system that has entry and exit built into the system rather than price action. what i mean by that is you dont go ok im in at .5 and ill exit at a predefined number … you watch the chart and when your macd hits 3 you exit boom no matter what the price action says.
You keep saying 0.5 to 3 but I dont understand. I myself took 4 years of data for Cable and EUR/USD to backtest (because i like the idea of the system! took me a whole day, i couldnt remember any of my EMA) and i cannot find any timeframe (obviously there might be…in 5 min data, 10 min data…but this is just noise for me) where there is a C range of 0.5 to 3. Even now i can look at my cable chart and its like 0.7 to -0.4 or something. (BUT OF COURSE, I WOULDNT FIND ANY VALUES, U SAY YOURSELF THAT THIS IS FOR EQUITIES. But then you maintain the 0.5 to 3%? there is some inconsistency. (clearly not for you given how succesful it is for you!!)) But the system says I just multiply these by 10? but of course the data fit then 7 to 4.
you have a coefficient for each pair in each window this coefficient is the clevel it changes this for you the histo will not change in anyway except the numbers… look at the big histo movements then look at the small ones you can tell which is trading range and when is good movement… make the good movement fit and by doing so youll eliminate the noise. because often the noise may hit .5 in trading range but you wont get price action crossing EMA of 20 period so its a filter…
So say I times USD/JPy by 3.1415927 just to get 5 to -5 . (Or maybe i say in my head i want to see 3 to -3, in which case i dont even need numbers or to change the Clevel parameter, i just imagine i see 3 to -3, i trade on the cross and then exit when the histogram reach the top…you see what i mean? i dont see this myself but maybe someone does…
this is what i do i dont look at the numbers i stack three histos different periods and wait for two larger periods to be and same direction and then the short period to move into the same direction and i enter and exit right on the crosses… you dont really need the numbers once you can just look at it and go ok thats what its doing.
Now you are well within you’re rights to say "hey guys, f*ck off! (sorry!) , its my system and now im sick of explaining!! but you are clearly one of the good guys and im sure you would like us all to understand. So maybe the solution is to show us how the c value relates to SPECIFIC currency pairs, GBP/USD, EUR/USD etc etc. to be honest I dont know the answer, only that I feel it is the histogram itself and not any numbers . ( That is why sometimes you say 0.5 to 3, sometimes you have said “trade on the cross” which is not the same exacvtly)
The fact is that for you the numbers dont mean much anymore throug your experience. But for us yes.
exit on cross is a safe exit if histo fails to hit 3 and your still in … its a good idea to exit on cross even if your sl isnt triggered. you still may leave with a profit.
I would like to help in any way I can, ur obviously a good honest guy and if you could help us to understand this system I could be eternally grateful
Many thanks Chris
Phil
hope that helps if you got more quesiton ask away.
im gonna be honest i dont understand your code cause im not much of a programmer… basically you normalized it… if you normalize it i tried this actually through much much much more code so its not as neat as yours … you can adjust it so the top fits.
the reason for tp is i always keep em up for spikes… so i exit into spikes if they just so happen to hit most of the time i would just exit on 3 tho.
Thanks, Pippy. I’ve been busy and working on some Point and Figure stuff as well, but I really like this system and have been waiting for exactly this to close any and all gaps.
ok first trade ill post remember money management with this all these trades will be live no trading in tha past here.
USD/CAD Short @ 1.2353
Reason - we notice from the chart long Histo is down intermediate is down and the blue had a bullish divergence that wasnt much at all and is not going to reverse down entrance would be at blue histo reading .5 and because the other long and short histos are both negative the blue should find good power in this negative move attempt.
for those interested in when you would enter and exit based on numbers i have this one set up from 10 to -20 meaning the value area here will tend to negatively develop also reinforcing our decision this is on a 4hour chart
also notice that the bar before decision to enter shows bearish engulfing formation giving added weight to movement
questions ask away
POF - probably 1.242732
Tp- probably around 1.21906
well it didnt setup like i hoped macd didnt cross down yet and we have def not hit -.5 yet
so here we are still waiting 99% waiting and 1% action
new chart attached still be a good setup if blue histo crosses down all those numbers will have to change for new entrance
also if red pulls up and crosses odds are good long position will open up on cross would be a HEALTHY increase in price/time without hard moves keeping macd-sig price ratio low but growing to 3 so well see
Hello…
Is anyone still using this system? Seems tyo have excellent potential… Chris banking over 2,000 pips in certain weeks. Chris or anyone else using this system are you out there? If so…lets start back this thread or form a new one. There is still great interest in this method…
Lets be honest, any system that gives a solid 500+ pips a week is worth supporting and improving if possible.
Please any of the former traders of this system respond back and lets see if we can continue to trade and develop this system…
Chris , it would be great to have you back if you are still around, please read this and reply back…
Thousands of views and no replies what is this? A system that can deliver hundreds of pips a week and no one responds? 98% of ths systems out there cant deliver this per week… I know I have seen hundreds of them…what are we all here for then? Lets just work our 9-5 jobs until we die… Sorry for the rant but all this potential and no one comes forward with any interest because it is more difficult than a simple MA cross over system…
Come on… lets get some talk started over this… why no one from the original thread has not responded is beyond me… Somebody has got to be using this system???
Hallo Theo,
just briefly, because I was also coming here sometimes ago, and you ask for people like me.
The only thing I would suggest you is not to be too attracted by +500 pips/week. I am now a sort of member of another “community” here, where people were showing >50% gains/month (on capital!) for several months in a row, and then lost their shirts in the last few.
2008 was a very “strange” year financially. I’m not saying at all that pipy’s system is not working, but perhaps some of the stellar results of the past months are not so indicative of the “average” or “expected” performance of a system.
Personally I must say this is not “my cup of tea”, probably because I have some big difficulty in fully understanding it. But then, most importantly, because I find it particularly difficult “formalizing” it in an ordered and precise way (made some efforts, and were not accepted much warmly, must say). I love arts, but for trading I prefer math and order.
