## Tuesday, October 30, 2007

### Pivot Level Support and Resistance: How Often Do We Hit Those Price Targets?

In recent posts, I have examined the odds of hitting high and low prices from the previous trading day and the odds of closing opening price gaps and hitting average trading prices from the prior day. This post will conclude a survey of intraday price behavior by examining how often we hit pivot-derived support and resistance points.

For the purposes of calculation, the pivot price is defined as the average of the previous day's High, Low, and Closing price. As with the earlier posts I am using the S&P 500 Index (SPY) going back to 2004 (N = 963 trading days) for my calculations.

The R1 resistance level is defined as (Pivot Price *2) minus the prior day's low price. The S1 support level is defined as (Pivot Price *2) minus the previous day's high price.

Thus, let's say that SPY closes at 150. Its high price is 151 and its low price is 149. The Pivot Level would be (151+149 +150)/3 = 150. The S1 level for the next day would be (150 * 2) - 151 = 149. The R1 level for the following day would be (150 *2) - 149 = 151.

Going back to 2004 (N = 963 trading days), we find that SPY hits either the S1 or R1 levels on 779 occasions, or a little over 80% of the time. We hit R1 on 490 occasions (a little over half the time), we hit S1 on 425 occasions (about 45% of the time), and we hit both levels on 135 occasions (about 15% of the time).

Thus, we tend to hit one of the pivot-derived levels during a given trading day, but which one we hit--on the basis of price alone--is pretty much a crap shoot. We'll want to look for indicators that help us effectively handicap the odds of hitting these levels.

Finally, let's look at the frequency with which we *close* above or below R1 and S1. Of the 490 occasions in which we hit R1, we close above R1 on 244 occasions, or about half the time. Of the 425 instances in which we hit S1, we close below S1 on 212 occasions, again about half the time.

Clearly, then, it is more common to trade above the R1 resistance and S1 support levels than to close above them. If we think of R1 and S1 forming a trading range, it will be helpful to have indicators that help us handicap the odds of closing outside that range vs. returning to that range.

In all, going back to 2004 (N = 963 trading days), a total of 455 days closed outside R1 or S1, a bit less than 50%. So about half of the time, we will close within the range defined by R1 and S1.

In my upcoming posts in this series, we'll begin the search for indicators that help us handicap the odds of hitting and closing above these benchmark prices:

* Previous Day's High
* Previous Day's Low
* Previous Day's Pivot Level
* Opening Gap (Previous Day's Close)
* R1 Resistance Level
* S1 Support Level

If we can find measures that assist us in handicapping those odds, we'll be able to trade more like a card counter, placing bets when probabilities are favorable and standing aside when they're not.

RELEVANT POST:

.

ainkurn said...

thanks for another great post. i plot support and resistance levels on the YM overnight session every morning before beginning my trading day. i don't really use the technical pivot point calculations (mostly breakout price levels coupled with volume levels at certain prices) but I will take another look at pivots after reading this post.

www.10kthrownaway.blogspot.com

As a firm believer in the power of pivots, I'm curious why you did not include R2 & S2 in your study. While I don't have Odds Maker software, I do closely monitor the pivots every day and expect that the probability of reversals off these levels is significant.
As an aside,I would also recommend another resource for appeciating pivots and that is Mark Fisher's book THE LOGICAL TRADER, which integrates the concepts of position sizing and pivot ranges.

F. said...

Another great article. Thanks, Dr. Brett.

F.

Jeff Pietsch CFA, Esq said...

Watch out for loosing streaks in this strategy! I realize that is part of the point of this series on a daily basis. Just that I've seen it go very wrong for months at a time.

Jeff said...

Thank you for another great post! Have you considered taking into account, in your analysis of pivots, where the open is relative to the pivot point? In other words, compare a case where the open is at or near the pivot, vs. a case where the open happens to be near, say, R1?

BTW, it seems to me that pivots seem to "work" better for the ES than for the YM. Do you have an opinion about that?

thanks again

Brett Steenbarger, Ph.D. said...

I'm jumping ahead of myself, but my research has not found as good an edge in hitting R2 and S2 as in the other levels. But I may post something about those target points as well at a later occasion. Thanks--

Brett

Brett Steenbarger, Ph.D. said...

Hi Jeff Pietsch,

I tried clarifying this in a subsequent post: these stats aren't meant as a strategy, just as a summary of baseline odds of hitting various points. A valid strategy would have to significantly improve these odds.

Brett

Brett Steenbarger, Ph.D. said...

Hi Jeff,

I have not studied YM vs ES re: pivot-derived targets; that would be interesting. And yes, it makes sense that where we open would affect the odds of hitting various target points. I may take a look at that shortly; thanks--

Brett