Skip to main content
Message Font: Serif | Sans-Serif
No. of Recommendations: 107
Making The Plunge, Is Mechanical Investing For Me?
Doc's Unofficial Guide to New Investor's-Part 2

(If you haven't read Part 1, please do. You can find it at )

Now that you have done your homework by reading the back posts, (And you did do your reading, didn't you?), no doubt your first question is,

Can it be this simple?"

Want the short answer? "Yes, Mechanical Investing can be simple."

But that's like an airline pilot answering "Yes, flying this 747 is simple."
Contrary to Hollywood hype, it's the rare flight attendant, who could step in and land that jet after the crew takes sick. Pilots study for hundreds and hundreds of hours before they are ready. And so probably will you.

If you are running annual screens, you can come in on a Friday, after that week's updates are posted, look up the stock picks, place your orders with your broker, and be back outside to flip the steaks on your barbeque before they burn. Getting the stock picks and buying them is simply the last link in a long chain, one that, once you understand mechanical investing can be done in a fifteen minutes, once a year. Sound familiar? The Foolish Four is a mechanical screen, just like PEG, Sparks or Keystone.

"But shouldn't I be focusing on buying a stock, and holding it for a long time? (Long Term Buy & Hold) Some of these screen want me to trade EVERY MONTH!"

Where main stream Motley Fooldom teaches you LTBH of stocks, mechanical investing is Long Term Buy & Hold (of screens). You're mechanical investing strategy should be to pick a screen you're comfortable with, and invest with that screen for the long term. Only by doing that, will you be protected from the ups and downs. Don't fall into a common new investor trap of chasing the "Hot Screen of the Month". That's a sure way to lose money.

You will find though, that most people who do mechanical investing ALSO hold LTBH stocks. Mechanical Investing is not meant to your entire portfolio, just another part of it.

"But shouldn't I look at the stocks? Check their fundamentals? Analysis them using the Rule Maker criteria?"

While I have the highest respect for the Gardner's, when it comes to deciding if a company's a good buy, most people, and I'm including myself right at the top of the list, couldn't do it if their retirement depended on it. (Wait a minute, your retirement does depend on it). Mechanical investing takes the guess work out of it. From the grumpy Master Rayvt himself,

"Time to repost my numero uno quote:
O'Shaughnessey, quoting from David Faust (The Limits of Scientific Reasoning):
Many tests & studies show that humans almost always show inferior predictions in the face of empirically derived actuarial formulas based only on a few attributes known to be associated with outcome. Intuitive forecasters are not only worse than optimal regression equations, they usually do worse than any simple formula. Intuitive forecasters don't lose to simple formulas because they apply invalid weights, but rather because they apply them unreliably--one time using a piece of data, another time rejecting it, depending on how they feel or the current circumstances."

OK, enough soapboxing for now. You and only you, can decide if mechanical investing has a place in your portfolio.

Making The Plunge
But you say, "Look at those returns! Of course I want that kind of profit! I'll be living on a tropical island before I'm thirty."

While I believe that mechanical investing has a place in everyone's portfolio, I also believe mechanical investing is not for everyone.

"Huh?" (heads scratched) "Ain't that a contradiction?"

Yes and No. :)

Two things will make mechanical investing unsuitable for you. Cost and Risk. One can be solved with patience, the other with honesty.

There's no way around it, its going to cost you money to make money. Commissions get you coming and going. Buy or sell at the wrong time of day, and the spread will take more than your commission. Do you hold for a year, or hold for less? And then there's Taxes! But you can keep your costs down by investing smart.

First off, read Moe Chernick's three part article on costs.
"Cost of Screens-Part 1"
"Cost of Screens-Part 2"
Part 3, "More on Costs"
And some questions at

"But look at my buddy, he has a RSOverlap that made him 100% gain this year, after he paid commissions and spread. With that kind of return, why should I worry about paying 10% on costs. Heck, even 15% is a deal."

The focus on costs, is not about what you made this year, but about what you MAY make next year. When we backtest a screen, we make the assumption that next year's returns will mimic the past. It may be alittle higher, alittle lower, but near the average. (the CAGR) We just don't know. But we do know our costs. Those are a constant, which we will pay whether the screen returns 10% or 100%.

If the postman stopped by tomorrow and gave you an envelope with $100 inside, would you tip him $20? I would. And if he came back every day next week, each day with another envelope containing $100. Would you tip him $20 each day? But what if you didn't know how much was in that envelope. One day $100, another day only $10. Would you tip him, without knowing? Suppose 4 days go by, each day with only $10 in the envelope. Wouldn't be long before you wanted another mailman. Or at least renegotiate that tip.

Costs are like that mailman's tip. And the screen, that envelope. If you commit yourself to paying high costs, what do you do if next year, the screen returns much lower than normal?

Commissions and Spread
You will have two types of expenses when investing mechanically, "Per Trade" expenses such as commissions and spread, and "Per Year" expenses, primarily taxes, both Federal and State (and perhaps Local)

With discount brokers, the "Per Trade" expenses have dropped in recent years. It's easy to cut commissions to the $5 to $10 range. American Express's recent offer of Free Trades may push that even lower in the coming years. If you need a discount broker, checking out the Discount Broker Board FAQ is a great place to start.

The official Motley Fool recommendation is keep your cost below 1.5%. That seemed to be based on Foolish Four-like portfolios with a annual return of about 20%. That 1.5% converts to 7% of CAGR.

And that means commissions AND SPREAD.

Think of the last time you went to see a movie. If you're like me and live in a big city, Where did you park? If you got there early, you had the chance to cruise around and find a parking place on the street. But if you were running a little late, you might have parked in the local Parking Garage and paid. Commissions are the tickets, spread is the parking. If you were figuring up what that date last night cost you, you would include parking wouldn't you? Both are money out of your pocket.

Take a look at this thread
then consider these two tables:

PEG 1-4, semi annual, 47.9% CAGR
start total commis spread drag TER actual
cash gain gain
$2000 $958/47.9% $160 $29 $30 10.9% $739/36.9%
$8000 $3832/47.9% $160 $112 $45 4.0% $3515/43.9%
$20000 $9580/47.9% $160 $301 $52 2.6% $9067/45.3%

RSOverlap 2. monthly, 62.2%CAGR
start total commis spread drag TER actual
cash gain gain
$2000 $1244/62.2% $420 $108 $193 36.1% $523/26.2%
$8000 $4976/62.2% $420 $508 $259 14.8% $3789/47.3%
$20000 $12440/62.2% $420 $1723 $73 11.1% $10224/51.2%
(TER-total expense ratio or costs/start cash)
Table assumes $10 commissions, spread of 0.41%

Now I admit it's alittle like comparing apples and oranges, since one is a 2 stock monthly, the other a 4 stock semi annual, but look at the costs. For small portfolios, commissions eat the lion's share. For larger portfolios, spread takes even more. And drag creeps in. "What's drag?" For every dollar you spend on commissions and spread, that's one less dollar compounding and making you money.

Here's a quick table:

Screen Cash needed per stock (1) to return 90% of CAGR
CAGR Monthly Quarterly Semi Annual Annual
20% $25000+ $4000 $2000 $1200
30% $13000 $2300 $1400 $ 800
40% $ 9000 $1800 $1100 $ 600
50% $ 5800 $1300 $ 800 $ 500
60% $ 4900 $1100 $ 600 $ 400
Table based on $10 commissions, spread of 0.25%

Look at how monthly costs increase as the CAGR goes down!

So if you wanted to start a quarterly screen, 40% CAGR, picking 4 stocks, you would need $7200 to return 90% of the CAGR. More on this table later

"What if I don't have that much money?"
You might want to consider Drip investing to get started. You can check out Drips at
Or if you have $500 to $600 a month, consider a "Dozen's" approach. Check that out at

All of this assumes you are trading in a tax free account.

The Dreaded "T" word
Somehow our government did something right, ROTH IRAs.

(I don't know how that happened, but I vote for space aliens.)

Probably, no single thing will save you the new investor more of your nest egg as will a Roth IRA. While you can only put $2000 a year in one, all of your future gains are tax free.

Let me say that again,All of your future gains are tax free.

(Unlike a regular IRA, where the contribution is tax free and the gain is tax deferred)

If you are a new investor and do not have a Roth IRA, I strongly recommend you open one. You can find out about IRAs at

(Quick Note: I don't know a thing about the many other IRAs available. If you think you're eligible to use one, be sure to check.)

What about the money I want to invest after fully funding my Roth IRA this year? How will taxes effect that money? First check out synchronicity's tax article
and comments

"Ahhhhh" (I can hear the gears grinding) "That's why everyone tells me to trade only annual screens in my taxable account! And since in my Roth, none of my gains are taxed, trade anything less than annual in it!"

You get a gold star on your report card.

You should also check out this thread
And concerning when to go monthly in a taxable account, see BAGoldman's thread starting here

(Cover the Author's Ass Time: Be sure to check with a good accountant when it comes to taxes. Read up on things. Your situation is not the same as mine, and the IRS doesn't seem to have a sense of humor when it comes to mistakes. Also taxes vary from state to state. Having your investments return 50% means nothing if you have to give most of it to the government.)

If you are a beginning investor with limited funds, mechanical investing may not be for you. Have patience, build up your money. These fantastic returns will be here when you get back.

If "Cost" can be solved with gentle patience, "Risk" can only be solved with the brutal honesty.

My old copy of Funk & Wagnall's defines risk as: A chance of encountering harm or loss; hazard; danger. In other words: "What's the chance of me losing all my money?"

Now be honest with yourself. Could you stand by and watch your portfolio drop 25% and not be on the phone to your broker shouting, "SELL! DAMN IT SELL!"

And its not just downside risk, but the upside as well. If you look at backposts, not a week goes by that someone says "XYZ climbed 80% so I sold to lock in my profit, ONLY TO WATCH IT ZOOM TO 200%."

IMHO, this one thing causes more problems than anything else in mechanical investing.

Lie to your lawyer, lie to your accountant, lie to your priest, but don't lie to yourself when it comes to the risk you're willing to take with your stocks.

CAGR - Compounded Annual Growth Rate - the annual growth rate of an investment, similar to the APR on a bank deposit or loan. The same as the average of the returns, only using log based averaging.

Standard Deviation vs. GSD
(Ode to the Statistically Challenged: I'm the first to admit, when it comes to this stuff, 90% goes right over my head. If you understand all the math, my hat's off to you. If you don't, NO WORRIES. Understand that we're talking comparisons here.)

When we talk risk, we are usually discussing volitilaty. Or "Deviation."

Standard deviation measures "how often, how far" - how often a return will fall a given percentage from the mean. In either direction. William Lipp give an explanation in
To complicate things further, investments don't behave in a nice linear manner. They are better modeled when using log based deviation, or Geometric Standard Deviation (GSD). It's an imprecise name, because there is no such creature in the statistical literature, but it best describes the procedure we have adopted.

First proposal to use GSD by William Lipp
in particular, how to figure it up

"But doesn't everyone use Standard Deviation? Why don't we?"

William Lipp wrote an excellent post concerning GSD vs. SD
And TMFElan does too in

"We're getting down to the definition of risk, and many people find the Sharpe Ratio's definition lacking. There's been beaucoup discussion of this subject on the MI board. We've discussed Down-side risk and the Sortino Ratio, the Ulcer Index, and a whole slew of other risk measures. It's one reason that the back test sites (bdfinney's and gritton's) have standardized on using GSD instead of SD and why they present the +- 1,2 and 3 standard deviation returns. People can look at those numbers and get an idea of the chance of a loss of any size they perceive as risky, and essentially evaluate risk for themselves."

"This is really returning to the question of "what do we mean by risk?" We seem to be comfortable that the risk we are trying to quantify is the random fluctuation of returns. We are pretty clear that random upward fluctuations are not "risk" in the sense we are trying to quantify. "

In statistical terms, the shape of return distribution is usually represented by a bell curve.

. Mean, or CAGR
. xxxxxxx
. xxx xxx
. xx xx
. xx xx
. xx xx
. xx xx +1 sigma
. xxx xxx
. xxxx xxxx
. xxxx xxxx
. xxxx xxxx+2 sigma
. xxxx xxxx +3 sigma

We make the assumption that stock returns are distributed in a similar manner. For discussion in statisics, "Sigma" is used to indicate the returns at 1,2, and 3 standard deviations around the mean, commonly called "Sigma." (1sigma, 2 sigma, etc.) We know the cumulative probability of returns that are less than 1, 2 or 3 standard deviations. Mean +- 1 SD is 68.27%, Mean +- 2 SD is 95.45%, Mean +- 3 SD is 99.73%.

So a CAGR - 2s -10.5% means that there is roughly a 2.3% chance that the screen will return less than -10.5% in any given year.

Why use log based deviation? Using traditional deviation can give us a less than 100% result, which can't happen. You can't lose more than 100% of your money. See this example

Traditional mean 46.8% average Log-converted mean 44.9% CAGR
1 SD range: +22.0% - +71.6% +21.8% - +72.4%
2 SD range: -2.8% - +96.4% +2.3% - +105.2%
3 SD range: -27.6% - +121.2% -14.0% - +144.2%

See also Dave Goldman's post

There has been some recent concern that given our small sample size, +13 years of stock data, whether the 2 and 3sigma numbers are accurately showing us the chances. Check this thread,

Consider the returns for a PEG5 annual screen.

1986 27.4
1987 30.3
1988 24.9
1989 91.8
1990 19.1
1991 137.3
1992 47.0
1993 11.2
1994 59.4
1995 57.0
1996 49.5
1997 47.0

CAGR for 86 - 97:46.9, GSD 27.5
But add 1998 returns of -9.3, and CAGR drops to 41.6, and GSD falls 15% to 23.6.

So if we have another bad year, GSD may fall even more.

Remember this table?

Screen Cash needed per stock (1) to return 90% of CAGR
CAGR Monthly Quarterly Semi Annual Annual
20% $25000+ $4000 $2000 $1200
30% $13000 $2300 $1400 $ 800
40% $ 9000 $1800 $1100 $ 600
50% $ 5800 $1300 $ 800 $ 500
60% $ 4900 $1100 $ 600 $ 400
Table based on $10 commissions, spread of 0.25%

Now take a look at this post of TMFElan
(it concerns seasonality, but his discussion on GSD variation is right on.
In particular this quote:
" Looking at January, (for PEG1-4 annual)we have CAGR=50.8% and GSD=22.3%. This tells us that there is a 68% chance (based on a Normal distribution) that the actual return will be between 23.3% and 84.4%."

So if you wanted to use this screen, you would need a minimum of $2000, (4 x $500), but consider that the returns may dip as low as 23.3% in 68% of the time, you might want to take 22.2% as a guide, starting the screen with $4800 instead.

Ideally we would have 40 or 50 years of data, with which to backtest. (Which yes, we're working on) Until then, consider the GSD number an aproximation of the deviation you can expect.

Some Final Quotes:
"For all you people panicking out there here is a fact:
If you had invested at the beginning of 1987 $10K in the PEG Semi annual with adjustments in Jan/July not including commissions, taxes and brokerage fees you would have entered 1999 with $1.2 million dollars.
However during that period you would have lost
29.4% of your money in the 4th quarter of 1987
25.3% of your money in the 3rd quarter of 1990
33% of your money in the 4th quarter of 1995
and of course 15% of your money in the 3rd quarter of last year.
So I know its hard to start on the down side but if you can't stomach 33% loses maybe this isn't for you because remember the above scenario is definitely too good to be repeated."

"You have to remember, statistics can only indicate chances of something happening, and it can only determine chances based on the data that is available. If the screen only has data during a bull market, then it can only make "predictions" based on a continued bull market. This is one of the reasons an extensive backtest is important, as if you didn't already know that. ;-)"

And finally I'll leave you with this quote:
"Classical statistics can do a good job of telling an investor what county he/she is in, in terms of risk. Some people extrapolate this out and want to claim that classical statistics can derive what address you're headed for, but it's more of a blunt instrument than that. (In other words, show me five high standard deviation screens, present the statistics about their risk, and I'll bet at least a few of the screens in question defy what the statistics want to predict--other screens will satisfy the predictions.)As blunt instruments, the statistical techniques we have are very good and handed down over long periods of time. But they are butter knives, not scalpels."

Addition Reading:
MakeItJake's "Risk & Resampling", the itialized section especially
Discussing of Downside Variation by Frank Sortino &
Peter Kuperman's thread on Downside Variation &
"Measuring Risk" by Michael Peltz
Risk: A Challenge to the Accepted Wisdom
from an NYU Stern School class on finance,
Random Numbers
and this for more
Bayes's theorem and MI
other interesting threads

As always a thank you to those people whose posts and work I've referenced
Print the post  


When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.