Large Scale Fibonacci in the Dow Jones

Nov 24, 2008: 2:53 PM CST

I wanted to thank a couple of readers who brought to my attention the Fibonacci retracement grid from 1975 to the 2007 highs in the Dow Jones Industrial Average which recently has interesting if not outright actionable information regarding the current structure and possible large-scale inflection point (support) we just tested.  Let’s look at this development.

Dow Jones Industrial Average Monthly chart from 1975:

Click on the images for larger than normal grids.

From the 568 low in 1975 to the 14,200 high in 2007, we get the resultant Fibonacci retracement grid if drawn from the highs to the lows.

From this, we’re able to see the significance of the 7,382 price level, which contained the 1998 “Asian Contagion” market crash as well as three successful tests (with the actual low breaking the price just slightly) during the 2001-2003 bear market, and recently a (currently) successful test of the level last week.

Let’s zoom in to see these developments in a more compressed chart-scale.

Dow Jones Industrial Average Monthly chart from 1996:

With the exception of an intra-month “nip” below 7,382, that level has held like a rock each time price tested it.  Thus, from a technical perspective, we cannot underscore how important it is that the 7,382 level held (actually 7,449) from a chartist’s perspective.

Coincidence?  That’s the largest argument against Fibonacci principles, but regardless, the level has provided support multiple times which – in and of itself – makes the index level quite significant and should not be ignored from a trader’s or an investor’s perspective.

What are the possible implications from this (currently) successful test?

Again, use your own analysis going forward, but it could mean at least a short-term if not intermediate term bottom in the market, should this level indeed be so significant.  It also means we have an extremely favorable reward to risk ratio from a swing to position trader’s perspective to place a stop beneath this level and ride a potential counter-trend wave up to the 10,000 or 11,000 level at the most.

It most likely means that any shorts need to cover if they have not done so already (Friday and today have been brutal for the shorts, and heaven-sent for the bulls).

What if we break these levels?  Again, trading is not about foreknowledge, but more about managing risk relative to potential reward – and finding high-probability ‘windows’ of opportunity amidst a sea of seeming randomness.

If we break these levels, look for the next downside target to be Dow 6,000.  But that’s… a whole other story.

Afraid to


17 Responses to “Large Scale Fibonacci in the Dow Jones”

  1. Svet Says:


    a lot of bloggers focus on the 2002 low and see the next area of support for the S&P500 in the 600 range given that the rise we saw in the 1990s was pretty uninterrupted so there’s no obvious area of support between here and there (there being the 600ish zipcode).

    I am fairly new to TA and can’t help but wonder – what made the index stop at 768.63 in 2002 from a TA point of view?


  2. Corey Rosenbloom Says:


    Correct – if we break the 2002 lows again, we’ll hit what we call an “air pocket” where there’s no clear support (or targets) until we reach much lower levels like the 61.8% retracement at 5,770. Of course, the old adage “In a bear market, there is NO support” comes into focus, so that’s in play too.

    In regards to why it stopped at this level multiple times, there’s universal theories that I don’t necessarily subscribe to (that Fibonacci is an internal part of human nature or cosmic nature and that we all move to a universal pulse) which caused price to stop at that “cosmically significant” level (such as Gann or others who use ‘celestial math’) but I suspect one could think on such things – I’m more of a realist but one has to admit, there’s some eerie peculiarities to Fibonacci, Gann, and Elliott. That ‘magic’ 1.618 number is the common thread that links them all.

    That being said, I don’t have a concrete answer as to what happened other than the cop-out “Buyers overcame sellers at those levels repeatedly.”

    We’ll just have to call it a mystery or strange occurrence until better or more complex mathematics can help explain these occurrences to us.

  3. Svet Says:

    Thanks. I guess in retrospect, you have to either believe that:
    a) the 2002 low was predicting the 2007 high or
    b) the 2002 low was a random number which would then put in question the validity of the next area of support if we were to break it conclusively this time around.

  4. Corey Rosenbloom Says:


    An interesting “either, or.” The best discussion – perhaps the only discussion – I’ve ever read on this topic comes from Constance Brown’s “Technical Analysis for the Trading Professional” book – it’s a must read that gets so little attention in the trading community.

    Brown discusses a unique method of drawing Fib grids and extending them just in this matter. Start at a low point and then extend them so that a recent swing-low (say, that of 2002) becomes either the 50.0% or 61.8% (or 38.2%) retracement where the 100% (top) will be well ahead of price – just like in this example.

    Did anyone in the world do that in 2002? Probably not – I’m not even sure Mrs. Brown did – but that’s exactly what she teaches (in the book) on a smaller scale (which apparently worked as well with the 38.2% and 50.0% retracement tagging the 2007 high (projecting) down to the recent Dow low – same methodology.

    If that’s the case – if this really does work – then it’s one of those moments in science that changes everything. It goes against everything we know and believe and – if proven effective – will literally shatter the TA world (I think).

    Of course, it’s far easier to accept postulate “B” for many reasons.

  5. Svet Says:

    Very interesting. If you try tthe same on the period from the 1932 low to the 1973 high (weekly candles), the 50% line is within spitting distance of the 1962 low (548.31 support vs. 539.19 low that week).

    Ditto for 1962 low to 2000 high – you nail the 1998 Russian crisis support level at 38.2% off the 2000 peak although one could say that the index simply extended to 1.618x the Russian crisis bottom.

    This resembles data mining but I definitely see where the theory is coming from.

  6. hankypanky Says:

    I have a problem with calling 7382 in years prior to 2007 a 50% retracement of the 2007 high. Prior to 2008 7382 should be considered a retracement of the 2000 high, which was 11,497.If my math is correct that comes out to 35.79%. Is 35.79 a Fib retracement level? I think the fact that 7382 held in 2008 was a case where the bulls were desperate to hold the old support level and is a coincidence that it’s at 50%.

  7. Corey Rosenbloom Says:


    I was careful not to call it such in the post – the numbers/grid was generated via TradeStation’s Fib retracement tool.

    No, 35% is not a Fibonacci retracement level, but this is applying Fibonacci unconventially using methods discussed (as far as I can tell) only by Constance Brown in that you take what you have currently and then set up your Fib grid so that the current support corresponds with a key retracement, meaning you have to drag the grid well above price (chart scaling) to achieve the result.

    It’s something I’ve never done before, but Connie gives an excellent explanation that covers a full chapter in her book with examples.

    Your terminology is correct (we can’t call it a 50% retracement until the grid is complete) but I’m becoming open to interpeting Fibonacci grids – which really become projections – in this unconventional manner.

  8. Forkoholic Serge Says:

    Hey Corey
    are we just in a bigger SPX fractal or is it just pigment of my vivid fractal imagination?

  9. Corey Rosenbloom Says:


    Those would be sort of ‘measured move’ patterns or I’d even go as far as to say bear-flag ‘like’ patterns. It’s possible we’ll get a measured move equal in stature to the previous move, but I think the support from last week will hold and tend to subscribe more to the “Wave 4 counter-move up” thesis that could be playing out currently. Plus, at least in the Dow, the recent level we tested was major (Fibonacci?) support so I’m one of those “I’ll believe it when we see it” or “I’ll change my view once we break last week’s support” in terms of immediate downside projections.

  10. David Says:

    Thats a yearly chart and the the year isnt over yet. We are soon to retest them i believe and I am not sure they will hold on the second try.

  11. Corey Rosenbloom Says:


    I do agree the level will be broken, but I see the potential for the Elliott 4th wave (ABC) which should play out into early 2009. There’s almost always a ‘holiday’ or ‘end-of-year’ rally which would play into the “Wave 4 theory” so I see the indexes rising through year’s end perhaps along with the ‘hope of change’ from an Obama administration, only to have something happen in early to mid-2009 which slams the indexes back down into a fractal 5 wave impulse down which takes out the lows at that point.

    But again, we try to take it day by day as new information becomes available.

  12. David Says:

    I am amazed at the name I have seen people call Bulls in the last few days from ” bullt*rds ” to ” mouth breathers ( lol ) but i suspect you are correct although one last push down over the next days is likely. I did trade that im in the SKF and SDS.

  13. Steven Says:

    Hi Svet/Corey

    In response to the question on why the S&P hit a low of 768.63 in Oct 2002, perhaps you can draw a price retracement using the monthly chart from the high in Mar 2000 (1552.87) all the way to the low of 768.63. It looks eerily beautiful how the the Fibb levels at 76.4%, 61.8%, 50%, 23.6% acted as resistance to the 4 rebounds when the S&P was on its way down.

    The same observation can be seen if one draws the retracement from the Oct 2007 high. The 76.4% and 61.8% levels acted as resistance & coincide with the rebounds in May & Aug 2008. However, after Aug 2008, the S&P continued to drop and failed to rebound but instead bypassed the 50%, 38.2% retracements. I wondered why??

  14. Man4urheart Says:

    Why is Fibonnaci grid not drawn from lows of 1930 to high of 2008 and also a long term trendline?

    If you draw long term line from 1930 to low of 1982, you will see the up trendline will give support by 4000! for Dow jones

    Check this

  15. Corey Rosenbloom Says:


    I heard it on TV this morning as I was waking up. Some TV guy asked “Dare I say it – have we hit a bottom?” I immediately turned the channel. That makes my stomach turn to hear TV people talking about bottoms. They’ve been doing so since early this year.

    The ‘bottom’ will come when we put in a higher low, higher high, and then (or simultaneously) break cleanly above the 20 and 50 day (at least) EMAs. Until then, it’s best to follow the trend and stop trying to outguess Mr. Market’s ‘wisdom.’

  16. Corey Rosenbloom Says:


    The more I do work with Fibonacci, the more I’m impressed. I don’t think I’m data mining, as some of the things I’m discovering are occurring by ‘accident’ or otherwise. It’s quite phenomenal. I don’t do much with the Fib retracements other than 38.2, 50, and 61.8 but I should branch out. I just don’t want too many lines on the chart if possible.

    I’ll check those out! Thanks for the tip!

  17. Corey Rosenbloom Says:


    Excellent analysis! You have done an excellent job of capturing the long-term move complete with Fib reactions of the large-scale move, which clearly have more significance than any shorter term Fib grid off daily or weekly charts, or even since 1975!

    I recommend everyone to view his post at the link above.