Key to this is the idea of special numbers. Not lucky ones as 7, or unlucky 13, but ones which, from number crunching, are considered significant. We are not going into the realms of mysticism here, as those of you who are aficionados of Fibonacci retracements and projections will understand. Proportions of 0.61875 and 0.38125 are also special, but not magical (although some do believe 1.618 is divine).
The Timespan Principle has three simple numbers: 9, 17 and 26 (remember, 9 and 26 are the days used for the moving averages) and are considered the most useful.
• 9 is known as the ‘basic unit’
• 17 (9+9-1) is two of these basis units (the –1 is explained later)
• 26 is three basic units (9+9+9-1), or one ‘term unit’.
The next set of numbers are compound ones, arrived at through rough combinations of the first three. These are:
• 33 (one term unit plus one basic unit), 26+9-1
• 42 (one term unit plus seventeen), 26+17-1
• 65 (is known as one ‘super big unit’, 33+33-1
• 76, (known as one ‘cycle’ as well as being three ‘term units’, 26+26+26-2
• 129, 65+65-1
• 172, 65+42+42+26-3
• 257,129+129-1
You can see that the word ‘roughly’ is very appropriate.
At first these compound numbers may seen a bit confusing, but the following may help. Let’s say we start with a series of nine ascending candles which we have labelled 1 to 9. Then one observes a series of descending candles labelled A to I. You will note that candle 9 and A are one and the same. The whole of this inverted V Wave pattern is completed at the seventeenth candle.
Now let’s look at an N Wave which starts with nine ascending candles forming the first leg. The down move also consists of nine candles, as does the last leg higher. These have again been numbered for rallies and labelled with letters on the way down. As before, the first candle 9 andAare one and the same, and similarly I and the second 1 are also the same Hammer candle.
The whole formation takes 25 days to complete and in order not to double count the candles at the turning points the formula is: (9+9+9-2)=25.
Ichimoku day counts are plotted from intermediate highs and lows and we watch for where they cluster in the future. The more counts that end at, or close to, the same day in the future, the more likely that that day will see a trend end and reverse. Again, all these numbers were discovered through trial and error - in what must have been a nightmare job.
From the completed N wave the next move might be an I wave lower which should take about 26 days (one of the preferred day counts). This is so that the I formation most closely mirrors the size of the N wave which took 26 days to develop.
When working out day counts these do not necessarily have to succeed each other perfectly. Hosoda-san allowed for gaps in the count: say, one to five days, around important highs or lows. As already mentioned markets are more likely to reverse/accelerate around the time when the Clouds become very thin, so too the day counts start from around the time of significant highs and lows. Of course, this makes it a lot easier to fit the 9, 17, 33 day counts to what has actually happened. It also makes it far harder to decide exactly on a date in the future when the move will end.
Intuitively, the 9 day and 26 day numbers make sense, nine being just under two weeks and 26 what had been the average working month. These are similar to ten and twenty-day moving averages used by manyWesterners.
Western pattern analysis also gives hints at dates ahead when things are about to change. For
example, a head-and-shoulders top: the perfect (idealised) example of which will be symmetrical and have a horizontal neckline. If the left shoulder took three weeks to form, and the top of the head was two weeks after the first neckline point, then the rest of the formation will probably take five weeks to complete - two weeks to the second neckline point and three weeks to form the right shoulder.Again, in a perfect world it will be the same height as the first one. In other words, as in Ichimoku charts, the timing aspect is also inherent in some conventional pattern analysis.
Plotting the number of days ahead gives us some idea as to where intermediate as well as major market turns might occur. It is a rather laborious process though.
Also remember that over time markets and instruments change and some methods of analysis will work better under certain conditions. Do not be afraid to discard methods when you feel the market has changed. For example,
• For very explosive markets which are trending very strongly, I immediately discard the Relative Strength Index.
• When markets are in very strong bull or bearmoves, I tend to discard all wave counts, and therefore most price targets. Nevertheless, I keep an eye on time targets.
Likewise in sideways moves I will discard moving averages and the Ichimoku Cloud.
Explosive move higher and very dynamic despite a thin cloud.
The chart of sugar above is a good example of a massive move where rigid analysis ought to be discarded, be it Ichimoku Kinko, RSI or pattern analysis.
That’s the end of looking at the three principles of Cloud analysis.We’ll next turn to applying all this in practice and work through some real-life case studies.
The Timespan Principle has three simple numbers: 9, 17 and 26 (remember, 9 and 26 are the days used for the moving averages) and are considered the most useful.
• 9 is known as the ‘basic unit’
• 17 (9+9-1) is two of these basis units (the –1 is explained later)
• 26 is three basic units (9+9+9-1), or one ‘term unit’.
• 33 (one term unit plus one basic unit), 26+9-1
• 42 (one term unit plus seventeen), 26+17-1
• 65 (is known as one ‘super big unit’, 33+33-1
• 76, (known as one ‘cycle’ as well as being three ‘term units’, 26+26+26-2
• 129, 65+65-1
• 172, 65+42+42+26-3
• 257,129+129-1
You can see that the word ‘roughly’ is very appropriate.
Avoiding double-counting
One digit is deducted because: “it has one attachment part”. This is in order to avoid double counting the candle that lies on the last day of the first section and is the first day of the second part of the sum. So for compound number 172=(65+42+42+26-3). we deduct three days’worth of candles, one for each day between the four counting periods.At first these compound numbers may seen a bit confusing, but the following may help. Let’s say we start with a series of nine ascending candles which we have labelled 1 to 9. Then one observes a series of descending candles labelled A to I. You will note that candle 9 and A are one and the same. The whole of this inverted V Wave pattern is completed at the seventeenth candle.
Now let’s look at an N Wave which starts with nine ascending candles forming the first leg. The down move also consists of nine candles, as does the last leg higher. These have again been numbered for rallies and labelled with letters on the way down. As before, the first candle 9 andAare one and the same, and similarly I and the second 1 are also the same Hammer candle.
The whole formation takes 25 days to complete and in order not to double count the candles at the turning points the formula is: (9+9+9-2)=25.
Projecting ahead
Just as economic cycles are said to range from 3.5 years all the way through to 54 years, highs and lows on the path to prosperity lie at regular intervals along the way. Similarly Fibonacci numbers can be used to predict certain dates in the future. Say the latest move from a high at A to a low at B took 10 days, then Fibonacci time projections would mark 6 (0.6 x 10), 10 (1 x 10), and 16 (1.6 x 10) days ahead as being days where the trend may change again.Ichimoku day counts are plotted from intermediate highs and lows and we watch for where they cluster in the future. The more counts that end at, or close to, the same day in the future, the more likely that that day will see a trend end and reverse. Again, all these numbers were discovered through trial and error - in what must have been a nightmare job.
Kihon Suchi
Timespan numbers are used to project how many days ahead an interim high or low is likely to occur, known as Kihon Suchi, or “the day of the turn”. So, from an important low of what should be an N pattern, if A to B took 9 days then point C will be 17 days ahead (9+9-1) of point A, and point D, 25 days from the start date (9+9+9-2). The N formation should then finish when the third wave matches the height of the first one and around 9 days after point C.From the completed N wave the next move might be an I wave lower which should take about 26 days (one of the preferred day counts). This is so that the I formation most closely mirrors the size of the N wave which took 26 days to develop.
When working out day counts these do not necessarily have to succeed each other perfectly. Hosoda-san allowed for gaps in the count: say, one to five days, around important highs or lows. As already mentioned markets are more likely to reverse/accelerate around the time when the Clouds become very thin, so too the day counts start from around the time of significant highs and lows. Of course, this makes it a lot easier to fit the 9, 17, 33 day counts to what has actually happened. It also makes it far harder to decide exactly on a date in the future when the move will end.
Not to be treated with too much precision
Note again the qualifying use of the word “about”. All of these are indications and are not cast in concrete; one must allow for a day or two either side of each and every number, coupled with the fact you are subtracting for the attachment parts. Just as the Clouds give you some idea when the trend may change, so too do these projections. For example, after an important high one might expect 33 or 42 days while the market pulls back and consolidates.Intuitively, the 9 day and 26 day numbers make sense, nine being just under two weeks and 26 what had been the average working month. These are similar to ten and twenty-day moving averages used by manyWesterners.
Timing is implicit in the pattern analysis
Beyond that, I feel there are simply too many potential turning dates to cope with. Cyclical analysts will beg to differ, but I feel that a more flexible view suits me. For example, if a dramatic Shooting Star reversal candle forms, I would rather go with that and forget about all possible day counts.And what if the move turns out to have an extension?Where do I then start my subsequent day count? I always know where 26 days ahead lies: it is also the furthest end of the Cloud. Thirty three days ahead is just a few more, so again you might start guessing what the Cloud might look like on that day. If, say, it has been getting thinner over the last week or so, then I assume it will become more so by then. So here, while I am looking at where important turning points might lie, it is the Cloud itself rather than the wave counts or price targets that are the basis of my view; if and when the two coincide the likelihood increases.Western pattern analysis also gives hints at dates ahead when things are about to change. For
example, a head-and-shoulders top: the perfect (idealised) example of which will be symmetrical and have a horizontal neckline. If the left shoulder took three weeks to form, and the top of the head was two weeks after the first neckline point, then the rest of the formation will probably take five weeks to complete - two weeks to the second neckline point and three weeks to form the right shoulder.Again, in a perfect world it will be the same height as the first one. In other words, as in Ichimoku charts, the timing aspect is also inherent in some conventional pattern analysis.
9, 17 and 26 – the most useful day counts
I think most Japanese traders would agree that the dates that are most useful for the timing part of this analysis are the 9, 17 and 26. Often 44 days seems to work quite well, but in my opinion not often enough to make it truly crucial.A great deal of subjectivity is involved
I know that some Japanese traders feel there is too much subjectivity involved in combining the day counts and the wave counts. One dealer may feel that the formation starts on the exact day of the time analysis, while another would prefer to start with the last significant high or low. Depending on the initial choice, and add in the fact that one can use some leeway with the day count, the two dealers would end up with different price and time targets. A similar question arises with the numbering or lettering of the waves.Where interim highs and lows are close, but not exactly the same, should these be labelledA’ or B, 2’, or 3?My gut feeling is that this might lead to the kind of lengthy discussions most often heard in ElliottWave circles.Also remember that over time markets and instruments change and some methods of analysis will work better under certain conditions. Do not be afraid to discard methods when you feel the market has changed. For example,
• For very explosive markets which are trending very strongly, I immediately discard the Relative Strength Index.
• When markets are in very strong bull or bearmoves, I tend to discard all wave counts, and therefore most price targets. Nevertheless, I keep an eye on time targets.
Likewise in sideways moves I will discard moving averages and the Ichimoku Cloud.
That’s the end of looking at the three principles of Cloud analysis.We’ll next turn to applying all this in practice and work through some real-life case studies.
Comments
Post a Comment