Do Very Bad Fridays Set Up Crash Mondays?

Many traders who are aware of the history of the ’87 crash may often think after a bad Friday, “Will this get substantially worse on Monday? Are we setting up for a crash like ’87?” It’s an interesting question. Was 1987 an anomaly or does a really bad Friday often carry through into the next week? Below I looked at all Fridays since 1960 that closed down at least 2.5%.

The “Average Trade” column on the far right is skewed thanks to the ’87 crash which saw the market drop 20% on Monday. It appears in the almost all of the cases that the market was set up for a bounce based on Friday’s action rather than a crash. Of course while the last week has been bad, the market does remains in a long-term uptrend. I decided to filter the above results again to examine the bad Friday’s that appeared in long-term uptrends.

Instances are low here, but for the short-term they really couldn’t be more bullish. Again they also suggest the bounce should basically come immediately.

Extreme Weakness Never Before Seen By This Measure

The McClellan Oscillator uses advance/decline data to calculate the strength or weakness of a move from a breadth standpoint. The value will vary from provider to provider as there are often slight differences in advance/decline data. Worden Bros. is one data provider I use. Their measure of the McClellan Oscillator hit -381.49 on Wednesday. This is the lowest reading since they began tracking advance/decline data in 1986. (Others I look at are low but not quite all-time lows.) Below is a chart of the McClellan Oscillator over the entire data period.

One notable about this chart is that breadth readings have become more extreme over time. Whereas moves above 100 and below -100 were rare from ’86 – ’93, they are fairly ordinary today.

For more information on the McClellan Oscillator you may visit the link below:

Weak Closes Since The March Bottom

Monday’s selloff saw the market close poorly and near its lows for the day. Below is a study that examines weak closes since the March bottom.

I consider this particular study to be environmental. It is indicative of the strength of the rally of the last 7 months and not necessarily an all-weather setup. While I wouldn’t base a trade on this study I do think it will be important to see how it plays out over the next few days. An all out failure to bounce could suggest a change of character for the market.

Back to back 7-day reversals – a rare setup

More of an oddity than a quantified edge this morning…

The last two days we’ve seen opposing reversals. Wednesday the market made a new high but closed down on the day. Thursday it hit a 7-day low before reversing to close up on the day. A reversal off a 7-day high followed by a reversal off a 7-day low would seem a bit unusual. I looked back to 1978 and found out just how unusual it was. Below is what I found.

It’d be dangerous to trade based off of just a sample set of 5, but I was still fairly amazed that there wasn’t a single instance of a profitable close within the next 4 days.

What Happens In Vegas…

The International Traders Expo takes place at Mandalay Bay in Las Vegas on November 18-21, 2009.

I’m pleased to announce I’ll be speaking on Thursday, November 19th at 1:15pm. The topic of my presentation will be “Quantifiable Edges for Swing Trading”. I’ll be discussing some of my favorite edges and most interesting research.

Registration for the event is free, and you may do so by clicking here.

I hope to get the opportunity to meet many readers and subscribers at the Expo.

The Last 4 Days Price/Volume Pattern

Price/volume the last 4 days has done the following. Thursday the SPX closed at a 50-day high on lower NYSE volume. Friday SPX closed lower and NYSE volume rose. Monday we got another 50-day closing high on lower NYSE volume. Tuesday another market drop with rising volume. That certainly sounds like a bearish price/volume pattern. I took a look.

Going back to 1970 I was only able to find two other instances with the same 4 day pattern where 50-day highs were being made. The 1st was 3/26/81 and it was followed by a decline of nearly a year and a half. The 2nd instance was 6/6/95 and that was followed by a 3-day consolidation and then a continuation of a massive bull market. Nothing to learn there.

But what if we look at the 4-day price/volume pattern on its own and not require new highs be made? Based on common knowledge it would still seem to be bearish. Below are stats going back to 1970:

It could be argued that the above results suggest bullish tendencies, especially over the 4-7 day period. I don’t see any evidence that suggests the current 4-day price/volume pattern is bearish.

Quantifying the Value of Historical Research

The most common type of post here on the blog is one where I’ll show a setup along with a statistics table examining how the market has performed based on similar setups in the past. On the blog, most studies are examined independently. In the Subscriber Letter I’ll take a holistic approach to viewing the studies. The tool I use to do this is the Quantifiable Edges Aggregator.

The Aggregator takes a measurement each evening that estimates what all of the currently active studies are projecting over the next few days. This number is plotted and used on the Aggregator chart, which is published each night in the Subscriber Letter. Along with the estimates the Aggregator chart also shows how the market has performed relative to expectations over the last few days. This is helpful in establishing whether the market is overbought or oversold. I have claimed substantial upside edges typically exist when expectations are positive and the market is oversold versus recent expectations. Also, substantial downside edges typically exist when expectations are negative and the market is overbought versus recent expectations.

While many of the index-oriented trade ideas in the Subscriber Letter were based on the Aggregator chart, I’d never quantified the Aggregator nor used it as a mechanical entry…until recently.

Now that we have nearly two years of historical values I decided it was time to take the concepts above and show exactly how a mechanical strategy based on the Aggregator would have performed. The results were even better than I expected.

Since 2/25/08 (about 20 months), the reinvested return (not inclusive of commissions, slippage, or interest on cash) of trading the SPX based on the Aggregator System signals would have been 106.34%. The system has struggled more recently and is currently experiencing a 3.94% drawdown. Over the full time period it has been invested a little over 60% of the time, with the remaining 40% of the time spent in cash. Both long and short trades have contributed fairly equally.

In my mind, the success of the Aggregator as a tool and as a predictive indicator has cemented the value of historical quantitative research. It demonstrates that incorporating quantifiable edges does indeed provide a quantifiable edge!

More details about the Aggregator System may be found on the systems page of the Quantifiable Edges website. Details include an 11-page working document that reviews the results, discusses recent performance and evaluates alternate entry and exit techniques. Additionally there is a spreadsheet available to all trial users that shows summary statistics, an equity curve and details of every single trigger since 2/25/08 (when the Subscriber Letter began).

Gold level subscribers are able to download the full history of the Aggregator and Differential values in a .csv file. This allows them to more easily integrate the tool into existing strategies or to build their own strategies based on it.

Anyone who wishes to trial the Quantifiable Edges subscriber services may do by clicking here. If you have previously trialed or subscribed to Quantifiable Edges, but would like the opportunity to trial again and see details of the Aggregator System, feel free to drop an email to support @ quantifiableedges.com (no spaces) and you will be set up with a new 1-week trial.

Strong Drops From Highs

When strong moves down occur from high levels as happened on Friday, there is often a bit more downside follow through. Below is a study that exemplifies this.

Certainly not an overwhelming edge, but a hint that there could be more selling before a bounce occurs.

BTW, watch out this week for a few exciting announcements from Quantifiable Edges!

Large Gap & Go’s To Intermediate-term Highs

Wednesday’s move may look especially strong on a chart. Historically when large gaps continue higher intraday and make new intermediate-term highs it has most often led to a pullback over the next few days. Below is a study that examines this.

Instances are a bit low but notable nonetheless. Below is a list of all the instances using the 3-day exit criteria.

This would appear to suggest a bit of a downside edge over the next few days.

Columbus Day Performance

While the markets are open on Monday, banks, schools, and government offices are closed. In past years, with the bond market closed, the stock market has done quite well on Columbus Day. Last year Columbus Day saw a gain in the S&P of over 11%. Many traders will likely recall it was the week prior that the market suffered its multi-day crash. Stats associated with Columbus Day are now a bit skewed thanks to 2008. Below is a chart from 1961 – present that shows Columbus Day performance.

As you can see it has generally been a positive holiday for the market.

Considering This Morning’s Gap

The SPY is looking to gap up over 1% as I write this. It has also closed higher 3 days in a row. Back in April I looked at large gaps up after the market ha already risen. This study suggests risk/reward favors the downside should SPY gap more than 1% this morning. The fact that the SPY closed within a range rather than at a 10-day high makes it a little less encouraging.. Overall, should this gap hold until the open, I would estimate a mild edge to the downside from open to close.