You need an investing system

Investors seem to have an intrinsic drive to classify themselves.  People will say something like "I'm a mix of Graham and Buffett with a dash of Rockefeller and the temper of Carnegie."  Sometimes these classifications border on ridiculous, other times confusing.  Even still investors continue to classify themselves.  We use these heuristics because often it's easier to identify with an investor's system, rather than developing our own system for investing.

I finished reading two books recently, The Signal and Noise by Nate Silver, and  How To Fail At Almost Anything And Still Win Big by Scott Adams.  I think both books are worth reading for different reasons, but both have applications to investing.

Scott Adams outlines his belief that successful individuals use systems rather than goals to accomplish things in life.  With a goal you are in a state of perpetual failure until you achieve the goal, then you need another goal to move onto.  He believes we need systems, continual patterns that are sustainable and drive us towards our desired outcomes.  An example of this might be losing weight.  A goal might be that we want to lose 20lbs.  To achieve the goal we decide to not eat candy or have second helpings.  The weight starts to drop and eventually we hit our goal.  What happens then?  Do we continue to not eat candy and smaller portions?  Is it sustainable?  If we sneak a Snickers bar are we failures?

A system looks at the problem differently.  It might be that we can eat unlimited quantities of certain food groups, but keep others to a minimum.  This is a sustainable system.  Maybe I am free to eat as many vegetables as I want but keep candy to a minimum, such as special occasions.  If I'm hungry I can have carrots instead of denying myself anything to eat.  Being able to eat to satisfy hunger is fine as long as it's something healthy.  My own view on this is that I've never seen someone gain weight by eating too many fruits and vegetables.

Nate Silver's book is much different, he discusses how to view the world probabilistically.  He believes by applying Basyean statistics we can increase the accuracy of our forecasts and enhance our forecasting outcomes.  Silver discusses all sorts of forecasting problems from weather to earthquakes to the stock market and politics.  The book is long, but it was an enjoyable read.  There were two take-aways that I believe apply to investing, application of the power law, and thinking about investments probabilistically.

It might be helpful to read this post as a follow-on to my post on diversification.  I want to talk about systems first, and then how the power law and probability fits into the system.

I think it's critical that investors create a consistent system for their investments.  What I mean by this is I think we need to approach investments, measure investment success, and view investments in a consistent and repeatable manner.  Sometimes I'll encounter investors who say they do a little GARP investing, plus some dividend investing, plus some value stocks and a few moat companies.  To me that seems like they're throwing things at a wall and seeing what sticks.  It's hard to find investments if we don't have a structure to view things within.

We need to find styles that fit our personality.  It doesn't matter if it's investing in small cap growth companies, or distressed credits.  There is something out there that will make sense to you and will be almost second nature.  Don't invest in a manner because someone else does, or because someone famous has made a lot of money investing like that.  Invest in a manner that makes sense to you.

When I meet someone who wants to talk about investments, but isn't that knowledgable this is how I explain what I do in non-investing terms:

I like to go to the older part of town and look for businesses that look like time has passed them by.  I will buy them for some amount.  When I visit my new property I find that 80% of my purchase price is sitting in cash in the cash registers, and that I can sell the inventory for the other 20%.  I can also sell the building for a gain as well, along with the fixtures.  Sometimes I let the business run because the cash it generates pays me back in a year or two, but not always.

People understand this, it makes sense to them.  The second question is naturally "how do you find these places?" But there is no confusion as to my process.  They key is that my explanation is also the  system I use while investing.  I am looking for things at egregiously low valuations.  I'm not buying a Mercedes at 10% off sticker price.  I'm buying a Chevy Cavalier on Craigslist and reselling it a week later at double the price.

My system is consistent and easy to apply.  I can apply it to stocks, or bonds, or real estate, or literally anything that can be bought or sold.  If I consistently apply my system I also know that I will consistently earn a satisfactory return.  I might not earn a return on every investment, I make mistakes, but over the long term I will earn a consistent return.  When I look at a new potential investment I view it through the lens of the system I'm using.  If someone is pitching me a product that will take over the world it just don't fit with how I view things.  That doesn't make it bad, or wrong, it's just not something I have experience with.

That brings me to one of the points from Nate Silver's book on thinking probabilistically.  One thing many investors struggle with is how do we know something is truly worth more than what it's selling for now?

Nate Silver's book flipped the lightbulb for me on this issue.  It helped me recognize that what I am doing when looking at companies is handicapping them, or thinking about the probability of them being worth more.  This is different than extrapolating the future.

Here's an example.  Barrons might profile a company and say they're trading at $20 and the paper thinks they're worth $22.  The reporter spends a page of text explaining why they think this.  The paper will probably make a great argument, but I ignore that and look at the gap between the two numbers.  They are saying that the stock has the potential to rise 10%, or put another way they're 90% of their fair price.

I look at a thesis like that and think that with a gap that small there must not be much uncertainty as to their real value, or the estimate needs to be right.  Is the probability of them being worth 10% more greater than the probability of them being worth less?  As the value gap closes between where a stock trades and what they're worth the amount of uncertainty needs to diminish for the investor to be right.

The opposite of this would be some of the investments I look at.  Take for example Conduril.  They were selling for 40% of NCAV and 2x earnings when I found them.  There is a larger probability that they're worth more, than worth less.  Maybe they're not worth 10x earnings or 5x earnings, but the probability of them being worth more than 2x is greater than them being worth less than 2x.  In a case like that it makes sense to invest.

This is why it's easier to invest in a much cheaper company compared to a company that's closer to full value.  If you think of two probabilities, one that it's worth more, and one that it's worth less, the probability that it's worth more decreases as the price goes up.  When comparing two companies side by side I will take the cheaper one in most cases.  The cases where I don't do this are times when there is more certainty that the more expensive one will rise.  An example of this might be two cheap companies, one with lousy self-dealing management, and one with honest management.  It's more likely that the company with honest management is worth more, there is less uncertainty.

I know a lot of fund managers use this thinking when looking at lottery ticket type of investments.  They look at the probability of something positive happening and the resultant gain, and then the probability of something negative happening and the resultant loss.  If the gain outweighs the loss then it's worth taking a position.  Over a single data point the probabilities don't mean much, but over a set of investments over a period of time the probabilities should hold true.  Referencing back to my diversification post, I could understand someone making five or ten investments like this at a time, but making one or two seems reckless.

The last thing I want to discuss is the power law.  Silver talks about this within the context of earthquakes.  People often misinterpret the probability of something greater happening because it hasn't happened in the past.  He uses the example of an area that's only experienced earthquakes measuring 5 or 6 on the richter scale.  Instead of thinking that nothing higher can happen because it hasn't happened in the past, the fact that any earthquakes have occurred at all should be taken as a warning that something greater can happen, and it's likely the magnitude will be greater.  That an earthquake measuring 5 on the richter scale happened is indication that an earthquake measuring a 7 can happen as well.  It's more likely that an area with prior earthquakes will experience a size 7 earthquake than it is for an area that's never had a earthquake to get one measuring a 2 or 3.

How does this translate to investing?  I believe the past is indicative of what might be possible in the future.  If a company has never been profitable is it likely that suddenly something will change and the company will make fantastic profits?  The lack of profits most likely indicates that it'll be harder for them to shift course and be profitable.  Likewise a company that's been profitable in the past, but is dealing with a temporary situation will likely be profitable again in the future.

We can't extrapolate the past mindless into the future, but we can look at the past as some indication of what could be possible in the future.  If a company has a management team that's done well previously then there's reason to believe they can be profitable again in the future.  If a company's management is experienced in losing money and constantly issuing shares what could prompt a change in the future?

An example application of this might be biotech investing.  It's more likely that a company with previous drug approval experience will get a new drug approved over a company with no experience in getting drugs approved.  That's not to say it's impossible, just not as probable.  If a company has experience with radical transformations and has come through successfully then we might expect that when faced with a radical transformation it will be possible for them to succeed.  If a company has never done anything transformative and then they attempt it we should view their chances of success with skepticism.

None of these concepts were entirely new to me, long time readers will recognize that I've been applying these concepts for years.  But what is new is having a name for them, and being able to identify exactly what I'm doing.  Sometimes we do things without being able to label it, I've now been able to label a portion of my process.

For an even more detailed look at my personal system, go here.

Disclosure: I receive a small commission if you purchase something from through the links provided.  The price you pay is not marked up, Amazon builds this commission into the cost of all of their items.


  1. Succinct and informative. Thanks for sharing your reflections from these two works and applying them to investment philosophy. Like you said, these ideas aren't new but thinking about them in this concise, organized way helped drive the fundamental concepts home.

  2. Great post. Agree with Taylor above, very clearly formulated and well-written. Next step is to execute on these ideas, and then repeat over years...

    1. Thanks, yes that's the key, find the process and repeat it through all market conditions.

  3. great intuition when you mention: "I think it's critical that investors create a consistent system for their investments." I've got a related piece on "Are you Trying Too Hard" which makes a similar point, but in a much more long-winded way with multiple references to academic literature that support your basic thesis. I like your shortened version--ha!