How many different businesses in your portfolio?

[Liberty]

I just want to thank everybody who has voted in the poll and shared their approach. It's very interesting.

 

Obviously this board seems to have a pretty strong bias for relatively concentrated portfolios. Close to 60% of the people who voted have between 1 and 10 businesses in their portfolio.

[twacowfca]

http://cornerofberkshireandfairfax.ca/forum/index.php?topic=2753.120

 

Continuation of discussion from the above link.

 

 

"There is just no way, that you can consistently get better results from your 20th to 50th best idea, than from your 1st to 10th best idea.  Not on a regular basis.  And if those 10 ideas are not correlated industries, you will do perfectly fine over the long-term.  In my opinion, better than any other method that can actually be replicated by more than one person!  You don't have to take my word on that...just take it from Buffett and Munger:"

 

Again, check the Validea link. It's not even a debate. 20 stocks perform better than 10 in Graham's system.

 

You're right, it shouldn't be about someone's "word," it should be about evidence.

 

 

Remember, Harry, when you say that "20 stocks perform better than 10" in Graham's system, you're saying that the 20 stocks in the 20 stock model portfolio performed better than the 10 stock model portfolio.  I know it seems obvious that you're saying this, but you have to think about the assumptions you've made:

 

(1) That the ordering of the stocks is based somehow on the ordering of the stocks in between the two portfolios.  For instance, best ideas #1-#10 are in both portfolios whereas best ideas #11-20 are only in the 20 stock portfolio.  This may be true from a quantitative standpoint (there's some arguments against that as well), but you're neglecting any qualitative assessment of the stocks in either portfolio.  In other words, what if the "best" idea was actually in the 20 stock model portfolio but not in the 10 stock model portfolio?

 

(2) I might be making an assumption here, but I think that in using the 10 and 20 stock model portfolios in your discussion, you're assuming that if 10 stocks are good and 20 stocks are better, then your 100+ stocks are even better.  Of course, taken along its logical trajectory, one would assume 1000+ stocks would be the best, though that would fly in the face of common sense because the more stocks in your portfolio past some point, be it 100, 500, 1,000 or 3,000, the more likely you are to be modeling an index.

 

Oddly enough, despite what I just said, I don't disagree with the argument that a portfolio with 100+ names can outperform.  I just think it's much harder to outperform with 100+ stocks than with <10 stocks.

 

I wouldn't make too much about the outperformance of the 20 stock portfolio.  Both sample sizes are small.  Was the outperformance statistically significant?  At what confidence level?  In general, the smaller the sample size, the higher the probability that a given sample will "outperform" the distribution of the larger sample, while the volatility of the larger sample generally will tend to be be lower.  However, this cuts both ways.  In general, the smaller sample will also have a tendency to underperform as well!

 

There is also the problem that extremely small sample sizes may select a higher percentage of companies where the data that passes the screen may not precisely reflect the true state of the business.  For example, using a screen to find the stock that has the lowest P/E in the universe may vomit up a fraud or a company in deep trouble.  

 

Harry's argument is right from this point of view.  But Sanjeev's argument has great merit if the analyst is very good and very careful.  The BG screen is among the very best.  Yet WEB's performance has been far better, especially before he was limited to megacaps and when he sold companies as their stock prices approached IV.

 

By the way, the Validea Buffett screen is a shallow reflection of what Warren really does, based on the superficial understanding of the site manager's interpretation of Mary Buffett and David Clark's book.  What Warren really does in picking stocks, other than special situations, is more related to extension of criteria similar to the BG screen, usually through a longer time frame than 10 years, with additional inclusive and exclusive criteria.

[Liberty]

What is this "BG screen" you talk about?

[merkhet]

 

Harry's argument is right from this point of view.  But Sanjeev's argument has great merit if the analyst is very good and very careful.

 

 

I think it just boils down to how most people view the world.  For instance, when I was in law school a criminal procedure professor boiled down the entire subject into one phrase - "you either trust the cops or you do not."

 

On Harry's side, he is correct that if you cannot a priori sift through "value" stocks with some granularity and determine the "best" ideas, then wider diversification (in the sense of more positions and possibly industries) will protect your downside while not necessarily limiting your upside to a degree that would keep you from outperforming.

 

On Sanjeev's side (and Mungers), he is correct that if you can a priori sift through "value" stocks and with at least some confidence determine an order to stocks that ranks them based on some "best" criteria, then it is in fact irrational to put more money into your #20th+ position than to add more money to position #1.

 

So as the criminal procedure professor would say, "you either believe in a priori stock picking ability or you do not."

[twacowfca]

What is this "BG screen" you talk about?

 

Sorry, BG referred to the Validea Benjamin Graham screen that Harry used as an example.  :)

[Liberty]

What is this "BG screen" you talk about?

 

Sorry, BG referred to the Validea Benjamin Graham screen that Harry used as an example.  :)

 

Thanks! I'm usually pretty good at keeping track of acronyms, but sometimes there's one that could mean a few things and I can't quite figure it out  :)

[Guest Bronco]

4 Long Holdings

FRFHF 23.6%

SMBC 42%

FSFG  14.2%

JPM  15%

Cash 5.2%

 

__

Short IWM ~60% of portfolio.

 

 

What's your edge on FSFG? I never heard of it.  I am not being cynical here - I am interested in learning.  Seems like it trades well below BV.  What do you know about it and can share?  Seems like an interesting pick.

 

 

[Liberty]

I'm now in 7 public businesses, with my top 2 representing over 60% of my portfolio.