Last Couple of Notes on DV(2)

17Jul09

Last couple of notes on the DV(2) indicator that we’ve been exploring this week (read more here and here).

1. I’m adding the strategy that we tested in our first post (unbounded DV(2) above/below zero) to the State of the Market as a short-term indicator. Expect to see its impact in tomorrow’s report.

2. I received more than a few reader requests for an example showing how to calculate DV(2) (my initial description was admittedly half-assed). Click for an example in Excel. Note I also included two of the (frictionless) strategies I showed on the blog: unbounded DV(2) above/below zero, and bounded DV(2) below 10/above 90.

[Edit: click for a summary of all posts in this series on the DV(2) indicator]

Happy Trading,
ms

 

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.

Add to FacebookAdd to DiggAdd to Del.icio.usAdd to StumbleuponAdd to RedditAdd to BlinklistAdd to TwitterAdd to TechnoratiAdd to FurlAdd to Newsvine



12 Responses to “Last Couple of Notes on DV(2)”

  1. 1 Tom

    Michael –

    Feeling like a real dope, but using your spreadsheet to calculate DV values based on SPY Friday close (High: 94.32, Low: 93.54, Close: 94.13) I get an unbounded DV of -0.2. But your state of the market report says the DV is above 0. Any idea why there’s be a discrepancy?

    Thanks

    Tom

    • 2 marketsci

      RE to Tom: nope, we’re the dope. Bug in the program. The “prediction” was correct, but the description is reversed. We’ll get this fixed on the next go around. michael

  2. 3 Andre

    Michael –

    I am interested in allocating a portion of my investment portfolio to some of your strategies, but I would prefer to handle the process directly, as opposed to using a managed account.

    However, I am not always available around market close, so I am interested in whether you (or any of your blog readers) are aware of some sort of interface that enables the user to enter conditional market-on-close orders that are transmitted to Rydex, ProFunds or Direxion (the mutual funds for which YK and Scotty were designed).

    If no one is aware of anything like that, then are you (or any of your blog readers) aware of an online broker that enables the user to enter conditional market-on-close orders for ETFs?

    – Andre

  3. 5 djg

    I have tested the DV(2) and the scaling RSI(2) strategies with a $10 commission (but without other costs of shorting) and I am getting a result where the strategies are good moneymakers at the volume of about 1000 shares per trade, ok moneymakers at 100 shares, and start losing money at 50 shares. Volume appears to be key so as to not swamp the gains with trading costs. I used the TD Ameritrade tool for testing. NB: this tool backtests pretty sophisticated strategies but does not generate Sharpe ratios or allow enough data to easily calculate it.

    • 6 marketsci

      RE to djg: good stuff – a couple of comments:

      1. Stating the obvious here but it’s important for investors to test any strategy using their specific asset traded using their specific expected transaction costs. These are very trader-specific factors. So I’m always happy to see folks do just what you’ve done above.

      2. The painful drip of transaction costs is one of the biggest reasons I trade leveraged mutual funds.

      3. Someone attempting to trade ETFs in small size needs to use a strategy much more sophisticated than the simple one I provided in this post because they really need to target trades at those specific days where probability says they can overcome transaction costs.

      Thanks for the comment, michael

  4. 7 John French

    Michael, using NDO on NDX data (I note your comments vs ETF’s etc.) going back to 1985 (Yahoo data) I average ~-60%pa (thats a minus) up to and including 1997. Thereafter it’s off to the races i.e. ~+65%pa. 8th Sep 1998 appears to be the day of the reverse in the models fortunes.

    I am still no closer to finding what one can use to indentify the move from follow-through working versus mean reversion other than having eliminated a few ideas. I would like to think I could use something other than just simply monitoring results a la Equity Curve trading. Maybe the answer lies in some kind of phasing in/out strategy.

    John

    • 8 marketsci

      RE to John: good stuff sir. I’m oversimplifying a bit, but I think trading the equity curve using a phase in/out (or what I call a “confidence-based” approach…vs a binary approach) is about the best I’ve come up with so far. I think even if one could pinpoint the exact reason that daily MR went into effect late last century (which I don’t know if we ever really can), it would be moot – the next evolution is going to be caused by something entirely different. Having said all of that, damn I’d like to be able to put my finger on the cause of that MR jump in ’98. michael

      • 9 Jim

        That “confidence-based” factor isn’t part of RH, is it? Needs to be.

        Jim

      • 10 marketsci

        RE to Jim: I’m assuming you’re referring to RH’s abysmal performance the last few months. Here’s my take on programs that come up through our Timer Seeds program (RH being the only one that I talk about on this blog): it’s important that I don’t impose too much of my own opinions about strategy development the developer. If I did that, the strategy would start to look more and more like what I’m already doing and we would lose some of that diversification that inherently comes from different minds approach the same problem from different angles. Having said that, yes, something like the “abnormal market filter” that I use in YK and Scotty and have talked about quite a bit on this blog would have blunted a lot of RH’s recent pain. michael

  5. 11 John French

    Michael, what you said above “I think even if one could pinpoint the exact reason that daily MR went into effect late last century (which I don’t know if we ever really can), it would be moot – the next evolution is going to be caused by something entirely different.” is typical of why I like this blog so much. Nothing like a dose of true reality to temper those fantastic looking back-tests.

    John


  1. 1 DV2 Indicator for Amibroker « Quanting Dutchman

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com 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