Does that not directly contradict E&M?
Yes, I believe it does.
Do you have any further details on this?
Yep.
First you'll have to suffer through my opinion though. <g>
E&M, Murphy, and all the other TA "greats" that I've read, talk about "minimum targets" and such for patterns without ever, to my knowledge, backing up their assertions with DATA.
Now, Bulkowski, on the other hand, is not perfect in a lot of ways, but at least he gives you the data from which he makes his conclusions. So, here is some:
1. B studied 500 stocks from mid-1991 to mid-1996. This was, I think we can all agree, a bull market, but not yet to the mania phase. Nonetheless, I would intuitively expect that bullish patterns studied during this period would somewhat over-perform while bearish patterns under-perform.
2. He studied 431 H&S patterns. He also studied 141 "complex" H&S patterns, (which he deals with in a separate chapter). I only mention the complex H&S patterns because when I annotated the chart I showed you I accidentally (or maybe on purpose -- it was a long time ago, so I can't remember), used the stats from complex H&S patterns rather than "normal" H&S patterns. I have used the stats from "normal" H&S patterns below.
3. The 431 "regular" H&S patterns divide up as follows:
Upside Breakouts = 9 (2%)
Downside Breakout Failure, i.e., failure to drop at least 5% below the neckline, = 21 (5%)
Successful Downside Breakouts = 401 (93%)
Stats on Formations Which Broke Out Downward & Greater than 5%, i.e., "Successful Formations"
4. Expected "Measure Rule" drop = height of formation subtracted from the price at the breakout point.
Formations which met or exceeded the "Measure Rule" = 254 (63%)
5. Average Decline of Successful Formations = 23% (actual)
6. Most Likely Decline of Successful Formations = 20% (using a frequency distribution)
7. Pullbacks = 45%
A smaller absolute number of H&S would fall 50% and beyond, therefore lowering the mean. It implies to me that one would should normally expect declines of only 20% to 27%, with 50% being the exception that skews the average.
That is my understanding, except that I have to correct the numbers I gave you, so I think we should "normally" expect a 20% decline, with a few large declines bringing the average up to 23%.
So far so good, right?
BUT, remember, Bullkowski performed his study on INDIVIDUAL STOCKS DURING A BULL MARKET.
QUERY: How are we to relate his data to a formation which appears on an INDEX and marks the transition from a bull market to a bear market?
(In case you're wondering, I haven't a clue.)
if the most likely drop is implied to end at 720 and given that the October lows ended at 708, one could make a bullish case that the H&S pretty much fulfilled its measurement implications in October.
Yep, one certainly could, with a large caveat and some corrections.
First correction: The October low was 768, not 708, but the "six" looked like a "zero" because it was covered by a horizontal line on my chart.
More corrections: When I adjust the numbers and use "normal" H&S stats instead of "complex" H&S stats, the 20% target becomes 756, and the 23% target becomes 736.
Thus, it appears that the formation was nearly "spot-on" it's 20% most likely target, and just a tad above it's 23% average target. So, we're "there," right?
Well, maybe... (Here comes the caveat.)
BUT!
Let's not forget one thing. 63% of H&S patterns in Bulkowski's study made or exceeded their MEASURE RULE targets. That means, (I think), that there is a 63% chance that this formation is supposed to drop 618 points from the neckline of 957, i.e., to about 347.
How the heck do we deal with that?
Well don't look at me, I don't know! <G/NG>
FWIW, here's the corrected version of my chart.
stockcharts.com[g,a]maclynay[pf][J5062179,Y]&listNum=46 |
