So: techie question. I enjoyed the Hurst book that I was kindly linked to, and was especially impressed that he did all that figuring in the pre-computer era. It was about decomposing complicated oscillations into superimpositions of simple cyclic motions, a venerable mathematical technique pioneered by Claudius Ptolemy in his 1st century work on planetary orbits (he gets a bad rap in scientific history as the fool whose errors modern astronomers corrected; I like to give him a shout-out because he is actually one of my heroes), now mostly associated with the name of Fourier (French mathematician who rigorously proved under what circumstances it can be done, and how to do it).
Nowadays, there is considerable software available to do Fourier transforms on this kind of data-set, but I do not see anybody using indicators in that style. Does anyone know if there are programs out there that do this kind of thing? If not, it would seem to be a possible money-maker to write one.
Irrelevant story: this is NOT, contrary to some recently expressed opinions, an example of a contentious Internet board. My only experience as moderator of an Internet board was several years ago on the Middle East Abrahamic Forum, a board founded by a Muslim with one Jewish and one --------- assistant manager "as a place where all faiths can discuss regional issues with mutual respect". You can examine how well it went generally, but there were interims of calm interfaith lovey-dovey talk; I said I was a Buddhist (didn't want to say "atheist" as I had no interest in arguing religion) but was told I was welcome anyhow. Then Palestinians blew up an elementary school bus, the flame wars developed into a forest fire, "oh yeah, what about ISRAELI crimes?" with gory pictures etc. etc. The manager raised the assistant managers to "manager" rank (conferred the power to delete posters as well as posts) and quit, not responding to e-mails pleading for his return. The Jewish manager banned the majority of Muslim posters. The --------- manager raised one of the surviving Muslims to manager rank and he figured out that this board system even let managers ban other managers, so he banned the Jewish manager and all the Jews, to which the --------- responded by banning him, raising me to manager as the only neutral, and quitting. This all happened in a few hours while I was off-line, and I knew nothing about it until I woke up to over 300 e-mails. I unbanned everyone, including the one who had been on a tirade of personal attacks against me; naturally she started posting that I was the conspirator behind the whole debacle.
When I first read Hurst, I thought this is great - now how do you do it! 50% backward shifted MA's work great in hindsight but give you really no heads up going forward unless there is a trend in play to extrapolate lines. I agree with Paul on the trending %'s - I recon 20% ish trending say on daily/main but mostly back and forth, back and forth..... There is a company who have software that does 'it' and if you want further info, here is the link http://www.sentienttrader.com/cyclic-theory/. Not for me mind you but looks like they have it sorted in their own image.
However, all of my work since then has been on attempting to harness the oscillations across timeframes and trying to find the 'right combo' for me. If you follow my charts on EURUSD, I believe the weekly trend cycle is shown very clearly on a 2hr time frame and the intraday on a 10 min. The intraday cycle clearly 'works' the weekly one (wave within wave) and I have now moved on/added in, the daily and 2weekly as bigger versions of the intraday/weekly. I will continue to post my charts which are fully annotated if you wish to duplicate the waves or send me a PM as others have done if you would like extra info.
The key however is attempting to 'telegraph' forward the next move and this is where the journey through Elliot Wave and the like became important to me. Unfortunately, I have not found the EW work in itself, particularly accurate at all, bar the ending patterns and intra wave relationships. This is why I developed the MyWave concept. At points of maximum extension, whether impulse or corrective, a wave count of 13 is going to get you close to a turn of some note. As you have seen, I count this on both 10min (intraday ) and 2hr (intra week) to try and tell me where we are. These on their own are not accurate enough but I have found them easier and more accurate than EW, but I am biased.
The combination of the Stochastic Momentum Indicator and MACD alongside the MyWave counts, Hull and Attenuation MA's inside regression channels provides (I am biased!) a multi-dimensional approach to the market that at least exposes the cycle theory of Hurst's, if not exactly replicate his hand drawing. Key however, is the use of multiple time frames to capture the waves within waves.
However, I am particularly enamoured with Gann/Don Hall's work on time squaring significant pivots, which provides 'forward' information. Combine this with carving up the weekly waves with a Fibonacci time ruler and .....well, I think you get the point, it ain't easy. But, I think I am there. I have been posting/sharing cause I want to give the knowledge away - I never realised however, that this is the difficult bit...not learning, but trying to teach/share! I will continue to post everything I can including the 'what I am looking at and why' so that you and anyone else can at least look at what I am seeing, but I fear that is as much as I can do for now.
Last edited by Clivewaverider; 05-19-2012 at 12:39 PM.
Any longs would have been profitable. I did not participate... because I'm already very long!
As can see, never taken a short since the drop below 1.27. 7 trades. 2 days. 111.5 pips.
Will probably be able to clear my remaining longs next week. I secretly hope it will close those eye-hindering gaps as well.
It's been very enjoyable day. We beat the SSI today
Good luck!
It is always a relief to see the 'ping' back when all that you have worked on till now, says it will, but continues to disappoint. That relief is clearly visceral in the wave action. It will be interesting to see what happens next but hope for that zigzag: if this is the 'a' wave then a nice deep 'b' to release my covering shorts again (from 1.2740) and then up to 1.3055 ish would be my not so secret dream too. It is however, more than likely that this is the 4th wave in wave 3, not the 4th wave of the descent from up high, so I keep short covering to lock in those pips when possible - if yesterday's low holds, then that will likely see me out of longs too by end of week and with a reasonable profit, so keeping fingers crossed.
I'll post my charts on Monday but remain cautious of downside plays from oversold conditions, notwithstanding the herd stampeding.
When I first read Hurst, I thought this is great - now how do you do it!
Yep. In theory, theory and practice are the same thing, but in practice, they aren't.
Originally Posted by Clivewaverider
There is a company who have software that does 'it' and if you want further info, here is the link http://www.sentienttrader.com/cyclic-theory/. Not for me mind you but looks like they have it sorted in their own image.
Thanks. What it looks like is that Hurst came up with a whole bunch of ad-hoc tricks, that still require subjective human input to convert into workable "trading signals". Of course, he was working when computers were a lot slower and getting big blocks of computer time was expensive. Nowadays I think it would be more possible to make the computer find the variable-frequency-and-amplitude cycles from first principles with lots and lots of number crunching. I see the algorithm like this: first find a linear regression (best-fit line is better than a simple moving average; a quadratic term for a trend which is slowing down or speeding up is not that mechanically difficult either) for the logarithms of the prices (the "exponential" moving averages use logarithms-- the reason being that the reciprocals EUR/USD and USD/EUR really ought to generate the same best-fit) and subtract that trend-line out; the remainder is the oscillating part. Compute the Fourier coefficients assuming that one block of time repeats forever-- but ignore all but the highest-weight harmonic, as a best approximation to the frequency of the biggest wave. The key is to have a first-approximation to the Hurst function, a modified sine wave with both frequency and amplitude variable with time: regression equations cannot, from the raw data-set, immediately solve for all the coefficients in such a complicated function, but given a reasonable expression, partial derivative of the summed-squared-deviations by each of the coefficients gives a method of successively refining it toward best-fit. Then subtract off the trend-line plus first-Hurst-function, and repeat the process with the remainder to find the second-most-important wave, and so on until the point of diminishing returns. The whole rigamarole would only have to be done once (for each currency pair): as new data comes in, only re-adjust the partial derivatives by the effect of the new points, to obtain updated coefficients for the whole trend-plus-multiple-waves function. Display the future projection, and keep the old projections also visible as an indicator of the error margins.
I'd need a research grant, though, too broke to spend all my time on programming.
Yep. In theory, theory and practice are the same thing, but in practice, they aren't.
Thanks. What it looks like is that Hurst came up with a whole bunch of ad-hoc tricks, that still require subjective human input to convert into workable "trading signals". Of course, he was working when computers were a lot slower and getting big blocks of computer time was expensive. Nowadays I think it would be more possible to make the computer find the variable-frequency-and-amplitude cycles from first principles with lots and lots of number crunching. I see the algorithm like this: first find a linear regression (best-fit line is better than a simple moving average; a quadratic term for a trend which is slowing down or speeding up is not that mechanically difficult either) for the logarithms of the prices (the "exponential" moving averages use logarithms-- the reason being that the reciprocals EUR/USD and USD/EUR really ought to generate the same best-fit) and subtract that trend-line out; the remainder is the oscillating part. Compute the Fourier coefficients assuming that one block of time repeats forever-- but ignore all but the highest-weight harmonic, as a best approximation to the frequency of the biggest wave. The key is to have a first-approximation to the Hurst function, a modified sine wave with both frequency and amplitude variable with time: regression equations cannot, from the raw data-set, immediately solve for all the coefficients in such a complicated function, but given a reasonable expression, partial derivative of the summed-squared-deviations by each of the coefficients gives a method of successively refining it toward best-fit. Then subtract off the trend-line plus first-Hurst-function, and repeat the process with the remainder to find the second-most-important wave, and so on until the point of diminishing returns. The whole rigamarole would only have to be done once (for each currency pair): as new data comes in, only re-adjust the partial derivatives by the effect of the new points, to obtain updated coefficients for the whole trend-plus-multiple-waves function. Display the future projection, and keep the old projections also visible as an indicator of the error margins.
I'd need a research grant, though, too broke to spend all my time on programming.
A little bit over my head Robert but in theory yes. In practice, I seem to be creating the Hurst outputs using the previously mentioned indicators and I keep working to trim the 'flab' so to speak, so that I can automate it a bit more. Until then, I'll concentrate on making money as proof as ever is in the proof.
I have been full time since last November and part time since Jan 2008, so feeding the family is my priority. That is the beauty of these forums - you get access to all the hard work that everyone else is doing if they are willing to share. Whether it is worth sharing is another matter of course!
I'll start with the week ahead and what the possible Elliot Wave counts suggest:
Chart 1: 240min showing monthly pivots. Accepting the 'traditional' count that we are in the larger third wave down, having mapped out a 1-2, 1-2 before launching this baby from 1.3283, it would see then that we are in the 3rd of the 3rd.
Within that 3rd of a third, it appears that we maybe are in an extended 5th wave and making out wave 4. Given that wave 2 of the 3rd (see red marker) was flattish affair, I was anticipating a sharpish 4th and this is what we seem to have - but it looks to be the 4th wave within the 3rd, so be careful. My short-term (10min count) suggests that we may be nearing the end of an abc wave zigzag, so down then in the 5th of the 3rd before a proper retracement. I have shown 2 sets of retracement calcs on the chart; one for this little 4th (1.27705-1.28502) and then if the 5th down holds very close to the current low (1.2886-1.3038).
I can only guess, but as the 8hr MACD is turning up now, I would expect price to make a new low but the MACD positively diverge. This seems to fit the count, but with the 'ledge' below 1.2623, and the extended 5th wave already, I am not expecting the 5th of the 3rd to be too big. Alternatively, the small 4th wave here becomes more complex and this and the 5th wave all 'smudge' together. This certainly fits my 'MyWave' work where the 13 waves down are already in and the MA bands, Hull and pyrapoint analysis all suggest we are turning (see Friday's charts).
Chart 2: attempt at trying to count this move up. It seems that the market is doing its best to confuse the issue by representing this 4th within the 5th of the 3rd as something bigger. It is doing this by making the 'b' wave almost non-existent. A 5th wave down that then takes out the 1.2641 low from Friday would trigger stop losses, and then a sharp move back up again in the real 4th wave would eat the stops being put above the this current move. All in all, I expect some volatility then. All the counts could be wrong but it seems to fit the 'turn' at the 'ledge'.
So in summary, after a bit of tidy in Asian markets/early European, down then up. I have covered my longs from 1.2740 and hope to release again between 1.2641 and 1.2623, with a move up to nearer 1.3055. I'll then bank my longs there and mass up on the short side unless the story changes.
Chart 3: I am still not convinced at all about the Euro break up theme. I have marked out on this weekly chart, the personality style of the waves thus far. Wave A is made up of an a(impulsive) b(corrective) c (impulsive). Wave B in my opinion is still in play as we seem to have had a corrective up (a) and this whole thing down is such a twisted, overlapping wave that I just have to call this a corrective too. If the 1.2623 level holds, then a 'b' wave within a 'b' wave. If this is the case then we will have a 'c' wave up to complete the 'B' wave and then the all mama of a descent in the real 'C' wave. This is my preferred count and I am sticking with it, although the shorterm EW counts do seem to make sense - I just don't like the price action - it all seems an 'act' rather than the real thing. It copuld of course be that what I am tracing out was what was hoped for, but circumstances have altered the 'plan' somewhat and we really are going down now. This is why I keep covering my longs, cause I just don't know. Hey ho. back to MyWaves tomorrow.
Last edited by Clivewaverider; 05-20-2012 at 10:17 AM.
EURUSD is a very kind pair - it lets the bears take some money, and let the bulls take some money
Circumstances change all the time, participants change views accordingly, but EURUSD simply just does its routine job to make everyone happy!
This upmove has been expected for some time since it broke 1.27 down - I cannot remember when the last time I were courageous enough to challenge a daily support level that may become the starting point of a trend reversal. And if I'm correct in my analysis, a short squeeze is in the making....
My cloud has finally come into picture again! Support at 1.275, it may break lower but should not close lower for 8 hours. Target will be 1.296
It's a retracement .... until it's not! Good luck!
I have been posting/sharing cause I want to give the knowledge away - I never realised however, that this is the difficult bit...not learning, but trying to teach/share! I will continue to post everything I can including the 'what I am looking at and why' so that you and anyone else can at least look at what I am seeing, but I fear that is as much as I can do for now.
I realize I can count your waves objectively with my objective zigzag lines!
There are only 11 waves, to speak objectively, with a minimal swing of 65 pips to count as a wave ... do let me know if you need the algorithms to generate the zigzag lines for you, which I think will be helpful in your analysis work. Free of course! Good luck!
I realize I can count your waves objectively with my objective zigzag lines!
There are only 11 waves, to speak objectively, with a minimal swing of 65 pips to count as a wave ... do let me know if you need the algorithms to generate the zigzag lines for you, which I think will be helpful in your analysis work. Free of course! Good luck!
Thanks Paul - I got Passavento patterns that does the same and that counted 11 too - There is a small 1-2 at the top that is being missed or we have the next 5th of the 3rd down as per my earlier post. The early waves set the tone for the rest of the wave me thinks, so objectivety is a tad too much expectation. I have tried different bar lengths on my indicator and sometimes it is accurate and sometimes not - retro fitting, perhaps. Anything that helps though and if someone else is at least taking me seriously, then fab.
It no matter though; anything after 9 and you have to assume that things are getting 'squiggly'. All I am trying to do with the MyWaves is avoid consensus (it ain't never going to go up , it is different this time, etc etc, etc). Down then up for me, but we are on the same half the pitch so I hope you are good company!
I have tidied up my weekly wave chart now and added in the daily MA band to the 2hr - quite easy to see where the retracement will house itself until price bonds to the underside of the daily wave and makes its next move. For now though, my stoch moms on 2hrs are looking like repreive is due to the downside - well soonish - and the pink 2 day hull will be the key. Price will want to test this over and over like a crocodile roll until it flattens and turns up and the 2hr MA band has got to rollup too. They are still facing downhill, so no escape for price yet without a news announcement and even then, I'd expect (can not guarantee) some sort of sharp return to pick up the stragglers. We shall see though - I am just an observer!
Last edited by Clivewaverider; 05-20-2012 at 12:51 PM.
Clivewaverider I noticed your Elliott count had the last low as III.
Then your last Mywave count had a 1 on the last whole up leg. Is that to imply you anticipate a certain number of waves to fulfill your Mywave observations or a matter of allowing for any potential to present itself and thus mandatory to begin at 1??
You mentioned an expectation for a down move after your indicators allow for such, so my curiosity makes me wonder where you anticipate your count starting with 1 to top out at (3,5,7,9)?
Thanx for your consideration
_______________________________ Just cleaning up all the leftover pips
Clivewaverider I noticed your Elliott count had the last low as III.
Then your last Mywave count had a 1 on the last whole up leg. Is that to imply you anticipate a certain number of waves to fulfill your Mywave observations or a matter of allowing for any potential to present itself and thus mandatory to begin at 1??
You mentioned an expectation for a down move after your indicators allow for such, so my curiosity makes me wonder where you anticipate your count starting with 1 to top out at (3,5,7,9)?
Thanx for your consideration
Good to see you Captester - you been quiet lately? I run the MyWaves alongside the EW counts in an attempt to keep my options open. In EW world, we are stuck with not really knowing whether this is wave 4 of 3, wave 4 or something different, like my start of C wave up and the whole count is not aggressive to downside as per standard thinking. MyWave however is a tad more objective - the longest impulsive wave (double extended or Hopkiss's long wave) or corrective (triple) consist of 13 waves as per banlan, Hopkiss and EWP books. Rather than worry about form intra wave, I am just looking for timing the end to what 'was'. When it comes to triangles at the end of a corrective triple, there will be some grouping of waves as you can start to get higher wave counts than 13, but not higher highs and lows.
As to what it will become, well, I'll just start counting again. As you saw, I have started with 1 because 1 follows 13. The 13 in the current intermediate count suggests a turn was imminent on a weekly basis, although nailing that 13 was not straightforward as the overall wave down lacked a certain definition at times (Paul - I'll think again about using the zigzag/Passavento...). Once we get to 7, (potential abcBabc) then we can look at the style/personality of what made that up and how the oscillators look and try and make a decision as to what is going on - so does the corresponding EW counts make sense. I have found counting the waves in the short term chart however is the one to trade with, rather than the intermediate - not all counts will make 13, but the higher ones from 9 upwards will make up the impulsive moves generally on the intermediate chart - fractals just like EW.
Just a tool Captester along with the rest - I think it adds something, if only a slight edge.
Last edited by Clivewaverider; 05-21-2012 at 12:53 AM.
So here is the short term MyWave count as I see it this morning - was 11 enough or is there a 13 up? have I got the counts right? ...don't know till we see the reaction to wave 12. If we do not make a higher high then 11 was enough but it certainly looks like an Ending diagonal at the right place.
I think for me the key here is that I would not be thinking of going long at this point (even if there is a possible 13 up) and would be waiting for a retracement. Obviously this fits into the longer term work I have done in previous posts, but as a heads up, the short term chart is a great start for me.
I have closed my longs at 1.2795 and remain naked short through the retracement (or otherwise ) back downhill.
Last edited by Clivewaverider; 05-21-2012 at 01:20 AM.
So here is the short term MyWave count as I see it this morning - was 11 enough or is there a 13 up? have I got the counts right? ...don't know till we see the reaction to wave 12. If we do not make a higher high then 11 was enough but it certainly looks like an Ending diagonal at the right place.
I think for me the key here is that I would not be thinking of going long at this point (even if there is a possible 13 up) and would be waiting for a retracement. Obviously this fits into the longer term work I have done in previous posts, but as a heads up, the short term chart is a great start for me.
I have closed my longs at 1.2795 and remain naked short through the retracement (or otherwise ) back downhill.
Matters seem to be going according to plan this morning and closing those longs on the previous count looked timely. I am only counting MyWave 5 so far on this retracement (or not!), so holding my shorts for now and hoping for 1.2670-1.2680 (S1 + Hull + channel exit zone etc etc). If I can count 11-13 by then, I'll be looking to close shorts and reapply longs as per game plan.
I have left fibonacci time slices on the chart in case anyone wants to look at them.
Last edited by Clivewaverider; 05-21-2012 at 11:05 AM.
I usually don't put much value on EW, but use it mainly for wave characteristics. The down move looks every inch like a wave c of B to me. So this is the scenario I favour. I know Clive would be very pleased with this potential...
I usually don't put much value on EW, but use it mainly for wave characteristics. The down move looks every inch like a wave c of B to me. So this is the scenario I favour. I know Clive would be very pleased with this potential...
EURUSD is a very kind pair - it lets the bears take some money, and let the bulls take some money
Circumstances change all the time, participants change views accordingly, but EURUSD simply just does its routine job to make everyone happy!
This upmove has been expected for some time since it broke 1.27 down - I cannot remember when the last time I were courageous enough to challenge a daily support level that may become the starting point of a trend reversal. And if I'm correct in my analysis, a short squeeze is in the making....
My cloud has finally come into picture again! Support at 1.275, it may break lower but should not close lower for 8 hours. Target will be 1.296
It's a retracement .... until it's not! Good luck!
I don't know why I'm posting again when everything is just moving as planned .... maybe just to pump up the number of posts for a new baby thread!
The rush for time not to close more than 8 hours below 1.275 is quite obvious here Good luck!
Disclaimer: Trading foreign exchange on margin carries a high level of risk, and may not be suitable for all investors. The high degree of leverage can work against you as well as for you. Before deciding to trade foreign exchange you should carefully consider your investment objectives, level of experience, and risk appetite. The possibility exists that you could sustain a loss of some or all of your initial investment and therefore you should not invest money that you cannot afford to lose. You should be aware of all the risks associated with foreign exchange trading, and seek advice from an independent financial advisor if you have any doubts. Any opinions, news, research, analyses, prices, or other information contained on this website is provided as general market commentary and does not constitute investment advice. Forex Capital Markets LLC. will not accept liability for any loss or damage, including without limitation to, any loss of profit, which may arise directly or indirectly from use of or reliance on such information.
English
Español
Français
中文
Deutsch
Italiano
العربية
Sverige
Ελληνικά
Originally Posted by Robert Eckert
Reply With Quote
