Famous Trader

Index and Portfolio Updates

The S&P 500 closed up 14.96 to 1473.91, for a daily gain of 1.03%  The IBD-EQ Index closed at 5939.60 up 162.72, for a gain of 2.82%  Not only did the IBD-EQ exceed the S&P in terms of percentage increase, it closed above it's close of July 26th, erasing (in terms of points) the damage done on Friday.   This may just be a by product of the weekly re-balancing, and the boost of new additions to the 100 list.  On Friday, I'll calculate the weekly changes for this week's new additions, and compare them to the overall changes for the components.

The IBD 100 Portfolio is at $92530.92 in equity, down 7.5% since it's inception.  For the same period, the IBD-EQ index is down only 6%.  I'm not even counting commissions, so a real portfolio, following this strategy, would lag the index even more.  This could also be caused by a boost effect to new additions.

The index is re-balanced every every Friday after the close, using that day's closing value as an initial value.  Each stock in the re-balanced index accounts for 1/100th of that previous value.  The new additions are weighted to be 1/100th, and the old components are weighted to be 1/100th.

The portfolio does not re-balance.  We look at the cash available after "selling" the components which are dropped, then divide that amount equally among all the replacements.  This past week, there were 22 stocks dropped, and 22 new additions.  After selling the 22 dropped components, I divided the cash balance by 22.  This week, that left me with $1022.44 per new addition (Only whole numbers of shares can be purchased).

IBD 100 Composition Update

Investor's Business Daily considers the IBD 100 to be the "leading" stocks of market.  This blog is a forum for exploring what that leadership means.  But it's clear that inclusion in the IBD 100 means that the leadership they are talking about is primarily a function of relative strength.  IBD includes 14 statistics with the component table each week.  These include normalized values like 'Relative Strength Rating' and 'EPS Rating'.  We don't know the exact formula for determining the ranking of each component stock.

But broadly speaking, relative strength feels like the most heavily weighted parameter.  Stocks are included in the IBD 100 after they've gone up sharply, exhibiting high, short term, relative strength.  As I collect more data, I'll be able to analyze this more carefully.  Anecdotally, it doesn't seem that dropping from the list is a strongly dependant on negative relative strength. Consider Global Source, Ltd. (GSOL).  This IBD 100 component stock has declined more than 20% in the past 2 weeks, but it still ranks at number 38 in the table.

This focus on relative strength introduces a strong survivorship bias to the IBD 100 Index.  But it may be mitigated by the weekly re-balancing. I'll be discussing this in upcoming posts.

I presume the the IBD 100 components are the stocks with the 100 highest composite rankings, which probably also meet some filter criteria related to market cap or volume.  I'd be interested to learn how those rankings are calculated.  I'm also interested in what happens to the individual component stocks when they are added or removed from the list.  Do stocks get a boost when they are added to the 100?  Do they tend to weaken when they are removed?  It will be interesting to find out. On Friday, July 27, 2007, 22 stocks in the IBD 100 were replaced.

The New Additions this week are:

 

                                                                                       

(VDSI)	VASCO DATA SECURITY INTL
(AMZN)	AMAZON.COM INC
(ABB)	A B B LTD
(DADE)	DADE BEHRING HOLDINGS
(LAYN)	LAYNE CHRISTENSEN CO
(TNE)	TELE NORTE LESTE PAR ADS
(WFR)	M E M C ELEC MATERIALS
(SJR)	SHAW COMMUNICATIONS CL B
(WCG)	WELLCARE HEALTH PLANS
(NVDA)	NVIDIA CORP
(SNP)	CHINA PETROLM & CHEM ADR
(PDA)	PERDIGAO SA ADR
(NE)	NOBLE CORP
(TPX)	TEMPUR PEDIC INTL INC
(KCI)	KINETIC CONCEPTS INC
(RIG)	TRANSOCEAN INC
(UEIC)	UNIVERSAL ELECTRONICS
(B)	BARNES GROUP INC
(FSRV)	FIRSTSERVICE CORP
(MFRI)	M F R I INC
(WNR)	WESTERN REFINING INC
(GMCR)	GREEN MTN COFFEE ROASTRS

The Stocks which were dropped this week are:

 

                                                                                       

(TWIN)	TWIN DISC INC
(AFSI)	AMTRUST FINANICAL SVCS
(CENX)	CENTURY ALUMINUM CO
(NTGR)	NETGEAR INC
(DWSN)	DAWSON GEOPHYSICAL CO
(FCSX)	F C STONE GROUP INC
(ATHR)	ATHEROS COMMUNICATIONS
(DXPE)	D X P ENTERPRISES INC
(SPTN)	SPARTAN STORES INC
(MTW)	MANITOWOC CO
(SCHN)	SCHNITZER STEEL INDS A
(VCLK)	VALUECLICK INC
(OYOG)	O Y O GEOSPACE CORP
(HDB)	H D F C BANK LTD ADR
(HOC)	HOLLY CORPORATION
(NVEC)	N V E CORP
(GRP)	GRANT PRIDECO INC
(BWLD)	BUFFALO WILD WINGS INC
(RATE)	BANKRATE INC
(CTRP)	CTRIP.COM INTL LTD ADR
(MBT)	MOBILE TELESYSTEMS ADR
(WHQ)	W-H ENERGY SERVICES INC

Introducing The (unofficial) IBD 100 Index!

It's here!  The unofficial IBD 100 Equal Weighted Index (IBD-EQ) has been coded. The Index is first calculated for Monday, July 16, 2007.  Here is the algorithm for calculating and balancing the index:

1. Determine Initial value - take the sum of the closing prices each IBD 100 component on the day the composition list is updated (generally, every Friday, after the market closes).
2. Divide the previous index value by 100 to determine the dividend for the individual weightings. (For the first week, we use the initial value (sum).  In all subsequent weeks, we use the calculated index value for the prior weeks components on the day the new components are added.)
3. For each IBD Component symbol, divide the dividend in step 2 by that stocks closing price on the day the new Component Listings are published.  The quotient of each for each division is the weighting multiple for each respective symbol.  Those are the weightings for that weeks composition.
4. For each trading day of the following week, the index is the sum of the products of each component stock's close times it's weighting.
5. When the new IBD Component Listing is published, (go to step 2) re-balance the composition and weightings.

Now comes the fun part!  In the coming posts, I'll begin to analyze the index action and see if I can clarify just what the "leaders" of the market are telling us.  I'd welcome any comments or suggestions.

The index started out at 6257.75 on Monday, July 16, 2006.  It ended at 5776.87 on Friday, July 27, 2007, for a decline of 7.7%.  For the same period, the S&P 500 cash index declined 5.8%.  As I've previously noted, we should expect to see higher volatility from stocks which are high beta by design. Below, I've plotted the index against the S&P 500 for the same time period. 

I'll be updating and posting this graph regularly.  It's interesting to note that the IBD index made a new high on July 23, while the broader average had a lower peak.  In this tiny sample, not only did the IBD average fail to act as a leading indicator.  It notched a false breakout, only to exhibit a greater percentage decline than the S&P in the days immediately following.  We don't have enough data to draw any conclusions yet, but perhaps the IBD's lack of confirmation by the S&P will prove to be significant.

Blocks vs. Wealth-Lab

I use Wealth-Lab both Wealth-Lab, and Telechart. I think Telechart is a great value for the dollar, Although  I know it's sometimes referred to as the poor man's TradeStation.  (Not only am I not really famous, I'm not rich either. But I'm working on both.)

My equal weight IBD 100 index needs to be rebalanced weekly.  Index Builder in wealth lab has no built in way to accomplish this.  It will calculate the entire range of index values for the current components and weightings.  I need to access the composition and weighting for each week as I calculate the index, so It requires some external piece of code. 

To calculate the index in Wealth-Lab, it seemed that my only option was to write a COM object which accessed the components externally (presumably from a database or the file system).  This sounded like an incredible pain in the butt.  I opted to calculate the index in a completely separate program, then write it's values to a file, which will serve as an ascii based DataSource for Wealth-Lab.

I'm using TeleChart as my data source since the  data access SDK is dead easy to work with.  And I've implemented my own database to store the  price data imported from TeleChart, and the IBD Component Data which I import from downloaded IBD 100 spreadsheets .  I should have all the initial coding completed this weekend.

Anyway...that's all just background.  I really wanted to use this post to solicit opinions.  I use Wealth-Lab for back testing.  It has a lot of nice built in functions, and I can use Fidelity as a free source of intra-day data.  But I've got to say I HATE WEALTHSCRIPT! It's an ugly, pascal based scripting language. If-then-begin-end-else-begin-end-end-;  It's so old fashioned.  The placement of statement terminators is kind of arbitrary (sure it follows it's own rules, but the rules are arbitrary).  And some of the keywords, like var and Type are just retarded.  If they wanted a scripted environment, why didn't they choose javascript?

I can't help it.  I was a C++ programmer, who became a Java programmer, who became a C# programmer. I prefer the syntax of the C variants.  I understand that Blocks can be used for backtesting, and that it's .NET compliant. So I'm thinking about trying Has anybody worked with Blocks? (Does anybody read my blog?).

I'd love to hear some comments about working with Blocks.  If anybody has worked with both Wealth-Lab and Blocks, I'd love to hear opinions about their relative merits.  Thanks in advance for the input!

It's officially a correction

IBD has officially called a "correction".  They've dropped their distribution days count.  IBD cautions to never buy anything during a correction.  We are told to wait first for an upside breakout, then a follow through occurring at least 4 days later.  Of course, the IBD 100 portfolio and IBD 100 Equal Weight Index will ignore this.  At some point, I'll run some back tests on CANSLIM buy rules.  First, I have to complete the index coding.  (Sorry it is taking so long.  I have another job.)

After yesterday's sell-off, the value of the IBD equity portfolio is $95763.72, down 4.2% since inception, down .6% yesterday.  Since inception, the S&P 500 is down 4.5%. Interesting....the "leaders" have actually held up a little better than the broader market. 

S&P 500 down 51.78!

Wow.....what a vicious sell off!  Days like this always put a knot in my stomach.  Here's a harrowing admission: I was short OEX puts on Black Monday in 1987!  I learned a painful lesson about selling naked premium and have never done it since.

I wish I could say I was short or even flat going into today, but I'm not.  I'm roughly 85% long, with the rest in cash.  I bought EMC yesterday at 19.05, and bought some more today at 18.50.  Very nerve wracking.

As of yesterday's close, the IBD 100 Portolio had lost 3.6% since it's inception.  Not much change from the Close of the 24th.  I expect it will be sharply lower at today's close.  Once I get the index coded, it will be interesting to see if the price action of the 1BD 100 stocks gave any forewarning  of this  market decline.

IBD 100 Portfolio update

I'm still working on the equal weighted index.  I expect to start posting that soon.  In the mean time, as of the close of July 24, 2007, my portfolio of IBD 100 stocks has lost 3.6 percent since it's inception on July 13, 2007. In that same time period the S&P 500 declined by 2.7%

That the IBD 100 portfolio would decline more than the S&P 500 isn't surprising to me.  I'd expect the portfolio to have a high beta.  Still, that they would "lead" on the way down wasn't entirely a foregone conclusion.  What I'm more interested in is whether the IBD 100 stocks were in anyway predictive of the current broader market decline.  I'll be exploring this once I've complete the coding for my equal weighted index.

I have a lot of historical stock, end of day stock data, but I don't have the historical composition of the IBD 100.  I asked IBD for this information, but their support team declined, saying that they don't currently offer this data, but will consider it as a suggestion. Once I complete the index coding, I'll pester them again.  I live in NYC, so maybe I'll actually go to William O'Neil's offices. Without the historical composition data, it will be a while before my research can produce any statistically significant results. 

How to beat Google at Search....be Amazon

Great numbers from Amazon last night.  The obvious takeaway is that retail is still strong, although shoppers will be doing more of their shopping on the Web.  But notice that amazon's resurgence is coming on the heals of a dramatic shift in the way amazon does business.  Have you bought anything on Amazon lately?  Much of what Amazon sells is not actually sold by Amazon. Or should I say, it is not actually fulfilled by Amazon.

Amazon "sells" merchandise which is fulfilled by other stores (Amazon calls the "sellers").  Not long ago, I bought a soccer jersey on Amazon (I'm a big AC Milano fan, despite the scandals. Go Rosso e Neri!).  The jersey I got came from WorldSoccerShop.com. World Soccer Shop maintained the inventory.  World Soccer Shop fulfilled my order. 

There's nothing new about this.  Book sellers have been doing this with print advertisements for years. I believe it's called drop shipping.  You present yourself as a store, but really you are just selling someone else's inventory.

What's different about Amazon is that they are now a drop shipping department store.  How many different items does Amazon offer?  It must be thousands.  But Amazon provides an easy way to find what you are looking for.  You either browse their directory links, or you.....here it comes.....SEARCH for the items you want.

That's right, in some senses, Amazon can be thought of as a retail search engine.  Sure, their revenue stream comes from retail margins rather than advertising, but I think the analogy still holds.  You could say that their revenues come from pay per purchase advertising.  They do order tracking and payment processing as well, but I think these should be regarded as artifacts of their niche search business.

I believe that niche search...like Amazon....is the way to challenge Google's search dominance.  Why is Google so dominant?  I really don't know that much about the comparative merits of online advertising programs (Although I intend to learn, since I'd like to earn some ad revenues from this blog. If you don't mind, please click on one of my sponsors).  Does AdSense place text ads on pages that are more likely to drive clicks than other ad programs?  Does google have a better price structure than other ad programs?  Or is it better to place ads on Google simply because Google gets a bigger share of search?

As I said, I don't know the answers to these questions, but I suspect that the latter is the most significant factor.  This begs the question, does Google get a bigger search share because it's better at searching?  Maybe....for generalized searches.  I heard that there was a study done where search results from different engines where re-formatted to look like the other's results (so Yahoo search results were reformatted to look like Google's).  Users were then asked to rank the results of each engine.  As I understand it, they overwhelmingly favored the results that looked like the came from Google, even if those results came from Yahoo or Ask.

This suggests that people prefer Google because they're use to preferring Google.  This sort of dominance through momentum shouldn't be minimized.  It can be sustainable.  Consider VHS v. Beta (if you're old enough to remember).  These two weren't even equivalent.  Beta was empirically better than VHS.  But VHS won, at first because Sony dropped the marketing ball, but ultimately, because VHS gathered irresistible momentum.

So VHS used momentum to defeat the format that competed with it on it's own terms, but ultimately failed when the playing field itself was re-defined.  Right now, the search playing field defines a game of generalized search.  A search engine is a tool that searches for anything on the internet....web pages, blogs, news groups, images, video.....everything.  In this game, I think Google has won.

But the game will change.  As a niche retail search engine, Amazon is simply a better mousetrap.  I know, Google has some kind of Google store, or Google auction, but it's not as pleasant a shopping (or should I say SEARCHING) experience.

The way to be a better search engine than Google is to focus on niche search.  I think it's almost impossibly hard to beat Google at generalized search.  But search engines which focus on a niche can offer a better service.

I think there's more to this, and I'll be posting about it more.

Getting Started

The main focus of this blog is a place where I can share my trading ideas and research.  I hope that posting to a blog will impose a writer's discipline on my trading (see Hemmingway's 'A Moveable Feast'.  As I recall, there is a chapter titled 'Writing is good discipline').  Frankly, my trading has been undisciplined for years, and this blog is part of my commitment to change that.

The first topic I want to explore is the Investor's Business Daily 100.  I'm an online subscriber to IBD, and would recommend it to anyone interested in stock investing/trading.  I've read William O'Neil's books, and feel that they've improved my stock trading performance.  But here's the problem, I feel like O'Neil's books have helped me, but I haven't tested it systematically. 

Well...perhaps I can say that the improved performance is something stronger than a feeling.  Before reading O'Neil, I liked to trade stocks that broke out on above average volume.  (I'll go into the exact parameters in later posts).  I used to TeleChart to screen for them these breakouts and sort them into watchlists.  The watchlists performed fairly well, but the criteria generated too many hits.  So my process for selecting from the watchlists was sort of ad hoc.  The performance of my selections consistently lagged the broader watchlists.

Then I read O'Neil, and started applying some CANSLIM criteria to my filtering.  My results began to improve.  I think as a general statement, one can say that CANSLIM is a hybrid methodology using technical and fundamental analysis to select growth stocks.  Intuitively, if you believe in investing in or trading growth stocks, then O'Neil's methodologies ought to be helpful.

My performance did improve when I started applying some O'Neil methodologies.  But honestly, I didn't apply them systematically.  It could be luck.  What's more, I didn't start doing so fabulously well that I felt I'd found the holy grail.  So, I have to say that my gut feeling is that O'Neil methodologies work.  But I have no quantitative evidence to support this.  Furthermore, I couldn't say which parts of it worked, and how well.

So, I'm committing myself to a more systematic approach to evaluating O'Neil methodologies.  I'll be re-reading O'Neil's books, and presenting my ideas about what and how to test.  But my recollection is that O'Neil and IBD claim that their methodologies were born of extensive research, but they don't present their data.  Nor do they describe applications of their methodologies in ways that are testable.  They are presented more as heuristic frameworks, than as testable, mechanical rules.

I'm not one of those people who is committed to mechanical systems (although I may become one, depending on how this project goes).  But if you can only back test criteria that are algorithmic.  I'll be re-reading O'Neil's books and CANSLIM articles in IBD, to come up with ways to test their methodologies.

My first area of research will be the IBD 100 stocks.  Each week, IBD publishes a list of 100 stocks which it refers to as the market "leaders".  IBD recommends watching the leading stocks to gauge the market's health.  In general, IBD recommends buying the leading stocks, at proper entry points, when the market is rallying. 

But what exactly do they mean by leaders?  Is it just that these are stocks which will do better in a bull market, or is there something actually predictive their price action?  My gut feeling is that both are true to some extent.  I fell that in my own trading, when I've owned IBD 100 stocks in my portfolio, and those stocks started to perform badly, the broader market tended to follow.  But is that really the case, or is it just something I imagined? This is the first question I want to explore.

Starting last Friday, July 13, 2007 (good thing I'm not superstitious), I created a model portfolio of the IBD 100 stocks, with an initial balance of $100,000.  I allocated $1000 dollars to each stock, and bought (on paper) as many whole shares of each stock as I could for $1000.  The remainder stayed in cash.  When the new IBD 100 list came out on Friday, July 20, I sold the 13 stocks which were dropped from the list, divided my cash balance by 13, and bought that much of each of the 13 new additions.

Investor's Business Daily does not recommend doing this in real life.  I don't remember when, but I'm pretty sure they've explicitly pointed this out as something not to do.  In my next post, I'll explain why IBD says you shouldn't do this with real money, and some of the reasons why I think it may not be a good idea in practice.  (But who knows, maybe it will actually turn out to be profitable anyway?  Keep checking back and we'll find out together).  I'll provide details of the performance of the portfolio. 

I'm also planning to create an equal weighted index of the IBD 100 components.  I'm in the process of writing the code for this now (It's surprisingly tedious code.  It's not complicated, but there are a lot of little, time consuming methods to write).

So, look forward to portfolio and index results.  I'll also be working on improving the look and feel of the blog.