NCAV NNWC Backtest Refined

Stock Screen Strategy and Backtest Series

NCAV & NNWC Backtest Criterias Updated

Following on from the great discussions and feedback from the initial NCAV NNWC strategy, I’ve made some changes to the screen testing criterias to make it more realistic.

The best way to go about the screens would be to screen for companies that had enough daily volume. A reader, Chroma, also mentioned that Graham stated that it was best to buy NCAV stocks trading at less than 66%. Absolutely right. So I included this margin of safety into the backtest.

Purpose of Backtesting

With the backtest, I went as far back as 10 years, but I realized that analyzing a backtested strategy for more than 10 years is misleading. The purpose of a backtest is to simply show you how a certain strategy works over multiple periods of time rather than a straight 10 year period. If at the end of 10 years, the strategy is up more than the market, the immediate assumption is that it must work.

There are of course many studies which already show the performance of the strategy for each calendar year, which is the correct way of going about it, but many services only display a single graph such as the one below from Magic Formula Investing.

These types of graphs are misleading. Why? Because your eyes immediately follow the blue line to the top.

However, to Greenblatt’s credit, he also includes the performance for each calendar year on another page.

I know that having a long term view is important, but to be honest, how many small investors in the world employ nothing but the same investing strategy for more than 10 years? Zero. Most people would give up after 3-4 years if a strategy didn’t work out which is very possible.

The time you start is also critical and changes the whole game. Had you invested $10,000 when Buffett took over Berkshire, supposedly you would have hundreds of millions, but this thinking is completely wrong. Besides, I wasn’t even born at that time. Instead, had I invested $10k when I was 20 years old in 2002 into BRK, I would be up about 50% after 8 years…

This is why comparing multiple short 3-5 years time frames would be much more accurate in a practical sense. You also get an idea of what strategy to follow in different market conditions. After all, the advantage of small investors like us is the ability to move quickly.

And if you want to move quicker in your investment career just click on the image below to download the best investment spreadsheet that will help organize your thoughts and make things easier for you.

download investing checklists

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.

But back to the results.

Updated NCAV NNWC Screen Criteria

  • Volume is greater than 30k
  • NCAV margin of safety included
  • Slippage increased to 1%
  • Rebalance frequency changed to 6 months
  • Test period remains at 3 years

NCAV & NNWC Stocks with Volume

2001 to 2004

2004 to 2007

2007 to 2010

Now that the NCAV stocks are only chosen when trading for less than 2/3 of it’s value, it acts much like a NNWC. This debunks my initial thought that NCAV doesn’t perform well. As long as you buy them with a margin of safety, NCAV stocks look to perform equally well.

Either way, if a recession ever comes around again, I better sell everything in my portfolio and load up on Graham stocks.


What is Old School Value?

Old School Value is a suite of value investing tools designed to fatten your portfolio by identifying what stocks to buy and sell.

It is a stock grader, value screener, and valuation tools for the busy investor designed to help you pick stocks 4x faster.

Check out the live preview of AMZN, MSFT, BAC, AAPL and FB.

28 responses to “NCAV NNWC Backtest Refined”

  1. tedk81 says:

    This is awesome, really really good. Great changes, great breakdowns, very impressive.

    Do you mind sharing where you got your data from? I can’t figure out how to find which stocks would qualify as a net-net in, say, 2002. There are quite a few things I’d like to backtest and play with if I can just figure out an easy source of old data.

    I know that having a long term view is important, but to be honest, how many small investors in the world employ nothing but the same investing strategy for more than 10 years? Zero.

    What % of small investors in the world should even be thinking about picking their own stocks or doing anything at all? IMO, it’s < 5%. People don't want to put the work in, get too emotional, don't want to use their heads. Most people aren't wired to invest well. So I don't think it matters that most change their strategy all the time, they do that because they don't know what they're doing and they can't help themselves.

    Most people would give up after 3-4 years if a strategy didn’t work out which is very possible.

    Sure, again though, that would be to their detriment because they don't understand the nature of markets/variance/automatic systems. Incidentally, I find the longer I'm at this the more I think the Magic Formula is kinda genius because when I look at the list of stocks it selects looking at most of them individually my gut reaction is "ew, don't want to own that." Strangely, I think that means it's probably doing the right thing.

  2. peekay says:

    This looks much better Joe.. Thanks for the post!

  3. zehua says:

    I am curious what the result is gonna be for buying EPV stocks with MOS of 50%.

  4. George says:

    Your findings are consistent with my own findings that 2/3 NCAV is superior to what you call NNWC.

    How did you come up with a 30k volume cutoff? Why not 10k or 50k? I’m asking because I’m also working on similar NCAV backtests.

    I’m trying to find out if there are any patterns associated with stock that are trading below 2/3 NCAV that never recover. I’m having a hard time identifying which stocks stopped trading and/or went to zero. It’s also tough to identify ones that also paid out special dividends as part of a liquidation. Do your backtest returns numbers include dividends?
    .-= George´s last blog ..Festival of Stocks President’s Day 2010 Edition =-.

  5. Jae Jun says:

    @ ted8k1,
    I’ll try to send you an email of all the data.

    @ peekay,

    @ zehua,
    Good idea. I’ll try to figure out how I can do a screen on it.

    @ George,
    From the NCAV article I linked to, it has the following statement.

    “with the AAII Shadow Stock Portolio: The average daily number of shares traded should be four times the amount needed for the position.

    Otherwise, it may be too difficult to get in and out of the position quickly. For example, if I had $5,000 to invest in China Dasheng Biotechnology (CDBT), I would be able to buy 6,330 shares at its January 22, 2010, closing price of $0.79. Therefore, I would like to see CDBT trade at least 25,000 shares a day (spefically 25,320, or 6,330 × 4).”

    Also from my experience trading low price and volume, I found that having at least 30k was a good amount of volume to move in and out quickly.

    @ Greenbackd,

  6. tedk81 says:

    Thanks Jae, look forward to receiving it. I’ll send you back something I’ve been working on lately that I think you might like.

  7. Mark says:

    All three lines are pretty close as of 7/09. The NNWC strategy seems to work best when the entire market is rising (so beat down stocks selling as if they are going to go broke rise the most). The extreme superior performance of small caps and value stocks (like RZV which went from like 10 to 35 between March and Sept) in 2009 accounts for a big part of the superior performance of NNWC stocks.

    I don’t know how much to read into this.

  8. KenKP says:

    Hi Jae Jun,

    I have previously done a number of backtests on NCAV strategy that takes into consideration the overall market P/B level, with the following criteria:

    Buy criteria:
    Price/NCAVPS -10%
    Average Net Income for past 5 years / book value > 10%
    P/B of S&P sector for this security < 3
    Maximum individal security investment: Median P/B value for the security in the past 10 years, or P/B of S&P Sector > 3

    I have tried this combination with good success that provides ample down side protection, maybe this idea can help you in some way.

    Let me know if you are interested in discussing further: [email protected]

  9. KenKP says:

    Sorry, format issue:

    Buy Criteria:
    Price/NCAVPS -10%
    Average Net Income for past 5 years / book value > 10%
    P/B of S&P sector for this security 3

    Rebalancing period: 6 months

  10. vernadel says:


    i’m french and excuse me for my approximative english !
    so between january 2001 and january 2004 result of ncav strategy was arround + 500% ?
    between 2004 and 2007 = arround zero %
    and to finish since 2007 = arround 75%
    are you agree with me?

    if you agree with me ,where can i found stocks with criterions used to do ncav list ?.

    Where can i found results and exacts criterions for very long time period ?
    Thanks and excuse me for my poor english !

    Sébastien From french

  11. Evan says:

    Hey Jae,

    Big fan of your posts. I love how you’re diving into backtests of Graham’s net net strategy. I have a couple questions and a couple comments for you:

    Q. You mention (somewhere) in your posts that you limit the portfolio to 10 of the “best” net nets. How do you determine what is best? Do you mean the cheapest? Do you look at fundamentals such as current ratio, etc? I have no clue what you’re doing here.

    C. I love your backtests. When you’re posting them could you include a chart of yearly returns (eg. 2006-2007: 20%) because that would go a long way to understanding the behaviour of the portfolio in individual years.

    Q. Could you do a couple backtests testing things like the smallest net nets, net nets where insiders are also buying stock, net nets where the company is repurchasing stock, etc?

    C. Like I said, I love your posts. I think you could help people understand the significance of the tests if you spell out exactly what you’re doing.

    C. I would absolutely love some guidance in the areas of stock symbols and SEC documents. Boring, yes, but important. I don’t know much at all about the suffixes on NASDAQ symbols, what all the SEC forms are for or how to properly read them. Little help? 🙂

    Keep posting and I’ll keep reading!


  12. Jae Jun says:

    Hey Evan,

    1. I just ranked them according to certain ratios and the % discount it was trading at, can’t remember which. All written down in my secret notebook somewhere 😉

    2. I did that with some other screens but it doesn’t reflect the strategy well because all these backtests are not designed to analyze 1 year at a time but rather over a timeframe.

    3. I could try but can’t promise anything. I’m working on other backtests that are taking up a lot of time. I’ve been working on one for about 3 weeks but I’m still not getting what I think I could.

    4. thank you

    5. Try these links which will describe everything you need to know.

  13. Jim says:

    Hi jae Jun. Great Analysis. I was wondering if you had a count of how many companies met the 66% of NCAV (with vol over 30K) criteria, tracked by year. For example, how many companies met all the filters for jan 2001, jan 2002, etc. Thanks again!

  14. James says:

    What is your methodology for conducting backtests? Where do you get your historical stock data?

  15. Krystofo says:

    This is the first clear modern long-term backtest I have found for NCAV-type screens. In addition, you have clarified for me the differences between NCAV and NNWC or “net-net.” Thank you also for clarifying that the investor either should be sure he is using net-net, or if using NCAV, to insist on Graham’s original standard of being below 2/3 value. I surmised in the past that stocks today might be more overvalued than in Graham’s time, and therefore I believed that an NCAV of 1 should work just as well, but you have proven this to be an incorrect assumption!

    One small suggestion. You begin by pointing out why year-by-year tabular results are better than linear graphs, but then you only provide linear graphs. So I am having to make up my own tables, based on estimating where your lines end up. Some tables would be appreciated, but I suppose it all averages out. Thank you in any case for what you have done!

  16. Krystofo says:

    P.S. I find very different data for 2003 on two different pages of your site. I am eyeballing the approximate average of your two lines for NVAC and NNWC.

    2003: Approx. 150 to 500 = +233%

    2003: Approx. 130 to 290 = +123%

  17. Jae Jun says:

    Yea I didn’t provide those tables like I did in the first post. But the more important point is that rather than looking at it from a long term graph, it’s more relevant to look at it via snapshots of 3 years which is more reflective of how an ordinary investor may go about doing it.

    As for the different results, I have tweaked the parameters since to make the screen a more realistic representation of the results. Trying to eliminate chinese companies and playing around with slippage and volume for an accurate simulation.

  18. fanfan says:

    I tried to use backtest.org to simulate ncav with the following rules :
    I did not found any stocks ?

  19. Jae Jun says:

    backtest.org has a learning curve but it should display results. Try one of the predefined ones and then adjusting the code from there.

  20. fanfan says:

    What do you think about “stock investor pro”, the software proposed by AAII ?

  21. Jae Jun says:

    It’s a good screener but there are no backtest capabilities.

  22. Jim says:

    Does backtest.org account for survivorship bias? For example, are these backtest looking only at companies that actually survived? (and eliminating all companies that went bankrupt – a real possibility with net nets)

    Ideally any backtest would be able to look at the companies that existed at the time, and not only the ones that survivor today

  23. Barney Stinson says:

    Are these for only stocks that trade on a stock exchange?

  24. the stocks from the article? Yes because OTC stocks were excluded.
    However, the screener page includes it for idea purposes.

  25. Yesrxx says:

    http://www.netnethunter.com has lists for 7 markets but theyre all english speaking markets.

  26. brett says:

    Does it make sense to just allocate a portion of a portfolio to a basket of NNWC stocks? Thanks.

  27. Yes that is something you can do also. But within that portion, you should still buy a diverse basket. It may eat up fees though.

Pick Winning Stocks and Fatten Your Portfolio