Objective research to aid investing decisions

Value Investing Strategy (Strategy Overview)

Allocations for April 2024 (Final)
Cash TLT LQD SPY

Momentum Investing Strategy (Strategy Overview)

Allocations for April 2024 (Final)
1st ETF 2nd ETF 3rd ETF

Review of IntelligentValue’s Retracement-Value Portfolio

| | Posted in: Technical Trading

A reader poses the following question: “Have you looked at IntelligentValue? They claim pretty impressive results, apparently certified by FinancialContent. Their Retracement-Value portfolio, particularly, shows impressive results [“508% in its 1st 18.5 months!”]. Are you able to evaluate this newsletter service?” Based on the information provided on IntelligentValue’s web site, especially the closed trade analysis for the [now removed] Retracement-Value portfolio (171 round-trip trades, apparently from portfolio inception on 5/16/06 through 1/9/08, with a starting portfolio value of $10,000), we conclude that:

For a detailed review, we examine the Retracement-Value portfolio closed trades and accompanying assumptions. Key assumptions/process features [details now removed] are:

  • Every Sunday evening, IntelligentValue transmits “a newsletter that provides market commentary and analysis,” including “exactly which stocks to sell and which to buy and in exactly what quantities.”
  • Apparently, the Retracement-Value portfolio closed trade analysis assumes that the prices for “both the purchased and sold stocks” are “the average of the high and low of the stocks” the next trading day.
  • “The closed positions” in the trade analysis “DO NOT include the currently held stocks, which would change the results dramatically.”
  • “Total returns are price-related only, and the effects of dividend reinvestment and advisory and trading costs are not included.”

Next, we take the closed trade analysis at face value, with two minor corrections: purchase date for BDY 1/3/07 rather than 1/3/06, and purchase date for UCTT 4/30/07 rather than 4/30/17. Over the course of the 171 closed trades listed, this portfolio generates an impressive total profit of $44,119. The average (median) profit per trade is +5.2% (+1.7%), and 64% of all trades are profitable. Average trade duration is 15 calendar days.

Looking only at the 67 closed trades since “Public Introduction of Portfolio” on 7/3/07, the portfolio generates a total profit of $11,142. The average (median) profit per trade is +1.7% (-1.5%), and only 48% of all trades are profitable. Average trade duration is eight calendar days. For detail on the apparent deterioration in results after public introduction, we show in the following chart the average as-stated (raw) percentage profit per trade by calendar quarter for all 171 completed trades. The chart shows a marked decrease in average trade profitability starting in the second quarter of 2007. The fourth quarter of 2007 is roughly breakeven.

Presumably, trades since public introduction are out-of-sample with regard to development of trading rules and therefore free of data mining bias (good luck for trading rules identified during the prior in-sample period).

Where do the profits come from?

The following scatter plot relates individual trade profitability to per-share purchase price for all 171 completed trades. A best-fit line slopes downward, suggesting that trades in low-priced stocks are more profitable than trades in higher-priced stocks.

For a closer look, we calculate the average profitability of closed trades by range of stock purchase price.

The next chart shows average closed trade profitability for various ranges of stock purchase price. It shows that trades of stocks below a $10 share price tend to be much more profitable than trades of higher-priced shares. In other words, low-priced stocks drive portfolio profitability.

Since “Public Introduction of Portfolio,” the average stock purchase price for the Retracement-Value portfolio closed trades is $46, compared to $30 for the overall record (and $20 before public introduction). The tendency to recommend higher-price stocks since public introduction (perhaps due to market conditions) may to some degree explain declining profitability.

Next we examine the assumptions used in the raw analysis and make some adjustments.

The adjustments we make are as follows:

  • We subtract round-trip trading fees, ranging from $10 to $20 for sensitivity testing, from the net proceeds of each trade.
  • Based on the average number of trades per month and a $40/month subscription, we subtract $5 from the net proceeds of each trade.
  • We add a percentage to the purchase price and subtract a percentage from the sales price, ranging from 0.2% to 1% for sensitivity testing, for each trade to reflect the bid-ask spread and impact of trading.

The last adjustment potentially carries the greatest force. It addresses optimism in IntelligentValue’s assumption that midpoints between daily highs and daily lows on specified trading days reasonably approximate both the buy and sell prices for the portfolio. Because of the bid/ask spread and the impact of trading, the distribution of daily transaction prices for traders initiating buys differs from that for traders initiating sells. For example, a trader initiating a buy likely produces the daily high, and a trader initiating a sell likely produces the daily low. Moreover, when transactions are substantial compared to normal trading volumes, large buyers (sellers) or groups of buyers (sellers) drive share prices up (down) during execution. The more subscribers to the Retracement-Value portfolio, the greater the impact of associated trading on target stock prices. Summing up, the mean and median daily prices of a stock tend to be higher for those initiating buys than for those initiating sells.

This asymmetry between buyers and sellers is largest for illiquid stocks, disproportionately found among small companies and low-priced shares (which drive Retracement-Value portfolio profitability). The degree of asymmetry tends to grow when stock price volatility is high, as would arguably be the case for Retracement-Value portfolio trades.

Given the short average duration of Retracement-Value portfolio closed trades (one or two weeks), we assume that accrued dividends are negligible.

The following chart depicts the distributions of trade profitability after applying high-end adjustments ($20 round-trip trading fee and 1% each-way liquidity penalty) to the Retracement-Value portfolio closed trades. Adjusted portfolio performance (as for the unadjusted case) depends critically on a few extremely profitable trades, probably for relatively low-priced stocks, in the long right-hand tails of the distributions. For example, missing the best four (12) trades in the adjusted overall distribution would reduce average trade performance from +2.7% to +1.3% (0.0%). Missing the top five trades since “Public Introduction of Portfolio” would reduce recent average trade performance from -0.6% to -2.7%.

The following table summarizes profitability metrics for Retracement-Value portfolio closed trades during 5/16/06-1/9/08 for four scenarios:

  1. No trading fees/frictions
  2. Trading fees $10 per round-trip trade, IntelligentValue subscription fee $5 per round-trip trade, liquidity penalty (bid-ask spread and impact of trading) 0.2% each way on each trade
  3. Trading fees $15 per round-trip trade, IntelligentValue subscription fee $5 per round-trip trade, liquidity penalty (bid-ask spread and impact of trading) 0.6% each way on each trade
  4. Trading fees $20 per round-trip trade, IntelligentValue subscription fee $5 per round-trip trade, liquidity penalty (bid-ask spread and impact of trading) 1.0% each way on each trade

Scenario 3 is roughly breakeven since public introduction. Scenario 4 is unprofitable since public introduction. Note that open positions (five to ten?) marked to market could affect these results positively or negatively. There may be a tendency to hold unprofitable trades open hoping for recovery.

In summary, the extraordinary performance claimed by IntelligentValue for the Retracement-Value portfolio is likely due to a combination of early good luck and a favorable assumption about trade prices. Investors/traders considering such active trading services should carefully evaluate their trading costs/frictions.



Christopher Michaels, Publisher of IntelligentValue.com, sent the following comments on the above review:

General opening comment: “First let me say that after a thorough review of your site, it appears that you have set yourself up as an expert to take potshots and low blows at legitimate advisories. Admittedly, many advisories aren’t worth the paper (or air) they are written on. However, some make very concerted efforts to help subscribers, enhance wealth, build retirement funds, and finance college educations.”

General response: The intent of CXOadvisory.com is to offer skeptical analyses, not potshots and low blows. These analyses sometimes involve making judgments regarding information that is not clear or not available. The judgments may be imperfect, but they always include assumptions so that readers with different judgments can interpret results accordingly.

Specific Comment 1: “First of all, you do an analysis of my record, stating that ‘The average (median) profit per trade is +5.2% (+1.7%), and 64% of all trades are profitable. Average trade duration is 15 calendar days.’ This record of 64% winners places IntelligentValue at position #5 in your ‘Guru Grades’ section. Please add my newsletter in this position as soon as possible. If you don’t, we’ll know what that list is about…

Specific response 1: Guru Grades is a study of the ability of experts to forecast the behavior of the broad U.S. stock market. It is not a study of stock picking expertise. The intent of the section is to convey a sense of how much investors/traders should let such broad forecasts influence trading decisions. As noted there: “This kind of forecasting ability is different from, but may be related to, stock picking expertise.” To keep the Guru Grades reviews comparable, it focuses on broad market forecasts.

Specific Comment 2: “All of my trades are documented by an independent, publicly traded (FCON) third party organization that provides financial data services to many well known companies, such as CNET, San Francisco Chronicle, Denver Post, CBS News, ABC News, etc., etc. They are contracted to maintain all prices and issue Trailing Stop Loss (or Profit) when necessary on stock that reach a TSL(P). This organization can be contacted for verification of all trades.

Specific response 2: The review does not challenge the data as presented, except for a couple of obvious date errors.

Specific Comment 3: “You state: ‘We know of no research that reliably predicts financial markets using Fibonacci analysis. Rather, its highly judgmental application to financial time series is based on loose analogy with observations from other fields.’ This information is an incorrect depiction of the character of the portfolio. The Fibonacci numbers play a very tiny role in the incredible success of our portfolio. It was a marketing decision to name the portfolio ‘Fibonacci-Value’ because of the mysterious aspect of those numbers. When creating the scientifically designed ranking system and buy and sell rules for the portfolio, I found that by slightly changing some key numbers from .50 to a Fibonacci number, i.e., .618, etc., I was able to achieve slightly better returns. Since reading your article, I have changed the name of the portfolio to ‘Retracement-Value.’ Anyone interested can read the story of the development of this breakthrough portfolio.”

Specific response 3: The name “Fibonacci-Value” was your choice for marketing reasons. We changed “Fibonacci-Value” to “Retracement-Value” in the above review per your renaming. We have also deleted the comment on Fibonacci analysis since it is such a slight aspect of your model. The slight enhancement you found from using Fibonacci ratios may be an artifact of data mining that does not survive in out-of-sample use.

Specific Comment 4: “You state: ‘If anyone could reliably generate returns such as those claimed by IntelligentValue, why would they spend time/resources (and perhaps exhaust their own specific investment opportunities) by offering their critical information to the public via nominally-priced monthly subscription?’ I have stated in the marketing materials very clearly that I will only allow a small number of subscribers to participate in the IntelligentValue newsletter and this portfolio. I consider this an alternative to a hedge fund, but I only receive a small fee compared to the high fees charged by a hedge fund. Your statement rings of jealousy and envy. My personal plan is to fold the profits of the newsletter into my monthly contributions to my own portfolio of these stocks. With the magic of compounding, when I add an extra $5000-$8,000/month of the profits to my Retracement-Value-based portfolio, my wealth grows at an enormous rate. An investment of $10k over 10 years at 200% becomes $590 million. With a monthly addition of $5,000, that same $10k becomes $5.9 billion. Surely you are smart enough to figure that out for yourself. At the same time, I am providing wealth to those individuals that follow the portfolio. The reason I sell the newsletter is twofold – to make myself rich and at the same time make subscribers rich. I am not charging some outrageous fee; it’s only $39/month. In the near future, I will remove the subscription page and allow subscribers to leave at their own pace. I won’t force anyone out. In the meantime, I am limiting the number of subscribers to this portfolio to 400. It is my intention to retire when I reach my financial goals, but I will not leave subscribers high and dry. I am setting up an autotrading system that remaining subscribers will be able to participate in.”

Specific response 4: Since you have answered it here, we have removed the question from the review above.

Specific comment 5: “The portfolio has a liquidity minimum of price X avg. volume (20 days) > $1,000,000. This guarantees that investors are able to get into and out of these positions without problem. You state that, ‘We add 1% to the purchase price and subtract 1% from the sales price for each trade to reflect the bid-ask spread and impact of trading.’ Based on my experience, subtracting 2% from every round-trip trade that a portfolio makes essentially destroys any reason to invest at all. You state that I have had 171 round-trip trades, which would subtract 342% from the profits of the portfolio. I see this claim on no other newsletter you have evaluated (I admit I didn’t look at that many as it was easy to get the gist of your effort to take down each one), but to subtract 2% from every round-trip trade, after I go to all the trouble to adjust prices to the average of the hi and low for the day (the standard used by the Hulbert Financial Digest, the ‘Consumer Reports’ of newsletters) for all new sales and purchases, is outrageous. It appears to simply by your last, desperate effort to eliminate create an arbitrary 1% ‘fee’ that wipes out any returns that my portfolio shows. There are no ‘trading costs’ such as 2% associated with investing. If there were, it would behoove all investors to avoid the stock market like the plague. In 15 trades a year an investor would be set back 30%! This is simply asinine and you know it.”

Specific response 5: Trading frictions are far more important for frequent trading than for occasional timing or buy and hold investing. Our discussion of the difficulty of systematically buying or selling at the daily high-low midpoint stands. Moreover, there may be a significant impact to bid-ask spread and price when your 400 clients are all buying or selling at the same time. For example, 400 clients each attempting to buy or sell $10,000 of a stock at the same time produce a volume spike of $4,000,000. Unless the stock’s historical daily trading volume is much, much larger than $4,000,000, this coordinated buying (selling) activity is likely to drive both bid-ask spread up and price noticeably higher (lower). Your minimum average daily liquidity requirement of $1,000,000 in trading volume is not adequate to allay this concern. However, to address different scenarios, we have added a range of market frictions so that readers can compare them with their personal trading experiences.

Specific comment 6: “You state: ‘This asymmetry between buyers and sellers is largest for illiquid stocks, disproportionately found among small companies and low-priced shares (which drive Fibonacci-Value portfolio profitability).’ As I mentioned, the portfolio uses a buy rule of price X volume (20 day) > $1,000,000. There are no Over-The-Counter stocks or other illiquid stocks. Please list the stocks in the closed positions that are illiquid. There are none that I know of. I had no problem buying or selling any of the stocks listed within 1-5 seconds at the market price.

Specific response 6: See the response to the preceding comment.

Specific comment 7: “Finally, you take careful effort to push the results since July of 2007. If I may remind you, this is when the stock market began what I consider to be a bear market. I warn investors in the ‘Retracement-Value’ portfolio is designed to work during upward market environments, which occur more than 2/3rds of the time since 1880. When our timing indicators show that we are headed towards a significant downturn, we go to cash, or implement magnified, inverse ETFs in the portfolio. The may account for the change in prices of stocks during the period of August 2007-present. I know you have many articles on your site written by people who have failed to time the market, but where are the ‘balancing’ articles you hold in such high regard. As evidence that the market can be timed, I present the graph below. Which would you rather have, my graph or the S&P500 index? Even with the market in bear mode, the ‘Retracement-Value’ portfolio has beaten the market by about 53-55% as of yesterday. The graph below is a real-time graph that shows the returns since late July 2007, showing that the portfolio, although not gaining the 150% it might have gained in the same period the first year, it did gain about 50% at a time when the market was declining and the economy is in, ostensibly, a recession. As you can see, the ‘Retracement-Value’ portfolio beat the market by 53-55%.”

Specific response 7: The focus on your results since 7/3/07 (“Public Introduction of Portfolio”) is not driven by any broad stock market trends but by an assumption that these results are out-of-sample with respect to development of portfolio trading rules. Out-of-sample results are not subject to data mining bias (lucky parameter selection) and are therefore a better indicator of actual portfolio performance. With regard to your chart, if the underlying trade data excludes trading costs/frictions, it may not be representative of actual results.

General closing comment: “I won’t go on with more errors you have made. It is clear to me from the massive amount of articles on your site that you are dedicated to the fact that an investor cannot make money in equities. You are doing a great disservice to many, many people and are misusing your education by creating a site intended to prove what you consider to be your life’s ‘point’ that the market can’t be beat. Give me a break – how many times is Fama going to append his ‘Efficient Market Hypothesis’ until he finally admits on his deathbed that it was a big mistake. Warren Buffett is the most famous living evidence that the market can be beat and beat big ($50 Billion).”

General closing response: CXOadvisory.com is not “dedicated to the fact that an investor cannot make money in equities.” The work of many academics and practitioners supports a belief in financial market inefficiencies that investors can reliably exploit. This work also supports a belief that much of the variation in financial market returns and investing results is due to randomness (luck).

Suggestion: “I suggest you re-write the article as soon as possible. It would be difficult for you to uphold this slander in court.”

Response to suggestion: See the responses above.

Login
Daily Email Updates
Filter Research
  • Research Categories (select one or more)