Comments on: Walking the talk: Applying back-tested investment strategies in practice

By: A Miscellany of Deep Value « Greenbackd

A Miscellany of Deep Value « Greenbackd — Tue, 14 Dec 2010 05:03:03 +0000

[...] Shleifer, and Vishny. For me, this is cognitive dissonance defined. Aswath Damodaran also has some interesting contributions on this topic. Second, some real-world received wisdom in value is no wisdom at all (or at least, that is what [...]

By: The Small Cap Paradox: A problem with LSV’s Contrarian Investment, Extrapolation, and Risk in practice « Greenbackd

Wed, 27 Oct 2010 04:01:31 +0000

[...] an earlier post, Walking the talk: Applying back-tested investment strategies in practice, I noted that Aswath Damodaran, a Professor of Finance at the Stern School of Business, has [...]

By: stockmanmarc

stockmanmarc — Thu, 18 Feb 2010 18:13:39 +0000

Great Post!

By: NCAV NNWC Screening Strategy

NCAV NNWC Screening Strategy — Thu, 18 Feb 2010 08:05:59 +0000

[...] Greenbackd also wrote an article perfectly timed with my first one where he analyzes Damodaran’s discussion of how fees eat away the real returns of backtests. [...]

By: Back-testing the performance of net current asset value against net net working value « Greenbackd

Thu, 18 Feb 2010 05:02:06 +0000

[...] basis, which may be difficult to do in practice (as Aswath Damodaran pointed out yesterday, many a slip twixt cup and the lip renders a promising back-tested strategy useless in the real world). Here’s the performance of the 15 stock [...]

By: bcinvesadvisor

bcinvesadvisor — Thu, 18 Feb 2010 02:12:36 +0000

Damodaran’s logic is circular. If active portfolio managers on the whole own what makes up a reasonably large portion of the market they are going to own a representative sample of the market and as long as there is even one dollar of expense involved on average they will under perform.

The average is the wrong measuring stick. Nobody cares about the average. The average author sells sells 25,000 books while the median author sells 6,000. A few extremely popular authors drag up the average. The median is a better indicator of how many books an author can expect to sell. If we carried over the examination to the height of adult males the median and average would be nearly identical.

The average is of little value when looking at book sales because of a few outliers. Those are the folks whose performance is of interest when looking to see who outperforms the markets. If they are using value strategies and beating the market then it may be an indication that value strategies, but not a confirmation, that these strategies are successful. (That is unless all those who fail to be the market are using value strategies also.)

Damodaran’s premise that we need to look at the average portfolio manager is absolutely useless. It is amazing that someone so well connected could say something so dumb.

By: DeLano

DeLano — Wed, 17 Feb 2010 15:42:57 +0000

Excellent post. I have been contemplating these same thoughts. The friction associated with sub NCAV stocks is a major problem, however I do not believe it is one that cannot be overcome.

Using http://www.stockscreener123.com I backtested a strategy of selecting the cheapest 30 stocks based on the market cap to NCAV ratio, that were not OTC and not ADR’s. The portolio was rebalanced annually. From March 31,2001 until now the strategy is up approx. 1700%. While the S+P is about even. Even with taxes, and spreads, and transaction costs, I think this strategy could easily outperform the market.

Definitely check out http://www.stockscreener123.com. It is easily the most powerful free stock screener right now.

By: Parker Bohn

Parker Bohn — Wed, 17 Feb 2010 09:32:32 +0000

This is exactly the problem I’ve been wrestling with recently.

I’ve been doing fine vs the market, but I’ve been dramatically underperforming my system, which on paper yields market-crushing returns (several percent outperformance per month).

I should have no back-testing (curve-fitting) bias, since my system is to score a stock based on its current data. In other words, all of my system’s hypothetical results are based on out-of-sample data.

This just goes to show that constructing a real portfolio and a hypothetical portfolio are very different things, especially when dealing with very illiquid securities like I do.