Skip to content

New Livermore Active Issues Index Hedged (Nasdaq 100)

March 11, 2010

Following the series of posts on the impact of volume, I decided to integrate this factor into the ranking methodology for the Livermore Active Issues Index. This change produced a significant increase in returns, and also produced arguably better risk-adjusted returns as well. The final index methodology is now even more similar in spirit to what Jesse Livermore would have looked for in the “leading stocks.” Not only are the index constituents a group of high-momentum and smoothly trending stocks, but they also show substantial volume activity relative to other stocks in the Nasdaq. The chart above shows the top 10 ranked stocks in the Nasdaq100  rebalancing with a 5-day holding period. Note that the avg trade statistics are distorted by having to exit and re-enter every 5-days. The turnover was actually much less frequent and the trade gains were higher. These stocks are hedged 100% of the time using the QQQQ throughout the entire backtest. No timing is employed on either the short or longer term time frames. Preliminary testing shows that using the 50 and 200 day moving averages reduced risk for both the individual stocks as well as the hedge position in the QQQQ.

16 Comments leave one →
  1. John French permalink
    March 11, 2010 7:34 am

    David, it’s interesting to see the effect of the Volume ROC. Is this simply defined as (Todays Vol – Vol five days ago)/Vol five days ago?

  2. John French permalink
    March 11, 2010 9:38 am

    Just to clarify will the post of symbols for the Livermore Active Issues Index incorporate this ROC change from hereon?


  3. david varadi permalink*
    March 11, 2010 11:55 am

    hi john, yes that is correct–volume roc uses that calculation. also, i will have two lists–one will be the old method and one will be the new just to keep tracking consistent.


  4. March 11, 2010 12:35 pm

    Can you explain how you are hedging in this example: equal dollar, volatility or cap weighted,or something else. I assume you are using an inverse, nasdaq 100 etf, not shorting the QQQQ. Single or double beta?

    We use a 50 day MA with our volatility based hedging and it has improved results. We use futures instead of ETFs for the hedging.

    • david varadi permalink*
      March 11, 2010 8:24 pm

      hi jerry, this example is very simple and just involves dollar matching the stocks with an equal position short the QQQQ. ie 100,000 in the stocks and -100,000 in the QQQQ for a $100,000 account. the dollars in the stock vs the hedge are rebalanced every week. your suggestions are definitely more in order of what would be more practical — volatility based hedging is a must, and dynamic hedging does test out well but is obviously more of a timing approach. indeed the futures are more cost effective.

      thanks and good suggestions,

  5. Kevin Smith permalink
    March 11, 2010 2:00 pm


    Are you rebalncing on Friday morning, Friday at close or on Monday mornings?



    • david varadi permalink*
      March 11, 2010 8:26 pm

      hi kevin, we didn’t have a fixed day for rebalancing in the test–just every 5 days and we tried various days to start. the actual livermore is on the site on fridays close to the close.

  6. March 11, 2010 3:25 pm

    Good work !! I had read another post here about using the AAII screens which I tried to use in the early 90’s but had trouble with . Many of the stocks on some of those screens had large bid and ask spreads so the profit was eaten up/ losses magnified during rebalancing. It would be productive to try to avoid this on the Livermore selections ( ie. select reasonably liquid trading volume ).

    • david varadi permalink*
      March 11, 2010 8:28 pm

      hi thanks con, the issue with the AAII screens is that they often contain smaller cap names as you are suggesting. the current list is focused only on Nasdaq100 stocks which have penny spreads and strong liquidity for the most part. so this issue is not a factor–we have tested with various commission assumptions in our background work. furthermore, the volume screens in the ranking further ensure that these stocks are fairly active.


  7. Kevin Smith permalink
    March 11, 2010 4:01 pm

    Preliminary testing shows that using the 50 and 200 day moving averages reduced risk for both the individual stocks as well as the hedge position in the QQQQ.

    David – are you saying that requiring the 50 ma > 200 ma improved returns further? If the 50 ma < 200 ma, would you simply hold cash for that 10% allocation or drop down the list to the the 11th ranked postion?


  8. david varadi permalink*
    March 11, 2010 8:29 pm

    hi kevin, i think that dropping it to cash or replacing with a stock that had the 50ma>200ma is a good idea for sure.
    thanks a lot,

  9. kostas permalink
    March 12, 2010 1:10 am

    Hey David, is your list of NDX free of survivorship bias (does it reflect the index reconstitution changes across history) or is it based on today’s constituents projected back as a group to 1999? Good work either way!

  10. david varadi permalink*
    March 12, 2010 1:14 am

    hi kostas, we went back and updated the names each quarter going back in time which very closely approximates the true NDX. thanks

  11. Kevin Smith permalink
    March 12, 2010 9:17 am

    David – giving more thought to using a moving average to filter both Livermore and short QQQ positions, it seems that if you could end up with a portfolio that is either all long with no hedge or all short with no longs. I could see this being an issue at trend reversal points like early March 2009. At that time, I’d bet that all of the NDX 100 was below their 200 day ma, and an inverse QQQ etf was above. Upon reversal, you’d have no longs and a 100% short position until later in June. Perhaps using a couple of ma’s to stage into and out of positions. i.e. –
    a) always maintain 1/3 positions in both long & short
    b) increase both long & short positions to 2/3 based on the 13/34 ema
    c) increase both positions to 100% based on 50/200 ema
    d) scale back using same criteria

    Yours or others’ thoughts appreciated.



  12. Josh permalink
    June 10, 2010 12:00 am

    Does this test include transaction costs and slippage assumptions? Or is this more of a proof-of-concept? Amazing results either way!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: