Friday, February 8, 2019

Some Divergence

The primary purpose of this post is to show an example of upside divergence in the Fisher Transform, perhaps in time for it to make a useful contribution. We have shown this same exact chart before, but with the 'potential' divergence. That is, when we showed it (see the third chart down at this LINK), the divergence itself could not even be confirmed because there was not yet a lower histogram bar. That has changed. Now there is a divergence - with fully two lower histogram bars.

S&P500 Cash Index - Daily - Confirmed Divergence

Over the last two days the divergence itself has now been confirmed. But like any oscillator or divergence indicator, having confirmed a divergence is one thing. Just like the lines of the MACD crossing over to signal a divergence, the actual price action must follow through. That would most likely mean that a price bar with lower low bar than today would likely be needed to better establish that the indicator is most probably on its way to the zero line, and / or below.

Today is a reasonably good example of why one doesn't necessarily want to be the first pioneer out on the prairie. The day provided plenty of 'snap back' in price movement, and there is no reason to have acted rashly on the first lower candle.

Yes, if you haven't followed all of the comments, in the S&P500 cash, we could count a truncation high where the Dow made a higher high two days ago. And, because of that truncation, it seems like the market is telling us, 'the upward wave ended'. Then, we could count five impulsive waves down to yesterday, and a potential Flat wave up today - which may only be partially completed. The 'b:3' wave of that flat was more than 100% the length of the 'a:3' wave, making the flat a very good candidate for a true 'expanded flat'. In the S&P500, the b:3 wave was less than 138% x a:3, while in the Dow, the b:3 wave was 162% x a:3. That's a pretty impressive set of measurements, but hopefully it will help answer some questions for you about how deep b:3 waves can go, and still be valid. The two indexes are very different, and their b:3 waves did some different things. Hopefully, that is not too hard to swallow.

The divergence above lends some credibility to this count but it is still not definitive. The February highs would have to hold, and / or daily prices resume the gap-open and follow-through action lower to more clearly establish a down trend.

Have an excellent start to your weekend.
TraderJoe

87 comments:

  1. Thanks TJ.
    May we have video this weekend for long term count if possible

    ReplyDelete
  2. ET, had a question on one of the intraday chart you posted on earlier post.
    In https://invst.ly/9-jfe i and iv are overlapping by 0.25 point.
    High of i is 2690.5 and low of iv is 2690.25.
    Are we ignoring the overlap here to call ((1)) as impulse ?

    ReplyDelete
    Replies
    1. No E_D_T. I said in my post yesterday @ 3:23 pm, that first wave up (i - v) was a contracting leading diagonal. So, overlap was expected.

      Delete
    2. Ok that makes sense but isn't time taken by iii > i in that contracting LD. I thought along with price even time should for v < iii < i in contracting LD ?

      Delete
    3. Usually, not always. It's not a rule.

      Delete
  3. Joe -
    A Neely question. I hope you can follow the chart and the points shown. This would seem to be helpful in assessing our move today.
    Thanks, when time permits.

    https://imgur.com/KpGFeQG

    ReplyDelete
    Replies
    1. Possible to move A to the lower green bars just after the circle. Then the short down wave is another 'a', then the up peak is 'b', then the down wave is 'c' as a diagonal of B. In other words b is a flat of a B flat. Then B is longer than A, and it is a regular flat. And that is why so many people thought they were counting a diagonal.

      Delete
    2. Can we make that any more complicated? lol I follow what you're saying. So, to figure ab extension for "C" of larger flat, I would measure from your "A" that was right after my circle?

      Delete
    3. Also, at what point will "C" have gone too far, negating the flat scenario?
      Thanks

      Delete
    4. For measurement (not type of pattern determination) I would use the high in the cash market at the end of the day on Feb 7th. Then, beyond 2.618 would be too far. Should not quite be to 1.618 yet, measuring that way.

      Delete
    5. Thanks. If Im figuring correcting, that takes us within an eyelash of the prior high in cash. So, effectively, a new high negates as would be expected. Should that occur, it will be interesting to see how our assumed flat will then be "recounted".

      Delete
  4. The move back above 2700 is in my opinion a huge red flag for the bears. That reclaim of a prior pivot is often a warning that the anticipated change of trend has in fact not taken place. It has been my experience as a trader that the initial wave down at significant trend changes announces the arrival of the new trend by taking out MULTIPLE prior pivots, such as support/resistance shelves and open gaps. I originally thought the 3,3.5 expanded flat for a second wave was a good possibility but the collapse in VIX, low volume, and move back above 2700 suggests the bulls are not quite done.

    ReplyDelete
    Replies
    1. Depends what you are defining as the pivots and the 'multiple-pivots' since you have not shared those definitions with us. This leaves you in "opinionated control". It's not like saying "prices have gapped down over a trend line drawn from the Dec 24th low, though the Jan 29th low". This is a specific, quantifiable and verifiable observation by everyone. And a back-test of such a trend line, whether is makes a new high or not, is certainly possible.

      Delete
    2. Verne, look back to March 2018. There were a few days just like last Wed-Fri.

      Ditto in 1929 and 1987.

      The market makes it hard for bulls and especially bears. So when it goes, no one will be buying, hence it crashes.

      Plus Puetz, plus all the 7s at my blog. It's on. To sub-2000 by 25th.

      Delete
    3. I did ET. I cited resistance/support shelves, AND open gaps. Sorry if that was unclear...

      Delete
    4. Points well-taken watchman.
      My own trading approach assumes CB driven price rather than "herd sentiment". I have noticed over the years that how price behaves around contested pivots can offer a good clue about whether they are buying or selling, which the generally do with leverage.

      Delete
  5. Simply an observation. Looking at a 2hr chart of cash, we see 124 bars to date from Dec low. The AO pattern looks (to me at least) much more like what would be expected at wave 4 than an initial thrust down in a larger correction.

    ReplyDelete
    Replies
    1. That seems to ignore channeling. I can't get that count to either channel properly or wedge properly. Neither one. And it would make wave 2 very suspiciously like a triangle, which it can't be - although one could try to deny it by saying, "well, it's just a running two; ignore that later squiggle." All possible, but seems less likely.

      Delete
    2. I believe the DJI chart I posted below might address the channeling aspect. As to count, I've been going over and over this as well. It just seems odd to have such a nice "fit" while at the same time a count that does not. Admittedly, I'm employing a bit of the "if it walks like a duck, and quacks like a duck...." theorem, lol. Thanks for giving this consideration!

      Delete
  6. Here's a little weekend musing material. Struck me as interesting. Could certainly be off on my logic (or the lack thereof). :o)

    https://imgur.com/lsmXhSh

    ReplyDelete
    Replies
    1. Peak of the EWO on the 3rd wave is lower? Not higher? Third wave doesn't bust the channel to the upside? Very often a wave in a near-perfect channel is a zigzag.

      Delete
    2. I follow. The DJI chart I did (same time period as SPY above) shows a very nice channel. Perhaps this is actually telling us its a complex corrective move up (rather than just an abc)? Double/triple zz perhaps? This could account for the AO showing perhaps?

      Delete
  7. Hi Trader Joe,

    Do you have any recommendations to start learning EW from scratch, be it books, youtube or what have you?

    Kind Regards,

    Jackson

    ReplyDelete
    Replies
    1. Start, as always, with The Elliott Wave Principle, by Frost and Prechter.

      Delete
  8. Here's the DJI version of the chart I posted above. Perhaps this will shed add'l light.

    https://imgur.com/BScSXGT

    ReplyDelete
  9. Start with The Elliott Wave Principle by Prechter and Frost :)

    ReplyDelete
  10. A fresh look at the big picture. All comments welcome.
    https://imgur.com/fjYhtDn
    https://imgur.com/Vycc84i

    ReplyDelete
    Replies
    1. Blue 5 (smaller degree) is larger than all of Green 1 (larger degree). And that is a degree violation. Why is 'everyone' trying to make a new high?

      Delete
    2. because its a second wave high....

      Delete
  11. The following deals with the issue of time as it relates to the wave b in 3 segment patterns (zigzags/flats). This relates to our current flat in progress as I understand the labeling to be. If this premise does indeed apply as it appears to, then it seems some clarification may be needed (for me at least).
    Please note that my choice of degree labeling is used solely to differentiate between a lower/higher degree. Thanks in advance for clarification.

    https://imgur.com/ywlpACe

    ReplyDelete
    Replies
    1. I don't think you picked the low for 5. That changes the timing.

      Delete
    2. Here's why I selected low for 5.

      https://imgur.com/gecJTJf

      Delete
    3. Looks like you are just picking what seems 'visually' obvious. Did you actually try to count down wave 5, and maintain degree labeling within 5?

      Delete
    4. I'm just using a charting application put forth by Mr. Neely. His video is available at https://www.youtube.com/watch?v=WZWAaupdDkA It is possible I have misapplied, but should be easily verified.

      Delete
    5. Addendum: It was my takeaway from the video that the method he showed addressed where to begin a subsequent price analysis. So, in his example, he used the next higher low to begin his analysis of an anticipated move up. This may not be where the prior move down ended, he did not point this out. This is how I viewed it. So, on my chart, the 5 may indeed go on a prior low, but the next move (A) would begin as shown using his methodology (if Im applying correctly). Hope this helps.

      Delete
  12. Joe, I have seen a count where Jan 2018 was primary 3 and we have or are completing primary 4. Is that a valid possibility or is it a rule breaker?

    ReplyDelete
    Replies
    1. Depends where Primary 1 and 2 are.

      Delete
    2. Sorry, this count was proposed this morning by another poster independently.

      Delete
  13. Regarding the charts I posted for 2009 move up, all 3 showed higher AO peaks at our assumed wave 5. It was suggested that in this case, the AO should be adjusted to reflect % change vs absolute change. While I'm not sure I support this premise, nevertheless for discussion purposes, I have added a PPO Osc to reflect % change. I have used the same 5/34 simple avgs, hl/2 datapoint as the AO uses. The results are shown in the SPY, Nasdaq, and DJI charts referenced below. I believe our question remains.

    https://imgur.com/8RH4Fkb
    https://imgur.com/2LAtbCl
    https://imgur.com/DCNMGgM

    ReplyDelete
    Replies
    1. It's only one part of the equation.

      Delete
    2. If there are other determinants:
      1. Were they included in the write-up for Eight Fold Path? Did I overlook?
      2. What are they?

      We use the AO on a 5min chart and a daily without any adjustment. Yet, a daily chart is 78 fold higher timeframe (5minX12X6.5). My charts are 17 days vs 1 day. Seems odd we suddenly need to adjust for this as it has never been an issue prior that I'm aware of.

      I (as I'm sure many do) rely heavily on the EWO. If new evidence is available that calls into question the original tenets, it would obviously be beneficial to have the information supplemented in some way.
      Thanks!

      Delete
  14. ET, again with due respect, I did not say "everyone" is trying to make a new high. They do exist however. OEW and "Nate's Market Analysis" are two that come to mind. More importantly however, I am having trouble coming to terms with some aspects of the degree violation rules. For example, if wave 3 is in the vicinity of 2.618 times wave 1 then at least one of the internal waves of wave 3 (i.e. 1 of 3, 3 of 3 or 5 of 3) is virtually guaranteed to be larger than wave 1. This is very, very basic arithmetic. Case in point … bull market starting in 1942 (sorry I cannot supply a chart of the whole bull market, I would if I could but anyone can look it up) went as follows:
    0 - 7.47
    1 - 19.25 (11.78 points)
    2 - 13.55
    3 - 94.72 (81.17 pts & = 1 x 6.89 where 1.618 x 1.618 x 1.618 x 1.618 = 6.85)
    4 - 60.96 (This wave was an expanding triangle lasting from 1966 to 1974. I do have a chart for this wave and is attached.) Market then went to new highs.
    So adjudicating on degree violations solely based on points every time simply doesn't wash. There either has to be another workable methodology for degree violations or EW moves one step closer into the hemisphere of art.
    https://imgur.com/2xzMzV4

    ReplyDelete
    Replies
    1. You can add to your list, Daneric and Todd Gordon for a couple of others with blogs or videos. But, numerous people on this blog keep looking for the "up" count, as well. All I can tell you is I'm working on a formal structure for degree labeling, and it is not quite done yet.

      Delete
    2. Thanks for your reply. Sounds like a monumental task but I wish you a high degree of success with the formal structure as we all do.

      Delete
    3. its possibke to count that period without degree violations. wave 1 from 1942 ends before 1974 and 1974 is bottom of wave 2 ive done it many times no violations. any 2.618 extension is likely to be strong resulting in waves 2 and 4 correcting uoward like running flats. so take .618 for 2 waves 1 for another for up waves and you need upwatd corrections to make up small difference of 2.618-1.32-1=.3 approx... not unlikely or even unexpected.

      Delete
  15. https://imgur.com/a/p1vtCuF
    this addresses EWO from 2009 (use of log prices )
    and LT count (alternates exist)
    no degree violations
    joe can comment if he wants to validate any of this.
    as of now using very simple rules of degree violations everything is reconciled

    ReplyDelete
    Replies
    1. Marc - can you look at the DJI as well? How many bars from beg. to end?
      Thanks!

      Delete
    2. Addendum: Just curious as to where your highest EWO peak is on DJI. Dont need
      the chart or labeling.
      Thanks!

      Delete
  16. dji will still have higher ewo 5th wave
    count the months its monthly
    BUT its extended 5th wave, guidelines written from extenended 3rd at least thats how i interpret them.

    ReplyDelete
    Replies
    1. Thanks for looking. I trust your EWO is % change (like a ppo) rather than absolute change (like AO)?

      Delete
  17. its a difference of logs. so its an absolute change of logs which coincides with percentage changes in prices

    ReplyDelete
  18. Thanks Marc with regard to your input concerning log scale vs non-log scale. Extremely interesting and thought provoking. At the end of the day we are all looking for the correct count. What remains perplexing is using log scale does not invalidate my proposed count (2009 to present). Regarding the possibility that 1966-94 was a wave 4 expanding triangle is still plausible with 3 waves complete to 1966 (Wave 1: 1932-1937, wave 2: 1937-1942, wave 3: 1942-1966). Your count from 1932, particularly the use of running waves, is more creative but one cannot say that it is incorrect just as one cannot say the more conventional count that I prefer is correct. Interesting times and appreciate your input.

    ReplyDelete
  19. NB, total novice me, but let's see if I am picking up some useful information here:

    I see a 5-wave move up from ES 2680 on 8th February.
    I counted the bars of wave 1, and wave 3 was longer.
    But waves 2 and 4 were exactly the same length, with wave 4 terminating today at 2711.75 (triple touch of that level.
    So now we're in wave 5 up.
    If it's the same length as wave 3 it'll terminate at 7pm GMT (which is another 7). And it might get to the 50% fib of the entire move from the October-December move, which is at 2729.

    I've used these waves to hedge some shorts today, and it's working so far, so my thanks to TJ and other contributors.

    NB, there were 2 days before the crash cycle began in 87 and 29 that were clearly pauses, indecision, look at the candle shapes. I think we may have two such days now, and head off downwards on Wednesday.

    ReplyDelete
  20. If we go on to a new high/retest of last week, could the following be what is occurring?

    https://imgur.com/PT5dPMx

    ReplyDelete
    Replies
    1. This would help explain the AO.

      Delete
    2. As it appears the potential "b" waves in the chart <"a"s, to maintain the structure it seems these would be WXY (with wxy) for each leg.

      Delete
  21. Just noting a 'current' Fib relationship.

    https://invst.ly/a07b5

    TJ

    ReplyDelete
    Replies
    1. Cash gap just filled; bye-bye as w-x-y?

      Delete
    2. IF it truly fails here; this would be the w-x-y.

      https://invst.ly/a07fb

      TJ

      Delete
    3. Best Alt would be c:5, with this as a fourth. Nothing very conclusive either way until overlap of 2,700.

      Delete
    4. I'm watching the range closely. If they hold the range for the noon ET news, then a possible fourth wave triangle, going along with the c:5.

      Delete
    5. Cash range busted to the down-side.

      Delete
    6. A second up movement today is larger than the first. A leading diagonal is possible (lower) or a larger triangle still possible without overlap. Sorry, but they are valid alternates at this point.

      Delete
    7. For the larger triangle you would count the first three waves of a triangle, as follows:

      https://invst.ly/a085x

      TJ

      Delete
    8. For the diagonal, it would be best if there was only a slightly higher high, but lower than ((2)). This is why they are valid alternates.

      https://invst.ly/a087y

      TJ

      Delete
    9. As of 12:10, there is a sufficient higher high, for the fourth wave of the potential diagonal; OR the start of the (b) wave, up, of a triangle.

      Delete
    10. Here's the update ((4)) of the potential diagonal would have now taken more time and price than wave ((2)), and they are both zigzags. But, a larger triangle 'could' still happen.

      https://invst.ly/a08sq

      TJ

      Delete
    11. As a practical matter, that last high was about the limit for a wave ((4)) of a diagonal, with the c wave of ((4)) as a contracting diagonal itself. Higher than that, and it is most probable a larger triangle is on.

      https://invst.ly/a0921

      TJ

      Delete
  22. This comment has been removed by the author.

    ReplyDelete
  23. Thank you for the "play-by-play", Joe!

    ReplyDelete
  24. Just fyi - the DOW has a lower low and an overlap of prior waves.

    ReplyDelete
  25. A diagonal wave ((5)) was just validated as being the correct length below 2,703.37. And, it has wave ((5)) longer than wave ((3)) in time. And wave ((3)) longer than wave ((1)) in time. BUT now the real test begins. Any, diagonal can also just be its triple-zigzag counterpart: w-x-y-x-z as a fourth wave. While the triangle is no longer the alternate; the simpler triple zigzag fourth wave now becomes the alternate. The point loss is quite shallow at this time.

    https://invst.ly/a0aej

    TJ

    ReplyDelete
    Replies
    1. Keep in mind, the DOW just got another level of overlap lower.

      Delete
  26. Just one man's opinion folk, but they are clearly defending the round number pivots...they have deeper pockets.... :)

    ReplyDelete
    Replies
    1. Clearly I know what a round-number is. But, the only formal definition of a pivot I am aware of is this one.

      https://www.investopedia.com/terms/p/pivotpoint.asp

      So I continue to be unsure of your terminology, such as stringing together "round number" and pivots.

      TJ

      Delete
    2. You are quite right that my definition is not "formal".
      For trading purposes I define it as any price point around which I see contested "price action", which often happens to be round numbers.
      I know it is bit loose but it really has kept me on the right side of probable direction over the years. I simply want to make sure I can ring the register, however one defines that! :)

      Delete
  27. Cash did break the upper boundary of the diagonal. That's all it HAD to do. It also fell just short of a 62% retracement since this morning's top.

    https://invst.ly/a0avl

    There are insufficient waves to claim a higher high could be made as an ending diagonal. So, I won't at this time. The only other observation I have at this time is that, going into the settlement, the %D of the daily slow stochastic is now 79% (that is, less than 80%, and could lose its embedded status). That should be watched overnight to see if a reading lower than 80% persists.

    TJ

    ReplyDelete
  28. 5 out of the top 11 IBD Leaders made new highs today. Generally that's a bullish sign. On a bearish note, none of the 5 had a strong closing.

    ReplyDelete
    Replies
    1. Almost half of the IBD Top 40 Leaderboard stocks made a new high for the year today.

      Delete
    2. The name of the game is "confusion"...for bull and bear alike!

      Delete
  29. A new post has been started for the next day.

    ReplyDelete