The fallacy of the "best idea".

There is an idea floating around in the value investing world that investors should only purchase their five or ten best ideas.  The idea is that if a stock being considered for a portfolio isn't as good as the 'best idea' then why buy something sub-par? Instead investors should continue to pile into their best idea instead of finding new ideas.

The best idea line of thinking gets a lot of traction and has become incredibly popular after the 2009 market recovery where concentrated investors have trounced the market.  Investors are always chasing the best strategy and seeing as how concentrated investing has done so well recently it's no surprise that it's the strategy du jour.  I've posted in the past regarding my thoughts on diversification, and this post isn't intended to re-hash those thoughts.

I've always thought that there's a disconnect between academic investing and real world investing.  In an academic study an investor has a starting point, they pick a portfolio of some number of stocks according to specific criteria and then their performance is measured against a benchmark for a limited time range.  A study like this is intended to show how specific criteria, or a style of investing matches up against the market, with the market being the control group.  Numerous academic studies have shown that low-anything (P/B, EV/EBITDA,P/S,P/E) investing works.  An investor who purchases something for less than a chosen average metric and holds for a period of time will do better on average compared to an investor who purchases the market.

From these studies the notion that value investing, momentum investing, small stock investing beat the market is established.  The concept of value investing as defined through academic literature, or investment classics like like Security Analysis or the Intelligent Investor are usually the starting point for an investor defining their strategy.  There is a natural inclination in all of us to take something that works and think about how it can be improved.  We look at a study that shows buying all low-P/B stocks and think "what if I could remove the obvious duds, then I'd do better."

The idea of investing in a few best ideas is simply an extension of this concept.  What starts as investing in all low P/B stocks turns into only investing in the 'best' low P/B stocks, which leads to investing in only the 'best' cheap stocks.  On paper this makes perfect sense, it's intuitive.  It's been reinforced by the bull market coming out of 2009 where every concentrated investor has killed the market by investing in the best turnarounds.

Within this environment I'm asked continually why I'm diluting my portfolio by investing in my 10th idea or my 30th idea or my 50th idea.

I believe there are two problems with the best idea type of thinking: portfolios are built over time, and we don't know our best idea.

Not many investors start with a blank slate.  A new fund manager might start with a pile of cash that needs to be invested in a portfolio, but most managers start gradually as well as funding is received.  Individual investors create a portfolio through savings over time.  A portfolio is built over time, not in a single day.  During the building period new is released about holdings and prices are continuously changing.  What was the best idea initially might become a mediocre idea after a news release.  Or the best idea might experience a run-up that leaves it much less attractive.

The investment landscape is always changing, it's foolish to think that on the day one sits down to create their portfolio they will have they will already have their best idea.  Some investors go years ambling along before they find an idea that multiplies their portfolio many times over.  Other investors never hit it out of the ballpark, but consistently do well with most investments.  If the best idea isn't known on day one then what's an investor to do when they find this new best investment?  Should they sell everything and concentrate their entire portfolio in it?  Logic would say if you find a true best investment you should concentrate everything in it.  Why dilute your best idea with your second best idea?

The problem is we don't truly know our best ideas, even if we think we do.  Some ideas resemble situations that worked out well in the past, but there is no guarantee that they'll work well in the future.  There are no situations where one can be absolutely certain about an outcome.  Even an executive guiding their company through a merger isn't certain until the deal closes, the other party could pull out at the last minute.  There is always an element of chance involved in any and all investments.  We all know this, and this is why we don't see investors with one stock portfolios.  Of course there are exceptions, there are some super human investors who can pick five stocks that all triple and crush the market.  I don't know if they have a crystal ball or if they're just that much smarter, but they do exist.  The question one needs to ask themselves is whether they have that crystal ball ability.  I know I don't, and most probably don't either.  I know it's fun to think we're all above average, but math doesn't allow that.

When I say I don't know my best idea what a lot of people hear is "I don't know a good idea from a bad idea, they all look the same to me."  There is a subtle nuance here that needs to be clarified.  A company with high returns on equity selling at a low price has the potential to generate better returns compared to a net-net at 2/3 of NCAV over the long term, I know this.  When I say we don't know our best idea I mean we don't know what idea will perform the best.  It's possible a net-net will turn itself around and come roaring back with a magnificent gain.  On the other hand a great company with an indelible brand such as Coke could become yesterday's value if attitudes turn against sugary drinks and consumption drops.

Here's an exercise, go back through your portfolio statements from previous years and mark down what you thought was the best idea at the time.  Of those best ideas how many worked out exactly as you expected?  If the number is 70-80% then you should probably be concentrating in half a dozen stocks or less.  But if your ability to predict the future isn't that high then it's worth broadening up the portfolio and start hitting singles and doubles instead of gunning for home runs.  When I look at my portfolio I have had many companies I thought were potential rockets that did nothing, and other companies that were just simply cheap go on to double or triple.  It's much more common that I had a plain undervalued stock double rather than some turnaround story stock with a catalyst quadruple.  I find it easier to find stocks that are likely to gain 50-100% rather than finding ones that will return 400%.

I like to minimize the number of assumptions or moving parts necessary for a thesis to work.  A healthy company with an extremely low valuation is easier to purchase compared to a 'good' company at a reasonable valuation.  The low valued company just needs to survive long enough for their market value to correct, the good company needs to continue to execute at a high level without making mistakes to hold up their valuation.

For more ideas on how portfolio management fits into my overall investing system, go here.


  1. Excellent post, another related idea is the "20-hole punch card" established by Munger/Buffett. More of a concept than actual investment advise, it puts the "best ideas" concept to the extreme, where you're actually supposed to stay within a limited number of investments in your entire career.

    This has the strange implication that a 20 year-old should have the ability to compare current opportunities to those of a future investment lifetime. I suppose most people use this idea as "food for thought" rather than literaly, but its popularity as a concept is as undeniable as its impracticality.

  2. "What was the best idea initially might become a mediocre idea after a news release."

    Classic line there! I'm pretty sure I've had that happen the day after i bought something before.

  3. Hey Nate,

    You are talking about the classical shift from Graham's investment style to Munger's style (modern Buffett's style). Buffett has said many times that he wouldn't have done so well if he had kept investing Graham's style.

    Latest Buffett's remark on the issue from 2014 BRK annual meeting:

    "Graham emphasized financial factors for those birds in the bush, while Fischer
    emphasized business quality factors. I emphasized what Graham had said. Charlie came
    along and told me that I was wrong. He said that he had learned more in his law studies
    than I had learned in my business studies."

  4. Every statement on investing has to been seen in its context. Based on your blog, your context seems to be "investing in OTC companies, which do not report much". I have no idea how much access to the mgmt you have, but I assume that your decisions must be more quantitative than qualitative. If you would model the companies, talk with the management, analyze the competitors etc, you simply could not have more than 20 companies in your portfolio (assuming a one-man team).

    For me the effort to really understand my holdings is the main limitation on the diversification. It is not possible to diversify without diminishing the quality of the analysis. After 15 positions diversification become diworsification.

  5. As someone who manages a fund designed to outperform long term, with an eye on monthly vol, I agree with you 100%

  6. As someone who manages a fund designed to outperform long term, with NO eye on monthly vol, I disagree with you 100%. =)

  7. To me it seems like a fallacy to claim that you are capable of effectively selecting 60 undervalued stocks but not capable of effectively selecting 15 undervalued stocks. The analytical method for selecting your top 15 ideas is exactly the same as for selecting your top 60 ideas. There has to be a cutoff somewhere, you just choose to make it very lenient.