Famous Trader

IBD 100 Index Update

Sorry this post was delayed.  I was on vacation last week.

The IBD-EQ index swung wildly last week.  While the S&P 500 was down about a half a percent, the IBD 100 index was down over 5%!  This is the flip side of the IBD's outsized gains between 8/3 and 8/5.  Since inception, the IBD 100 index is down 9.48%, compared to the S&P's decline of 6.68%.  Here are the comparison charts

The closing changes are summarized below.  Notice how on the IBD 100 was down last Thursday's big reversal day, while the S&P was up slightly.  Already, we're seeing that momentum stock index has a different trading character than the broad index.  This would seem obvious, but the nature of that difference may be exploitable.

Changes for the week:

                                         

	7/16/2007		8/10/2007		8/16/2007			8/17/2007	Week % Change		Change since 7/13
spx	1549.52		1453.64		1411.27			1445.94	-0.53%		-6.68%
ibd	6226.13		5941.44		5456.75			5635.22	-5.15%		-9.49%

Responding to a reader's questions.

In the comments for an older post, a reader named Michael asked:

"Did you start your porfolio on July 13th as I can see on your chart? Do you use the big picture to manage your cash level in your porfolio: the "M" in CANSLIM? Do you think it could be a good idea to sell everything when the market is in a down trend? Waiting for your comments!"

I started both the index and the portfolio on July 13th.  I must apologize about the portfolio.  I haven't updated it this week. A bit of background.  I calculate the index myself.  I use fidelity data feed for price history and the weekly downloaded excel spreadsheet of the ibd 100 for composition.  I wrote a computer program to import the data, re-balance the weightings, and update the index values.  When you see the graphs of the index generated by Wealth-lab, it is Wealth-lab using a text file with my calculated index values as a data source.

For the portfolio, I was manually entering the trades into the Wealth-lab portfolio manager.  But last week, I realized that Wealth-lab was not calculating the available cash correctly after I "sold" shares from the portfolio (remember, this is just a paper portfolio).  This meant that my portfolio weightings were off. 

So, when I get some free time next week, I'll write a program to re-construct the portfolio.  I'll deduct commission as well.  Remember, the weightings for the portfolio are not the same as the weightings for the index.  The index is equal weighted, allows for fractional shares, and is re-balanced for equal weighting each week.  The portfolio should begin $1000 dollars alloted to each component, but only whole numbers of shares can be purchased.  Rather than re-balancing every week, I will sell the shares which are dropped, and use the cash proceeds to purchase the newly added components in equal dollar allotments.

I began tracking the IBD 100 on July 13th, but we have to remember that the new IBD 100 list is published after the close on Friday.  So the first close of the Index is Monday, July 16th.

I read The Big Picture every day, but I am not using it in either the index or the portfolio.  I guess I wanted to test the claim that the IBD 100 "trounces" the S&P 500.  I can't remember where I first read this, but if you google it, you'll find this quote frequently. 

I was intrigued in part because it's an odd claim to make.  Until I came along, the IBD 100 was not an index.  It was just a list of stocks which changed weekly.  The S&P 500 attempts to be representative of the broad market, while the IBD 100 is representative of high relative strength stocks, ranked by a proprietary algorithm.

If you look at the IBD 100 components, it is certainly astounding how many high percentage gainers there week after week.  But this sort of at a glance analysis is strongly skewed by survivorship bias.  The worst performing stocks in the list are dropped weekly.

So the index was created to try to compare the IBD 100 list to the S&P in a reasonable manner.  While I was at it, I decided I wanted to see if trading the IBD 100 list in a very naive way might actually be profitable. To be clear, IBD and William O'Neil's books all say not to do what I'm doing with the portfolio.
In fact...I'm not doing it with real money.  It's a paper portfolio.

But I wanted to see what would happen anyway.  In part it comes from the sense of frustration that the best performing stocks in the IBD 100 are often (and by definition) the most extended.  Mark Fisher, used to say that "the best trades are the hardest to take".  We see this all the time with the IBD 100.  The best performers don't give you a many chances to get in, or when they do, you feel like you're catching a falling knife.

Anyway...that's what I was thinking about with the portfolio. 

IBD 100 Index Update

It was a remarkable week for the IBD 100 index!  For the 5 day period, the IBD index outperformed the S&P 500, gaining 3.8% compared to the S&P's 1.4% gain.  But more notable was the action on Wednesday, August 8th.  Wednesday was the weekly high for both indexes. The S&P was 4.5% mid week, but the IBD-EQ was up a wopping 9%!  Wednesday's 6241.23 close was the highest the index recorded since I began tracking back on July 16. The changes are summarized below:

                                   

	Close 8/3/2007	Close 8/10/2007	5 Day Change	Close 8/8/2007	3 Day Change
IBD 100 Index	5723.67	5941.23	3.80%	6241.23	9.00%
S&P 500 Index	1433.06	1453.64	1.40%	1497.49	4.50%

Here are the comparison charts:

James Altucher's explanation of hedge fund morass

For the past six weeks, I've been a bit perplexed by the sub-prime crisis.  The default rate called into question the high ratings the agencies were giving to CDOs, but the default rates are not out of line with what one would expect from low grade mortgage paper.  Basically, it felt bad, but not surprisingly so.  Today, in TheStreet.Com RealMoney Columnist Coversation section, James Altucher posted a very concise explanation of the mechanics involved in the breakdown.

For those who don't subscribe, I've included the text below.  BTW -- I highly recommend subscribing to RealMoney.  It's very reasonably priced, and well worth every penny.

	James Altucher
	Can we please put everything in context?
	8/9/2007 9:32 AM EDT

Let me explain EXACTLY what is happening to these hedge funds. The "holy grail" for hedge funds is to return 1% a month with no volatility. The saying is, "if you can return 1% a month you can raise $1 billion" and thats a true statement. But its impossible to get 1% a month without enormous risk. Here's the trade that hedge funds started doing about 2-3 years ago to get that 1% a month. They would borrow money at the short end of the curve and they would then buy (lend) subprime paper. So they would borrow at 5.5% and lend as low as 6.5%. Only 100 basis points difference because that trade got VERY crowded. That was the problem. The problem is not that there are defaults. The defaults ARE exactly what one would expect from subprime people who never borrowed before. The problem was that that paper was only yielding 6.5%

But wait, why is that a good trade? Only a 1% spread per year. Well, leverage it up 12 to 1 and you make your 1% a month. And THATS the problem. An 8% decline in subprime paper and your fund is wiped out. WIPED. OUT.

Lets look at the rest of the news this morning, which is arguably more important to us than the few idiot hedge funds leveraged up 12: 1 to get the holy grail. WMT, TGT, BBY, all coming in with nice same store sales. Go Team!

The market is going to get cheap today. And people should start buying.

Position: none

IBD 100 Composition Update

After another volatile week, 20 stocks were replaced in the IBD 100.

Stocks added 8/3/2007

                                                                                       

(CRNT)	CERAGON   NETWORKS LTD
(VSEC)	V S E CORP
(DWSN)	DAWSON GEOPHYSICAL CO
(UA)	UNDER ARMOUR INC CL A
(FCSX)	F C STONE GROUP INC
(MTW)	MANITOWOC CO
(CTRP)	CTRIP.COM INTL LTD ADR
(DLB)	DOLBY LABORATORIES INC
(FTI)	F M C TECHNOLOGIES INC
(TRMB)	TRIMBLE NAVIGATION LTD
(HMSY)	H M S HOLDINGS CORP
(TKC)	TURKCELL ILET HIZME ADR
(MBT)	MOBILE TELESYSTEMS ADR
(EMC)	E M C CORP
(OII)	OCEANEERING INTL INC
(TTEC)	TELETECH HOLDINGS INC
(ESRX)	EXPRESS SCRIPTS INC
(ACM)	AECOM TECHNOLOGY CORP
(NVO)	NOVO NORDISK A
(SAM)	BOSTON BEER COMPANY

Stocks dropped 8/3/2007

                                                                                         

(RRST)	R R SAT GLOBAL   COMM NTWK
(GSOL)	GLOBAL SOURCES LIMITED
(DADE)	DADE BEHRING HOLDINGS
(AVT)	AVNET INC
(AXE)	ANIXTER INTERNATIONAL
(ARD)	ARENA RESOURCES INC
(CHE)	CHEMED CORPORATION
(TLVT)	TELVENT GIT S A
(ATN)	ATLAS ENERGY RESOURCES
(NE)	NOBLE CORP
(SII)	SMITH INTERNATIONAL
(PVG)	PENN VIRGINIA GP HLDGS
(SXC)	STANTEC INC
(TPX)	TEMPUR PEDIC INTL INC
(RIG)	TRANSOCEAN INC
(UEIC)	UNIVERSAL ELECTRONICS
(B)	BARNES GROUP INC
(FSRV)	FIRSTSERVICE CORP
(DECK)	DECKERS OUTDOOR CORP
(SIMO)	SILICON MOTION TECH ADS

IBD 100 Index Update

My apologies, I found 2 bugs in my index calculations.  One related to the first index calculation on 7/16/2007.  The second related to the handling of splits. I believe they are corrected now.  This week's results really surprised me.  Look at the comparison table

Notice that the IBD 100 index is UP for the week!  Because of my earlier bugs, I spent a lot of time double checking the calculations and I'm fairly confident that they are correct.  The day where the IBD-EQ visibly diverges from the SPX was Monday, July 30th.  On that day, the stocks of the IBD 100 index were up 4.1 % on (weighted) average. The SPX was up only 1.3% That day.

After the horrendous week we had in the broad market, I was genuinely shocked to find the IBD-EQ higher than the week before.  Here's the break down: for the week ending August 3rd, the IBD-EQ closed at 5723.67, up 1.4% for the week. The IBD-EQ is down 8.1% from July 16, 2007. (To fix the bug, I needed to calculate beginning the first trading  day after I started tracking the components). For the same period, the SPX is down 1.8% for the week, but down only 7.6% from July 17, 2007.

So while the high beta IBD-EQ is still down more than the broad market, it significantly out performed the SPX last week.  This is getting interesting!

Buying at the Close is Institutional

An old rule of thumb: volume on the open is retail, volume on the close is institutional.  The idea is that the  dumb money places market on open orders over night.  The smart money (and presumably, the big, institutional money) waits till the end of the day.

I think the first half of the rule no longer  holds entirely true. Mutual funds tend not to trade the open, but other big institutional players do.  Back in the 90s, I was a floor trader at the coffee, sugar, and cocoa exchange (now NYBOT).  In coffee, the managed futures funds didn't trade the open, but the commercials did. (commercials are growers, processors, and distributors, who are actually hedging risk, rather than speculating).  The commercials traded in size, and could not be thought of as dumb money.

Now, lot's of big players trade the open.  This, in no small part is related to obsolescence of the Opening Range Breakout trade in equities.  This system use to work in equities, but for the past few years, we've found that the majority of the range of trade tends to take place at the open.  If you're a short term trade of size, you have to trade the open.  That's where you get the requisite volatility and liquidity to make money.

However, I think the second part of the rule, that volume on the close is institutional still holds true.  The close is when, psychologically, the retail trader is spent.  But the mutual fund traders in particular, are focused on where they are marked to market after the close. This is just anecdotal, but thing should be evident: block trading on the close or otherwise is, by definition, institutional. 

For the past two days, we've seen big buying volume on the close.  Some institutions are betting big that (at least in the short term) the correction has run it's course.

More things I hate about Wealth-Lab

In an earlier post, I ranted about how I hate wealth-lab, and asked for input comparing it with Block, and or other options for back testing, custom indicator creation, custom indexing, etc...  Well, I've been working with wealth-lab some more, and now....I REALLY HATE WEALTH LAB!.

I'm writing a script to back test the Darvas Box strategy.  The strategy is explained in How I Made $2,000,000 In The Stock Market  , by Nick Darvas.At a high level, we can say that the algorithm for generating buy and sell signals is a state machine with 5 discreet states.  If you search online for the algorithm, this is the common description.  However, the transition rules for each state have two symmetric sub-states (reverse the order of evaluation, depending on the direction of the prior box break-out, if you're last breakout was up, you evaluate the states looking for a range high first, if last break-out was down, you evaluate for a range low first). 

So, to a human, the strategy can be described with 5 states, but for automated back testing, it takes 10 states.  But there's more. Remember that we use the previous box break-out to determine if we calculate for high-first or low-first.  But at the beginning of the price series you're testing against, there is no prior break-out, so there is a special case where we don't know whether to look first for a range high or low.  So, it's an FSA with 11 discreet states.

As an object oriented programmer, my natural inclination when faced with this challenge is to implement the State design pattern.  The Wealth Script manual says that Wealth Script is object oriented, supporting polymorphism.  Cool....if the language supports OOD, I can work with it.  I like to think that I'm language agnostic. But I noticed that in the sample scripts in the program, and the scripts you can view at Wealth-Lab.com, nobody uses classes.  All the code is procedural.

So I started to write a DarvasBox object, and tried to encapsulate the state logic in the State Design Pattern.  That's when I realized, once again, how lame wealth script is.  I tried to overload the constructor (using the stupid constructor method qualifier keyword, followed by the name "Create" which seems to be their constructor naming convention).  I got a compilation error saying that there is already a method named create.  Just to make sure you understand that I didn't miss the obvious, the two Create definitions have different parameter signatures.  Does wealth script actually support overloading?

Also, I tried to set a type reference to null, and got another compilation error.  Is it possible that this "object oriented language" actually doesn't support null assignments?

I really wish there were a better, alternative.  I took a look at Blocks, but it seemed too 4G-ish to me. You know....all visual drag and drop.  It looks like you can edit code after it is auto generated from the 4G designer, but I'd rather just write things myself.  I still haven't made my mind up on it.

Is there an IBD effect?

Does being added to the IBD 100 act as an upward catalyst for stocks? (If so, does this count as an 'N' in CANSLIM? ) Does being removed from the list act as a downward catalyst?  I'll be exploring these questions in upcoming posts.  Here is what the 13 stocks added on Friday, July 20th did on Monday, July 23.

                                                                                                                                                       

Stocks Added to IBD 100 on 7/20/2007
Symbol	Close 7/20	Close 7/23	Change	% Change
(AVT)	43.78	43.29	-$0.48	-1.10
(CAM)	79.14	79.29	$0.15	0.19
(FLIR)	49.00	49.65	$0.66	1.35
(GRP)	57.40	56.99	-$0.41	-0.71
(ISRG)	198.59	208.02	$9.44	4.75
(NVT)	48.38	57.33	$8.95	18.50
(OYOG)	84.60	87.58	$2.99	3.53
(PVG)	39.63	36.84	-$2.78	-7.01
(SCHN)	61.23	62.00	$0.78	1.27
(SII)	64.41	64.69	$0.28	0.43
(SLB)	96.68	96.31	-$0.36	-0.37
(SXC)	33.95	33.99	$0.04	0.12
(WHQ)	66.20	67.84	$1.65	2.49
		Average % Change		1.80

This weekend, I'll do the stocks added on the 27th.  I'll want to look at how these stocks do for the week as well as their first trading day on the list.  I'll compare that data with the results for all the stocks on the list, and the stocks which were replaced.

Does anybody have any suggestions on (or questions about ) how to approach this project?  I'd love to hear your input.