13 replies to this topic

[mortiz]

xD, In all due respect, although you have made some great market calls, your comment about days with over 95% down volume having negative long term price ramifications is simply not the case if you look back to 1950 going forward and examine the average returns for various time intervals following a 95% or greater down volume day. I will follow up with a more detailed analysis later this week, but by selecting arbitrary time windows, 1 month, 3 months, 6 months, etc, following a 95% down volume day, the average returns over the past 57+ years (58 qualifying events) are all positive except the 3 month window with an average loss of 1.1%. The average returns, as measured by the SPX, are all greater than 10% (up to 24%) for the 6 month, 9 month, 1 year, and 18 month windows. There were some events that did fit your negative returns statement, up to 24% losses, but of those 58 events, counting clusters of multiple 95% down volume days within one week as a single event, there were only 11 of those 58 events having at least one time frame (between one month and 18 months) greater than a 10% loss. I will conduct more rigorous due diligence for the impact of 95% down volume days in the longer term, but a five minute drive by analysis suggests the opposite is true, the vast majority of such events results in positive gains, on balance, going forward. If you are only including the April 2000 and October 1987 events, the longer term negative impact would be true, although longer term in the 1987 event would have be less than six months. However, there are far more 95% down volume events than those over the years that were followed by exceptional upside price gains. Respectfully, Randy N.

[A-ha]

Randy,
I saw your statistical study on trin days above %5, which is not the same but related to what we are talking about.
I know most of the cases resulted with an upside follow through. You generalized them as wash out days which is not correct.
As you know, this one didn't produce the expected results yet as it is still going down.
However this one is tracking 1987 very closely in terms of behavior not percentage wise.

So May I suggest one thing:

Could you discard the days that occurred after lengthly (multi-week) downside moves because they are capitulation days not initiation and they are naturally followed by upside reversal. What we are seeing here is not anything close to it. This event occurred after a lengthly upside , almost right off the new highs.
Such power moves usually implies a trend change over the long term.

You are right about cluster days. Do you remember those three %90 up days during the summer sell off? If this sell off is now retraced some and then followed by more %90 down days, will you make the opposite case of what we saw in the summer?

In short this is the situation that is likely similar to 1987 and 2000 imo. However like I said, it is entirely possible to test the highs or even to make new marginal highs if the market can stabilize without further damage, which will be part of the long term topping process.

Edited by xD&Cox, 04 March 2007 - 07:47 PM.

[mortiz]

XD,

Thank you for the filtering suggestions in analyzing the historical extreme down volume days, your ideas will be implemented. Since I have a day job unrelated to the markets, it may be next weekend, but the study will be conducted.

The motive was not 95% or more down volume days, but a study posted on Technical Watch today addresses 5% declines following 6 month price highs along with other filters, and can be found at this link if interested:

Study

As far as last week's extreme TRIN print, those are so rare one cannot apply much statistical signficance to such small sample spaces. I have since lowered the bar on the TRIN study to days above 5.00, but the sample space is still too small. The returns data are what they are, and one must take it FWIW. My intent is to only lay out historical precedents for those who are interested in trivia.

Without a doubt, there will be lower price lows going forward, how much is TBD. The internals MCOs for nearly every index I monitor, and that is about all of them and then some, are buried in the mud. Such negative MCO extremes almost always lead to retests of the price lows at the time the MCO extremes are posted, and usually those price lows are exceeded to some degree.

One glaring difference between the current market action and that of 2000 and 1987 is the level of the NYSE McSum. Both of those examples unfolded with the NYSE composite AD McSum well below zero. Currently, according to my data, the NYSE AD McSum sits at +690, a level that has insulated the market from crashes in the past. Of course, with the MCOs currently at extreme negative levels, the McSum could plunge to below zero within several days, and if the NYSE McSum falls below -250, the probability of a crash (over a 15% decline) would increase, IMHO.

Every effort is made to let the market decide position action, but until I see more serious degradation in the internals, the short side of this market will remain shorter term positions, and is the side of the market to be on at this moment. Longer term, I remain a bull until the market tools that have served well in the past say the bull is over... the market is always right.

Thanks again for the filtering suggestions, they will be used, but I will likely post the study on Technical Watch.

Randy

[A-ha]

Randy, I said this before on TW but I want to repeat here on TT. I always appreciate your comments and analysis and I learned from them. Thank you.