Sunday, January 25, 2009

The Importance of the Market's Opening Price

The market's opening price represents an important synthesis of information, including the trading of overseas markets and the impact of pre-opening economic and earnings reports. Simply locating the open relative to the prior day's trading range provides meaningful information to the short-term trader.

Burning midnight oil this weekend, fueled by caffeine and club music, I reworked my proprietary system for calculating pivot and price target levels. The results were satisfying: 75% of all days touch the daily pivot level, which is an approximation of the average trading price for the prior day. The Resistance levels R1, R2, and R3 represent upside targets; the Support levels S1, S2, and S3 represent downside price targets for the current trading day. About 75% of all trading days touch either R1 or S1; 50% touch either R2 or S2; and 33% touch R3 or S3, based on backtesting to the year 2000. (Daily pivots will be posted each AM prior to the market open via Twitter. Most recent posts appear on the blog page; subscription via RSS is free).

Interestingly, when the S&P 500 Index (SPY) opens above its pivot level for the day, the odds of it hitting R1 have been 56%. When SPY opens below its pivot level, we hit R1 only 21% of the time. Conversely, when SPY opens below its daily pivot, we've hit S1 about 59% of the time. When SPY opens above its pivot, we hit S1 only about 19% of the time.

Once we have the pivot levels in place, many other historical patterns jump out, including how the prior day's volatility affects hitting price targets in today's session and how prior weakness or strength affects the odds of hitting price targets in the future. One of the more interesting patterns for study is time-of-day analysis: how the odds of hitting R2/S2 and R3/S3 are affected by the time of day that the market reaches R1/S1.

All of this is fuel for future research and future posts. Bring on the caffeine and music.

.

13 comments:

Charts and Coffee said...

Interesting post. I use a lot of these techniques for setting my stops but have never quantified it.

dan said...

Why do you use the SPY and not the ES contract directly? Could you also provide the ES levels if it not too much trouble.

thanks,

dan

MarkB said...

Hey Brett,

Thanks again for your great work. I learn something from your posts nearly every day.

Have you ever looked into how longer time frame support and resistance, like 50/200 day MA's, large whole numbers, etc.) interact/interfere with your short term support and resistance levels? It would be interesting to know if the statistics around your support and resistance levels change when major longer term support or resistance is lurking nearby--if market participants shift their focus to those levels.

Mark

Don C said...

The biggest revelation in this post is Dr Brett's choice in music.
:)

Seriously, nice post.

Very true that once prices are framed (by pivots or any other R/S) many things stand out. And the volatility of prior day(s) definitely affect current prices (as in Crabel's range exp/cont). Sometimes, including a few more days may help, and rolling pivots with an x lookback can be useful.

On the proverbial other trading hand, the stickier rolling pivot levels may now lag runaway prices and you find yourself missing entries or leaving too much on the table.

E-Mini Player said...

Good stuff Brett! I personally plot the 15 min opening range, 1st hour high, 1st hour low, 2x opening range (for targets), the market's mid-point (average price), Prev day's high/low/close, Globex High/Low. I know it seems like a lot but this information really helps me stay on the right side of the market and identify areas of interest where I expect a buy/sell response. I'll make a detailed post about this later, and shoot you the URL.

Enjoy rest of the weekend!

E-Mini Player said...

Oh and check out http://www.pandora.com for more music :)

Zen said...

Great post!

Question - which calculation do you use for pivot levels? I know there are at least a couple out there.

Kevin said...

Zen: Question - which calculation do you use for pivot levels?

Excerpt from Brett's post: "I reworked my proprietary system for calculating pivot and price target levels."

Rick said...

Interesting work! Can I ask where you're getting your dataset for intraday data? I'm doing much simpler tests myself but only having access to OHLC historical data makes it hard to differentiate useful intraday moves. Thanks!

Brett Steenbarger, Ph.D. said...

Hi Dan,

As I mentioned in an earlier comment, I simply translate the SPY levels into ES levels since they are so correlated. I just use a ratio based on recent trading.

Brett

Brett Steenbarger, Ph.D. said...

Hi Mark,

No, I haven't looked at the possible impact of longer term moving averages, etc. I have looked at the impact of overnight range and prior day range and they are both key.

Brett

Brett Steenbarger, Ph.D. said...

Hi Don,

Always happy to share the tunes... :-)

One reason I've included R3 and S3 levels is to capture those runaway markets when they occur. Thanks--

Brett

Brett Steenbarger, Ph.D. said...

Hi Rick,

I get my data from e-Signal and download to Excel--

Brett