Issue # 78  April 20, 2003
NOTICE:  All contents © 2003 by George A. Ure, MBA except other authors as noted.  This document is intended for the sole use of subscribers and may not be transmitted, reproduced, or in other way used without the written prior express consent of the author.  This publication is by subscription:  $50/year for web browser accessed delivery to a password protected site or $150/year delivered by personalized email in HTML..  Please pass along a copy to your friends and suggest that they subscribe.  To subscribe, send a check to: Guru Press, 1355 West Palmetto Road, #281, Boca Raton, FL  33486-3383.  You may also subscribe using PayPal.  Your username and password are both your email address, in all lower case to access the protected web site, so don't forget to include it!   Address comments and correspondence to: george@ure.net.    Read the disclaimer: http://www.urbansurvival.com/disclaim.htm This report is based on sources believed reliable and makes no specific investment advice.  Before you invest in anything, seek professional advice and remember, you can only spend it once, unless you are a member of the Fed, in which case you can spend it as fast as you can print  it.

Good Math: Bad News >>> Sornette and SARS

Can Markets be Modeled?

This week, a few thoughts about modeling markets, Sornette's work, and the uncanny relationship between Sornette's low and the SARS death peak.  But let's start at the beginning.

I've been fascinated, as no doubt you have been too, with the possibility that markets can be modeled in a sufficiently robust way that we can make predictions worthy of trading.  This week, I want to run through market models a bit and see if together we can learn a little more about their behavior. People have been sorting through various mathematical approaches to markets for a long time, and obviously, the success of such systems is limited because as soon as a system becomes successful, it stops working.

We can begin by looking at different ways to divide up market models by looking at a few of them and trying to state just how it is they go about inferring future price information from past performance - a tall order in subjects like weather - and one must ask if financial markets are so complex so as to make modeling impossible.

Model: Elliott Wave

The Elliott Wave model, popularized by Robert Prechter's books and website ( www.elliottwave.com) can be summarized as the belief that markets generally move in 3 or 5 waves with the occasional 7 wave and X waves thrown in to make matters "fit" when 3 or 5 waves don't.  I've read a lot about Elliott Wave theory and it seems to fit some of the time, but its primary difficulty is that there is almost always an "alternate count" that usually shows up at exactly the right moment to gobble up my personal profits.  Elliott is a system based purely on price information over time - an important distinction as we'll see.

Model: Dynamic Gann Lines

Dynamic Gann Lines are geometric lines, again like Elliott, based purely on the price action of the market, which seem to project ceilings on upward moves and floors under corrective action.  Once a DGL is broken, the trader is supposed to find the new "geometry" of the price action and infer moves from that.  I haven't done much work in the DGL area, but there is a fairly active group, like with Elliott analysis, that believes in DGL.  I have a concern that I expressed in an earlier Inside Report ( http://www.urbansurvival.com/nldec27.htm) that Gann Lines may not work when we approach saturation levels in markets, which I told you last week, was where we seem to be about now.

Model: Bollinger Bands

The idea behind Bollinger banding is that markets tend to move is sections of parabolas, such that you can draw a smooth curve which will describe the "channel top", the "channel bottom" and about midway between is the middle of the channel.  By looking at this financial "channel", you're supposed to be able to able to predict the general price direction and make some real money.  But again, like DGL's and Elliott, it is a prive and time view.

Model: On-balance Volume

I have to admit a personal attraction to this model, not so much because of its creator (the legendary Joe Granville) but because it is a very simple way to interpret market action that utilizes something more than price and time.  It also works on the volume of trade in shares.  Essentially, what Granville did was observe that if there is more up volume when the prices moved up than down volume when prices moved down, the stock was headed in the up direction.  When the OBV flattened out, or when volume on the downside trended heavier than upside, then it was time to bail or look at going short.  The problem of course is one of time.  There's a lot of disagreement among traders over what time base to consider when looking at OBV.  If you look at a 5-minute chart, you might get a completely wrong-headed idea of a stock's direction, compared with a 60-minute chart.  That's where things get dicey, because most of us don't have the time or inclination to stare into a market monitor for more than about 30-minutes a day on our lunch hours and coffee breaks.

Model: Buying Power Indicator

This is very similar to Granville OBV, but it has been vari9ously branded by several outfits that claim to have unique ways to  see the divergences between prices and volumes by looking at it this way or the other.

Model: Custom Time Projection

Here's a market group that sells something called a Custom Time Projection which is a direction projecting histogram.  From looking at it, you're supposed to figure out what the next high and low dates are.  It looks similar in presentation to www.lookaheadcharts.com. Groups like www.marketpicture.com base their work in this area on "time cycle ratios" and "time count projections" which we don't even pretend to understand.

Model: Candlestick Charts

The Japanese have been great followers of a system of analysis which is similar to our bar charts, but which emphasizes the trend by drawing a "candle" which is "open" [white] if the close was higher than the open, or "closed" [black] if the close was lower than the open.  From staring at these candlesticks, some Japanese traders have supposedly divined certain patterns which are tradable.  But again, I don't get too excited about them as it's just price and time to me - with the granularity of a closing and opening price in each days data.

Model: RocketScience Trading

RST is Bill Shepler's proprietary trading model, and its of a general model class that we could label "highly complex" in that it takes in as many aspects of available trading information, such as price, time, volume, directional trends, and so forth, and then applies a "learned" set of behaviors in reaction to the market.  Some models of this genre are index oriented (as is RST) while others are single stock systems.

Where people end up spending a huge amount of time and money is in figuring out just which system apply to which stocks at any one moment.  Again, as best I've been able to observe, the advantage of an index oriented system is that something like Worldcom accounting blowing up will not have too much influence on an index, but if you were in that single stock and long at the wrong moment, you could have been wiped out.

Money on Analysis

I found an email from a fellow who said he was retiring from the markets, some time ago, and he had posted the following to a couple of discussion groups.  Look at the list of ways this fellow had collected in order to try and psych-out the market:

TRADESTATION PLUGINS AND INDICATORS $200 usd

: %c dt number 1 & number 2
: 12system
: aberration system
: abraham system and indicator
: adaptive indicators 2.0
: adk1floor
: alternate swings
: as s&p bear system
: as s&p raptor system
: asc trend pro deluxe
: asctrend 3.51 for ts2000
: bail3 for daily system
: bill williams chaos
: bill williams ela
: bill williams new trading dimension
: bill williams note
: binary wave study
: bob buran systems
: bollinger band differential
: bos bse ls mkt system
: bottom fishing
: bullseye indicators & paintbars
: candle forecaster expert
: candlestick patterns indicator
: catching waves with a multi cycle momentum
: catscan catscan number 1 system
: catscan catscan number 2 system
: cfb trading system
: chaikin money flow indicator
: chande's dmi & momentum oscillator ela code
: channel
: charles le beau 25 x 25 bond system + manual
: charles le beau big dipper bond trading system
: charles le beau crossbow
: charles le beau fsy2
: charles le beau mack trading system + manual
: charles le beau prudent s&p trading system + manual
: charles le beau roughneck co trading system + manual
: charles le beau strategic swiss franc
: charles le beau sward
: cherry picker
: clyde lee moon phase system
: coa ctg
: cole's range
: commando ii trading system
: confluence indicator
: coolspy for ms 6.5
: coppock curve indicator
: cup with handle
: cycle finder
: cycle surfer
: cycle trader 5.0 by walter bressert
: cynthia kase ela
: dale legan catching waves
: dale legan's confluence function
: david wright's spoon indicator
: day care system
: dbs trading system
: delta phenomenon notes and ela code
: dennis meyers indicators
: dennis meyers short term indicators v2 and long term v4 and 5
: dgl finder convergence indicator
: dinapoli indicators
: dmark key system
: dr. elder systems
: dualthrust
: dunnigan's way
: dynamic breakout system variation b
: dynamic gann lines auto for ts2000i
: ehlers indicator
: ela breaker
: ela manual
: elas
: ellery colemans systems and indicators
: elliott wave system
: entry22
: entrypoint 2.2
: entrypoint 2000i with documents
: equity curve ind
: ets dynatrend 2000
: ets trading system by michael mermer
: exhaustion systems
: gary smith systems gary smith 2 system
: gary smith systems s&p day trading system
: garys parabolic average
: greg morris insync index
: greg morris vol 1 volume and price vol 1 volume and price
: greg morris vol 2 time tested indicators vol 2 time tested indicators
: greg morris vol 3 japanese candlesticks vol 3 japanese candlesticks
: greg morris vol 4 different indicators vol 4 different indicators
: greg morris vol 5 market breadth vol 5 market breadth
: gunslinger
: gunslinger two
: hit and run by jeff cooper
: hitandrun
: inside advantage
: insight
: investor's dream for ts includes manuals
: jan arps swing box docs
: jan van arps ts indicators
: joe krut's 12 ultimate systems
: john clayburg systems & indicators
: jurik indicators for ts2000i
: jurik ock 7925
: kase
: kase and krut ela
: kaufmann
: larry williams 1.28 breakout
: larry williams ela's from seminar complete
: larry williams inner circle indicators
: larry williams system for ts4
: linda raschke indicators and systems linda raschke connors anti
: linda raschke indicators and systems linda raschke critical day
: linda raschke indicators and systems linda raschke lbr edge
: linear regression with calc
: mangeeze program
: market analytics exhaustion bars
: market analytics fractal toolkit
: market analytics ma_predict
: market analytics momentum toolkit
: market annihilator system
: market blueprint 7 commercial indicators and systems
: market blueprint trading system
: market facilitation index
: market reactivity system by al gietzen
: mbbo system
: mesa
: mesa 2000
: mindfire systems catscan catscan number 1 system
: mindfire systems catscan catscan number 2 system
: montecarlo
: murrey math indicator
: mwdxavg indicator
: mx capital trading
: natt 2000
: natt collection
: neal weintraub ela code ticklemeneal
: ninja turtle
: omnicom v5.17
: omnicom v517 for ts2000i
: paint alternating sessions
: paint day of week
: pathfinder trading system
: perry kaufmann
: phase calclulation
: phase technical analysis indicatr library & methodology
: pi trading system 2nd uppi system
: pinpoint systems
: plad system
: plad system mutual fund timing system
: playwave addon
: playwave by neurotrader
: polynom
: precisiontrader buysell point indicator manual
: precisiontrader buysell point indicator program
: profit trader 5.0 wallter bresset
: profitability test 2.0
: profittrader50
: prosuite indicator pkg by trade concepts
: psg sys number 1 intraday
: psg system number 1 intraday
: quantum leap
: r breaker system
: r mesa
: r mesa 3
: r mesa 4
: rangebrk
: rangebrk for ts
: reactivity indicator by dave deluca
: recurrence iv system
: reflection
: remarkable s & p trading system v1.3
: robert miner's stochastic rsi
: roos hook
: rpb bestcorrelation indicator
: ruggerio
: ruggerio indicators tttm disk number 1
: ruggerio indicators tttm disk number 2
: ruggerio indicators tttm disk number 3
: ruggerio indicators tttm disk number 4
: ruggerio indicators tttm disk number 5
: ruggerio indicators tttm disk number 6
: ruggiero stock traders toolbox
: safir 3.2 workingsafir 3.2.0
: sector wiz systems
: serendipity v1.1 bond trading system
: simple adv
: simplicity system
: sine weighted moving average
: sirtrade cycle
: sirtradecycle
: sirtsd98
: sophie
: spike 35
: spiral
: spoon
: stafford sp day trading system
: stenberg brothers ast indicators
: stochastic rsi chande
: street smarts
: street smarts ocean indicators
: street smarts swing trading indicators
: success system
: sup res
: support and resistance pivots
: swing activator system
: swing machine
: swing trader 4
: system trading & development club files
: td scalper ts
: the ela manual
: the mom144 day trading system
: thrust retracement
: tim tillson's t3 average
: time series forecast indicator
: times series forecast indicator
: todd mitchell indicators with instructions & examples
: tom demark new market timing worksheet
: tom demark systems
: tom demark td_scalper ts excel
: tomdemark systems
: toratrade for metastock
: trade cycles 2.0
: tradeworks random number generator
: trend reflexion
: trend rider
: trend stddev
: trend trader
: trendchannel pro
: true contrarian
: ts get v7 data transfer and manual
: tsci trendwatch
: user function like iff immediate if returns string rather than number
: variation b of dynamic breakout system
: virtual trader 2.05
: wa trading package
: walter bressert cyclefinder
: walter bressert indicator for msmetastock
: walter bressert indicator prosuite
: walter bressert stochastic indicators
: walter bressert the cycle trading manual & ela
: whisper system
: william tell coffee black box version
: winmidas
: winning edge
: wisdom of the ages
: wisdom of the ages manual
: wolfe waves
: wonderbars
: wwlink & sc
: z score

About the only thing I can tell you from looking at this list, and reading the rest of the note that went with it, was that whoever the fellow was, he managed to retire, but he made no claims about whether he did so in high style and luxury, or he was simply tapped out and was hanging up his trading shoes.

In December of 1964, Dr. Henry Kaufman talked with the New York Society of Securities Analysts, about his book "On Money and Markets" and described from the the biases that make market forecasting such a difficult job: ( http://www.nyssa.org/abstract/kaufman.html)

Besides Bill Shepler's work which is not widely available except here, there's been much public interest recently in Didier Sornette's work http://www.ess.ucla.edu/faculty/sornette/prediction/index.asp#prediction  which predicts the S&P 500 will be in the range of 660-670 by the end of 2004. That would be a 26% decline, and would bring the Dow under 6200.

Sornette's work, based on his study of math behind earthquakes, suggests that the outcome of the a complex system, like the markets, may be foretold by looking at some peculiarities that happen long before the peak.  But for trading purposes, Sornette's work is a little long in the tooth for our personal taste, as trading is more exciting that a buy and old, or in Sornette's forecast, a "sell and hold" strategy.

As much respect as I have for Bill's work - and that of Sornette - these approaches have always be a little troubling in that they are purely numerical and based on strictly "market-recognized" factors.  I personally prefer the kind of forecast that is based on not-necessarily-stock market events which can be quantified.

If you've been a reader for a very long time, you may recall that leading into the Y2K non-event, we published something called the Y2K Danger Index and correlated it to the performance of the Dow Jones Industrials.  You'll see in the chart below that as the Danger Index declined, the market went up, but the market also appeared to slightly lag the Index.

The Y2K Danger Index was made up of a number of items that I thought at the time might have some predictive value about Y2K.  It included the price of deep cycle batteries and lead time to get them from the local West Marine store in Seattle, where I was living on the big sailboat at the time. It also included the price and lead times on things like solar panels and inverters, which I reasoned would go up in price (and longer delivery times would become apparent) if the power grid was really going down.  The price of gold was a piece of the rationale, too.  I suggested gold would be a key indicator and it would be reflect the expectations of the "perfect" American family:

To this day I still dust off the notion of the Y2K Danger Index from time to time (including now) and wonder if there's maybe some great lesson we can all learn there.  The Y2K Danger Index seemed to run ahead of the market, which when you think about it really makes sense.  After all, statistics like the inventory levels of the nation's businesses change only after someone buys something.  Econometrics therefore that look at inventory levels are already a month behind the consumer.  Similarly, when something is a runaway hit with consumers, the markets may not know it until after a quarterly financial report or a news release.

To develop the model, and test it over time is beyond the scope of this week's update.  It would require a fulltime person a few months to figure out the methodology.  Everything like getting the Sunday papers from a dozen major cities around the U.S. and looking at the price of a 100 ounce bottle of Tide (as an index itself) and the price of particular branded items before consumers.

It sounds like tedious work, but in the end I think it might prove more useful than our purely numbers-driven approach - the quants versus the quals. If you're so inclined, I can tell you it is sure a lot of fun to come up with your own unique index, which you can correlate with anything that pleases you.  Ultimately, whether it was a waste of time or sheer genius on your part is something only the market will judge when you trade it.  I'll admit to being worked up about Y2K to excess and well in advance, yet the Y2K Danger Index certainly calmed me down and allowed me to take a more reasoned approach than if I had merely read Gary North, the Y2K Timebomb pages, and so forth.

As we enter the next phase down in the markets, the challenge will be to have spent the time right now to develop a trading indicator that you understand intuitively and give it enough time to run so that when the first bottom of the Second Depression gets here, sometime in 2004 by Sornette's work, you'll have your own toolkit and surveying instruments that will give you confidence when it is time once again to invest on the long side of markets.

SARS Update

So how does all this discussion about markets and models and the Y2K Danger Index and all that relate to Sornette's 2004 low???

There's booth good news and bad about SARS this week.  The good news is that the daily new infection rate has not gone much over 3% and the bad news is that the mortality rate has continued to climb and in the most recent daily report from the WHO is now over 5%.  At this rate, by August 21 of 2004 everyone in the world will have been exposed to SARS and the death toll from the disease will be just over 313 million dead.

Are you beginning to see it?  Here's the point of this week's report in a nutshell:

Sornette's predict low in the markets coincides with our projected worldwide exposure to SARS and the peak of the death toll at about 313 million dead based on the latest numbers.

Let me summarize by sharing an email:

Professor Sornette,

 

I appreciate that you are extremely busy, however I was writing a column for my readers this week and I noticed that your work (with Zhou) implies a low in the US markets in mid to late 2004.

 

Simultaneously, I have been tracking the SARS outbreak and note that on August 21, 2004 (or thereabout) on its present course worldwide exposure will reach 6.9 billion – effectively everyone on earth.  At the same time, about 313-million people will have died.

 

Because your work predicting a market low in late 2004 preceded global awareness of SARS, I would like to ask whether:

I would appreciate any comments you might share with readers as I have been engaged with colleagues in event predictive studies involving how subtle language shifts on the Internet may be event-predictive. 

 

It occurred to me that in a quantum sense, we may be looking at similar aspects of one phenomena.

 

Sincerely,

 

George A. Ure

www.urbansurvival.com

On to the Charts!

"Write when you get rich...."

Don't forget to visit Elaine's new site: www.independencejournal.com

George Ure (click here to  email a note)