Monday, January 19, 2009

Swing Trading With Weekly Price Targets

I recently wrote about how I use empirically-derived price targets to trade intraday moves in the S&P 500 Index. This approach builds on the pivot-based methods I described two years ago, in that the price targets are a joint function of recent price movement and recent volatility. I have tested these levels going back to 2000, with roughly a 75% hit rate for touching yesterday's pivot level in today's trading and a 75% hit rate for touching *either* the upside R1 or downside S1 levels derived from yesterday's trade. An example of how I use these levels can be found in this recent post.

Although my methods for calculating these levels remain proprietary for now, I post the levels for SPY each morning via Twitter prior to market opens (free Twitter subscription via RSS). Also included in my morning posts are data that capture strength/weakness in the previous day's trade. These data include the percentage of SPX stocks trading above their moving averages; Demand/Supply (an index of the number of stocks trading above/below the volatility envelopes surrounding their moving averages); and the number of stocks making fresh 20-day highs and lows.

As a rule, in a strong and strengthening market, I'll look for price to stay above the pivot level and test R1 and R2 levels. In a weak and weakening market, I'll look for price to stay below the pivot level and test S1 and S2. In a range bound market of mixed strength, I'll look for prices to revert to their pivot levels on moves toward R1 and S1. The idea is to use the ongoing stream of market data (volume/volatility; leading sector behavior; market themes; cumulative TICK) to gauge the odds of hitting these price levels and then enter the market at points that provide favorable risk/reward (e.g., your stop level is closer than your target point).

Often, not always, one level will serve as a target (say, S1) and another (pivot) will serve as my stop. In other words, if we're in a short-term downtrend, my trade says we should hit S1 and stay below the prior day's pivot (which is an approximation of average trading price). Many trade ideas can be crafted by knowing these levels, assessing the market's strength/weakness day over day, and gauging the strength/weakness in the current session's data.

Suppose, however, you are a swing trader looking to trade less frequently and take more out of market moves. Such a trading style is ideal for those that don't want to be married to the screen intraday. I've been working on an adaptation of the above trading methods for the wider timeframe and now have a backtested set of parameters based on weekly data. The weekly pivot has an 80% hit rate (i.e., going back to 2001, the current week has touched last week's pivot 80% of the time), and the odds of hitting either the weekly R1 or S1 levels is about 75%; R2 or S2 is 50%.

Once again, by gauging market strength/weakness day over day, the swing trader can play for multi-day moves to the R1/S1 levels and beyond. The weekly target numbers also enable short-term (intraday) traders to leave a piece of their positions on overnight to take advantage of the moves to the weekly levels.

These weekly levels in SPY will be published Monday before the market open and, again, will be free of charge via Twitter. For traders that prefer to not subscribe via RSS, the last five Twitter posts always appear on the blog page under "Twitter Trader", so you can simply check the blog prior to market opens for the target data. Over time, I plan to expand the targets to other indexes (NASDAQ 100, Russell 2000, sector/international ETFs), as well as other asset classes (bonds, gold, oil, etc.). Stay tuned!


netbsd said...

Hi Dr Brett,
i woul like to ask you, what is your opinon on why the Nasadaq Composite Tick ($TICKQ) has a quite persistent positive bias (very rarely goes ander zero, and when it does, it does it only by small values. The NYSE Tick is by far, more dynamicly oscillating regularly beetween extremes).
And for this reason do you think that the NYSE Tick col be more indicated to asses the intraday NASDAQ 100 sentment than the $TICKQ is?

miramar said...

Would you mind letting us know how to compute the ratio of SPY to ES? I tryed to use 10 X SPY but it seems a little off. Also, is this ratio a moving target? Thanks for your heip, I enjoy your blog.

GS751 said...

I just wanted to stop by and say even though I dont comment much I really enjoy your twitter updates and read most of the links you throw in there.


Kevin said...

miramar: Would you mind letting us know how to compute the ratio of SPY to ES?

I'll get you started: the value of an ES contract is $50 times the current price.

Some other factors to consider might include the SPY going ex-dividend and the rollover of the futures contracts.

Jorge said...

Dr. Steenbarger,

You're truly a gold mine. Pivot points for Bonds? Yes!

(I understand that the ratio of SPY to ES is simply calculated by dividing one by the other, and that it fluctuates slightly over time, if not, please advise).

Best trading,


Brett Steenbarger, Ph.D. said...

Hi Netbsd,

I can't say I've studied $TICKQ extensively, but my data don't show the persistent positive bias that you mention. I'd encourage you to check on the accuracy of your data source--


Brett Steenbarger, Ph.D. said...

Hi Miramar,

If you take the one minute closes of SPY and ES, you can divide ES by SPY and arrive at a ratio factor for each minute. Take the median value of those ratio factors over the past day of trading and that can serve as a conversion factor to translate SPY into ES.

For a quick and dirty estimate, simply pull up one minute charts of SPY and ES and look for bars where SPY traded at the price you're interested in and see where ES traded at those corresponding times.


Brett Steenbarger, Ph.D. said...

Thanks, George; I appreciate the interest and support--