Extremely negative breadth and extremely low volume is an unusual combination. On Monday we saw this combination. Below is a quick study which looks at it:
Not a huge edge, but it appears to suggest we should see a respite and a bounce sometime soon.
Last week I looked at trending vs. chopping tendencies for the market in both the daily and intraday timeframes. What I found is that in both time frames the market has evolved from trendy to choppy over time. Today I am going to look at weekly bars. I will also have some further follow-up later this week which look at the statistics of some of the “Trend vs. Chop” results.
I’m going to use the same “system” as last week to display weekly tendencies. The first test here buys on the close of any up bar and reverses to short at the close of any down bar. Like last week, the trades were at a constant $100,000 per trade for the entire period. No commissions or slippage are figured in to the test. A rising graph would indicate follow through is dominant while a falling graph would indicate reversals are dominant. Sideways action would suggest no tendency.
Unlike the daily and intraday time frames, which favored trendiness through much of their history until fairly recently, the weekly timeframe has been more prone to chop since the early 70’s.
As I did last week, let’s now break that down into upside follow-through and downside follow-through. You’ll see a large difference here. The chart below looks at the results of shorting down weeks.
Now let’s look at what happened if you looked to buy strength:
Upside follow through was strong through the 60’s and until January of 1973. That is where you see the peak near the left hand side of the graph. Since that time the market has gone through cycles that sometimes favored buying strength and sometimes favored selling strength. Thirty-five years later the equity of the system is right back near its peak 1973 level. It should be no surprise that the current spike down began in 2007.
These results shouldn’t come as much of a surprise to people who have spent time evaluating or developing swing-trading systems. When trying to take advantage of 3-5 day moves, it is generally easier to develop long-only strategies that perform relatively well over time than to produce short-only strategies that perform consistently well. Hence a big reason you see so many more long-only systems. Of course in 2008 so far shorting strength has been easier than buying weakness.
One relatively new blog I’ve begun to read and enjoy is MarketSci. Over the weekend I decided to check out some the past posts and studies on his blog. I was surpised to find one post eerily similar to my own Trend vs. Chop study that he published in July. Michael’s findings agreed with my own with some additional twists. He also looked at results above and below the 200-day moving average, and he looked at bonds as well.
I’ll be expanding the study again and publishing some results based on weekly bars a little later.
MarketSci also had an interesting post last week which expanded nicely on my recent SOX studies suggesting the SOX can act as a leading catalyst for the broad market.
Yesterday’s post looked at the propensity of the market to chop vs. trend. The basic conclusion was that the market has shown less tendency to trend and more tendency to chop over the last few years. This was especially evident when looking at a long-term chart.
Today I’m going to show the tendencies on an intraday basis. The break down will be similar to yesterday. The first chart shows results of buying any up bar and then exiting on any down bar. 15 minute bars are used for all the tests below. 
A rising graph would show a propensity to follow through and a falling graph would show a propensity to reverse. A flat graph wouldn’t favor either. One note about all these graphs is that the day is closed out flat. There is no overnight holding since we are only measuring intraday tendencies. Throughout most of the 90’s, the easiest way to make money intraday was to find an uptrend an jump on it. As with daily bars, that seems to have changed over the last several years.
What about shorting down bars? How has that worked?
Similar to buying strength, shorting weakness worked well in the 90’s. Over the last few years it has struggled. Even with the difficult stock market performance this year, downmoves have shown a higher propensity to reverse than to continue. The May-June period was a notable exception.
If you trade trend or breakout strategies and have found the market difficult in the last year or so, I’m about to show a prime reason for that diffuclty. Below is a strategy that buys the S&P 500 on the close of any up day. The position is closed at the close of a down day. Essentially your looking to buy strength and sit out weakness.
This strategy worked well in the 90’s, but since the market topped in 2000, it has performed poorly. Since the Spring of 2007 it has really accelerated lower. What this means is that performance following up days has been especially bad. It suggests the environenment has been especially choppy. Whereas strength begat strength in the 90’s, it has led to immediate weakness since.
Now let’s take a look at downside follow through. In this case the strategy is to sell short on any down day and then cover on an up day.
This was a break-even strategy through July of 2002. Since then it has done terrible. Interestingly, it’s done especially bad over the last year plus. Even though the market has fallen precipitously, the manner in which it has occurred has made it difficult to profit if you’re trying to short breakdowns. Down days have been followed by up days.
What about a combination strategy? Buy strength and short weakness in anticipation of further follow through. Below is a chart with the combo strategy:
Again, since March 2007 this strategy would have experienced an incredible freefall.
Perhaps it would be best to take a step back, though. In the next chart I show back to 1960 instead of 1993 in an effort to find other periods where choppiness has been so prevalent.
As you can see, buying after strong days and selling after weak ones worked well for 40 years. In 2000 that changed, and the last year and a half is the worst it has ever been with regards to follow through. This would suggest that strategies that may have worked well for forty years or more could be suffering greatly now. Traders should consider the current choppy market behavior when designing strategies. Buying weakness and selling strength is working better than buying strength and selling weakness. They could also monitor charts like these to see if tendencies begin to revert back to pre-2000. If tendencies do revert, adjustments may be needed.
—————————————————————–
Update: I just saw that Dr. Brett Steenbarger also addressed this topic in his blog today. He looked at it from an international standpoint as well. Click here to see his findings.
———————————————————————
The selloff from Monday followed through on Tuesday. For the second day in a row declining stocks outpaced advancing stocks by nearly 3:1. Also notable is that even with 2 days of strong selling the S&P failed to close at a 10-day low. I ran these observations through the wayback machine tonight and came up with the following results:
Instances are a bit smaller than I typically like but the stats are impressive enough for me to take the study into consideration.
A few additional bits of information about the study:
1) 12 of 14 instances (85.7%) closed above the entry day close within the first week.
2) If you look out 7 trading days that improves to 13 of 14 (92.9%).
3) The average maximum gain over the next 12 days is 5.2%. The average maximum drawdown is 2.2%.
4) After 7-12 days the edge dissipates. The setup is only short-term in nature.
Normally light volume on a pullback can be regarded as a good thing. When the pullback is fairly large and the volume is extremely light the expectations can turn bearish. For instance, the S&P dropped about 1.5% today on the lightest volume since July 3rd. I loosened the parameters slightly to get a decent sized sample and ran a study:
Of course volume this week and next week are expected to be a bit lighter due to traders taking vacations. Still, I don’t find the results terribly encouraging for the bullish case over the next two weeks.
The Nasdaq Composite has now closed higher 5 weeks in a row. In July we saw how negative persistence can be difficult to overcome. Now we have an example of positive persistence:
The kind of positive persistence we’re seeing often manages to continue higher over the following weeks and months.
Will strength in technology help to lift the market in the coming weeks, or will the negative influence of recent volume patterns lead to a move down? Should be interesting…
The market bounced on Thursday, but volume was anemic. On the NYSE it was the lowest volume since the July 3rd holiday-shortened session. When the market moves up, you’d normally rather see it occur on strong volume. Exceptionally weak volume, like we saw today, has been followed by some difficult market conditions over the last 10 years. Some statistics below:
One additional thought on volume for the back half of August. Over the next two weeks it is probable that volume will drop off due to traders taking vacations. When monitoring volume, shorter-term comparisons may need to be made to gain value from the information. For example, rather than comparing volume to its 50-day average, you may get a truer indication by comparing it to the last 1-3 days.
The unusual strength in the SOX continues. I wrote about this last week. Tonight another example of how the market has performed when the SOX takes a leadership position. This time the comparison is to the SPX rather than the Nasdaq:
The market made nice gains for the second day in a row today. Below is a recent chart of the S&P 500. Every time there were at least 2 up days in a row a purple dot appears.
In most cases over the last 2+ months 2 up days has quickly led to a selloff. The most the S&P gained on any 3rd day of this period was slightly over 2 points. A strong move higher could be a sign of a change in character for the market. Tuesday’s action may be worth noting. It could answer the question as to whether rallies will continue to be sold or whether they can start to accelerate their gains.
Time for a shameless plug.
The Short-term outlook from the August 3rd Weekly Research Letter concluded:
“This would imply that weakness early in the week could be buyable for a swing trade.”
That weakness came on Monday the 4th and those who bought into it should have done quite well over the next few days.
In last night’s Weekly Research Letter I showed a study that triggered on Friday the 8th at the close. When this setup occurred the S&P 500 had posted gains the following day 14 times in a row. Today made it #15.
Give yourself a Quantifiable Edge. For less than the cost of 1 buy and 1 sell order at most brokers you’ll get my most in depth research every weekend.
If you haven’t yet received a sample Research Letter, just send an email with your name and email address to weekly@quantifiableedges.com and I’ll send you the August 3rd Letter which includes the above mentioned system.
For more information or to subscribe, click here.
Even with the strong move to a new 30-day high on Friday new highs were unimpressive. In fact they were beneath new lows. I decided to examine the possible significance of this:
Results appear choppy and under perform a random sampling. The instances are quite small, and before jumping to conclusions it’s important to isolate the affect of the indicator. So below is the same test when news highs exceeded new lows:
These results are worse than the 1st case where lows exceeded highs. So while the market may pull back (which it frequently does after making 30-day highs), the blame shouldn’t be laid on the lagging number of new highs.
Today the semiconductor index (SOX) bucked the trend and rose over 1.5%. When semiconductors take on a leadership role, it’s normally a good thing for the market. I decided to look at times following other days where the SOX managed to rise over 1.5% while the Nasdaq dropped .75%:
I’ve showed before how light volume in a stongly rising market will often lead to a pullback. Wednesday’s market action set it up in a slightly different way. It does provide another nice example though.
Three bearish short-term studies all looking at different indicators the last two nights. Upside may be a bit limited over the next week or so.
