Smallest practically traded cycle


#1

One of the first points that Hurst makes in Profit Magic is that effort spent in improving timing accuracy is not wasted because the more often one can make a successful trade, the faster profits will compound.

I’ve noted, anecdotally, that many, if not most of the forecasts made on this forum, as well as with Hurst signals, and other Hurst practitioners are made on ‘longer’ cycles. I believe the magic numbers are 20 days and 20 weeks. In other words, particular attention is paid to the 20 week cycle troughs and peaks that occur within larger cycles, and resistance/support is sought with the 20 day FLD or higher.

My own experience has been mixed attempting to use a 5 day cycle. However this could be attributable to many things including inexperience.

Does anyone prefer to trade a cycle smaller than 20 days?


#2

I trade intraday frequently using David Hicksons intraday model. It requires a very great deal of experience and is definitely not as straightforward as the longer waves.

I generally trade the DAX or one of the more liquid forex pairs with a tiny spread.

Stick to the longer timeframes if you are a beginner.

Dave


#3

I trade intraday all the time. All of the rules are the same, though it takes a lot of discipline to follow them because things happen very quickly. If your time frame for “retirement” is long (i.e. 3-4 years), then I would recommend sticking to the longer waves. Your margin for error is much greater than it is on the very short waves. If you use fixed-ratio compounding you will have all the financial security you need plus a life in the meantime!


#4

Retirement? Life? I chose something else. :grin:


#5

Is that intraday model part of his course? Where can I learn more about it?

Sure, Hurst stated in PM that the lower the cycle the higher the yield but it gets extremely tricky intraday - of course Hurst had no access to this in his time. I would say there is somewhat more of a random walk intraday…but not totally random!

I’ve found that an FLD tuned to a 1 or 2 hour cycle at a relatively un-volatile time (1 hour after Sydney open before London open for example) is fairly effective as long as the underlying trend doesn’t take off, but this is where having a good idea of the bigger picture is vital.

So when you trade intraday, are you trying to stay onside of a higher cycle trend, or it doesn’t matter?


#6

For whatever it’s worth, John Ehlers does not recommend trading cycles shorter than 10 bars due to aliasing noise.

This brings up a question I’ve been wondering about recently. When trading stocks intraday do you recommend looking at cash markets or futures? I know cycles supposedly run 24/7 but they usually make more sense to me intraday when looking at cash or futures markets during active trading hours (ignoring the overnite futures session which is often just random noise).

Curt


#7

I wait until the CASH market has traced out at least 2 or 3 hourly nominal cycles before making any trades. I can then get a general idea of where the 5 hour and 160 min nominal cycles are with FLDs, ruler etc…

Sometimes it can be beautiful.

D


#8

Yes, I am very careful extrapolating cycles from data that is shown overnight in the futures session. Wise to wait several of the smaller cycles to get a good idea of the larger cycles phase.


#9

Even though I track the Globex session with my filters, one of my “rules of engagement” is that I only make trades during regular (U.S.) trading hours. My action signal waves are short enough that there will always be at least one opportunity during regular trading hours.

Ehlers comment is similar to one of Hurst’s recommendations in PM.


#10

Ruler? Nice! JM would be proud.


#11

The ultimate tool in any Hurstonian’s toolkit!


#12

Spectral dilation is lower frequencies (trend) showing up in a shorter term cycle analysis.


#13

Speaking of which, (lower frequencies showing up in shorter term cycles) and I’m sorry to bury it in this discussion but I can’t find the other thread where it first came up.

Someone (Curt or William I think) mentioned that Hurst only attributed 23% of market movements to cyclic activity, what he calls ‘X motivation’. The rest, I assumed, he attributed to the 'sum of all underlying cycles" or the ‘underlying trend’. However in the first chapters of Profit Magic he plainly states,

Foreseeable fundamental events influencing investor thinking regarding industry groups and specific issues account for 75% of the price motion of stocks. […] Unlike historical events, investor motivation stemming from fundamental factors affecting group and individual issues does influence price-motion to a large extent. The effect is of the nature of an underlying trend which is relatively smooth and slow to change direction - if the factors involved are foreseeable.

He goes on to mention the underlying trend isn’t really useful ‘for our purposes’, and I guess that’s why we never talk of fundamentals, much, in this group?

Did Hurst change his mind later, and decide the ‘underlying trend’ isn’t fundamental in nature but the result of unmeasurable higher degree cycles or am I just (typically) confused?


#14

I don’t think you are confused and Hurst did not change his mind. You bring up what I feel is one of the most important aspects of cyclical analysis with respect to financial instruments, that being the vast majority of the price movement is non-cyclical in nature. Hurst never explained to my knowledge how he came up with that differentiation. If he had, then cyclical analysis would be a snap, just subtract the non-cyclical portion from the price action and run your analysis on the high pass data. However, I accept Hurst’s statement as one of the underlying premises of my approach to the market.

What causes some confusion in my opinion is the fact, being time series data, the price action can always be decomposed into spectral components using one’s favorite algorithm, regardless of the length of the data. Hurst stated in PM that fact, in and of itself, tells you nothing.

Of course that leaves us with the question, what is the longest price wave? Fortunately for trading purposes I don’t have to answer that question because I am using relatively short waves. The sum of the longer waves plus the non-cyclical underlying trend is included in the high pass band of my band pass filter. Hurst makes the same statement with respect to the center line of his “dominancy envelope.”


#15

Beautiful


#16

That looks like textbook FLD action so far.


#17

Yes, that’s why I posted it. Not the dominant cycle so just a small piece of the bigger picture that is playing out.


#18

Very nice Curt, join me on twitter when you can and post these gems…


#19

Something I’ve whinged about before is not being sure what factors are the most important to consider when making trading decisions based on Hurst’s principles.

I mentioned that I thought some cycles are more ‘dominant’ than others, and that there is usually no point in trading against the dominant trend ‘caused’ by these cycles. This post is a reminder to follow up on an idea I’ve had, namely a chart that documents the average expected counter-trend move in certain selected instruments that would allow a trader to gauge the risk/reward of trading cycles of any length.

For example, as I’m quite interested in current CAD weakness and the expected future turning point I’ve examined the cycles in EURCAD and found that major reversals generally happen only at the < 18M+ nominal cycle peaks and troughs. Trading the 40 week nominal cycles is worthwhile however, as they can produce counter-moves averaging 80-100 days and with good amplitude. The 20 week cycles require more agility but still are worthwhile, while the 80 day cycles would seem too choppy on this trading horizon. However a trader who is focused on 20 week reversals would find the 80 day countermoves entirely acceptable.

Knowing the historical averages for each timeframe and cross-referencing would allow a trader to decide what is the sweet spot/limit for their selected horizon as well as helping identify the largest countermove that is likely for any given period. Please feel free to nag if I haven’t updated this with some preliminary results.