Monday was the 7th day in a row that the DJI closed higher. This triggered a study from the Quantifinder that looked at performance after 7-day win streaks in the Dow Industrials since 1980. I’ve updated the stats table below.
There is not much of an edge over the 1st few days. But once you get out a little further, the stats appear solidly bullish. The 16-19 day returns show a very high win %. Momentum tends to carry, and this study is just a simple example of that concept. Traders may want to keep it in mind as one factor when determining their intermediate-term market bias.
There is a seasonal influence that could have a bullish impact on the market this week. Op-ex week in general is pretty bullish. March, April, October, and December it has been especially so – at least until last March. S&P 500 options began trading in mid-1983. The table below is one I have shown and updated basically every year since 2008. It goes back to 1984 and shows op-ex week performance broken down by month.
March has a strong win rate, but the gains are not as high as the other highlighted months. Some people may recall that March used to have the largest average trade of any month. The table below is the one from last year.
So how did the average trade change so much? March 2020.
Last year really changed the look of the curve in a big way. A 1-week decline of 15% will do that. It is most likely just an outlier, and March opex still seems to offer a potentially bullish seasonal opportunity. But the historical stats sure weakened a substantial amount.
A couple of weeks ago I shared the March Seasonality Calendar for Long-Term Treasuries. I have copied it again below. I wondered whether the negative seasonality early in the month would even matter, since TLT appeared so strongly oversold at that point and perhaps due for some kind of bounce.
As of Friday morning, TLT is down nearly 5% on the month. (It is also down over 20% from its August closing highs – even including dividends.) So the oversold bounce did not materialize. But seasonality is now starting to turn positive. Numbers still appear somewhat muted for the next week, but starting on the 19th through the end of the month TLT should have a seasonal wind at its back to help with any rebound. Let’s see if it can mount one.
My friend and colleague Tom McClellan writes the excellent McClellan Market Report. In his report last night he noted that big spikes in QQQ volume are often a bottoming sign for QQQ. He shared a chart showing that the 10ma of QQQ volume was now at an extreme level. I decided to study this idea a bit more. So I took the 10ma of QQQ volume and compared it to the 100ma of QQQ volume, in order to normalize it for historical comparisons. On Monday the QQQ 10/100 Volume Ratio rose from 2.23 to 2.33. That is a very high level – more than twice “normal”. I looked back at all other times the ratio crossed above 2.25. This is only the 6th such instance since QQQ inception in 2000. Below is the list of others along with their 1, 2, 3, 6, and 12-month returns.
A few things to note here:
All 5 previous volume spikes occurred with a selloff over the last 10 days. Fear related to the selloff is typically what will cause such spikes. You don’t see big spikes in volume due to rallies. The current instance is no different. QQQ has declined 7% over the last 10 days.
Intermediate-term returns are bullish across the board. All 5 instances were higher 2, 3, 6, and 12 months later – and by substantial amounts.
Now this study is not to be taken as any kind of guarantee. The number of instances to examine is very low. Additionally, they have all occurred over the last 10 years, which is a period of time that the QQQ has rallied strongly. So multi-month gains would generally be expected. But even with flaws in my study, the concept appears to have merit. Similar spikes in QQQ volume have always occurred during selloffs, and have always been followed by intermediate-term gains. So Tom’s observation certainly seems worth keeping in mind. (Not a surprise – he has lots of interesting and worthwhile observations.)
Fridays are interesting in that they are the least likely day of the week for a selloff to end or a rally to begin. But when rallies do start on a Friday, they have shown the best odds of success of any day of the week. I’ve seen this a number of ways over the years. The study below describes the current market setup. It looks at times the market closed up the day after closing at a 21-day low. Results are broken down by day of the week, and also by holding period.
Looking out 10,15,20, and 25 days, Friday has the best stats of any day. And in some cases, like 15 and 20 days out, none of the other days are even close. So if you are looking for an encouraging intermediate-term sign based on Friday’s action, this appears to be one.
For the next several months, I have decided I will share one of the Quantifiable Edges Seasonality Calendars for the upcoming month. This month the calendar that caught my eye was the US Treasury 20+ Yr Total Return Index. This index is very similar to TLT. It is one of 10 seasonality calendars that we publish at Quantifiable Edges.
Treasuries have been in a big swoon over the last month. TLT closed down 5.7% on a total return basis in February, leaving it down 9.2% so far in 2021. That is a rough start to the year for TLT.
The Quantifiable Edges Seasonality Calendar uses multiple systems to measure historical performance on similar days to those on the upcoming calendar. The systems look at filters like time of week, month, year and so forth. Over the long run, staying out of the market on days that do not appear in green, would have been beneficial. To appear in green the date needs to show a historical Win% of 50% or more and a profit factor of 1.0 or more.
So Treasuries look like they will face a seasonal headwind for much of the first two weeks of March. After that, they may have seasonality on their side. Whether seasonality will matter much in a market that is so overdone right now will be interesting to monitor over the next month. But it is an input that has mattered over the long haul, and traders may want to consider it when forming their trading plan.
Perhaps the most astounding chart I saw this week was one that Helene Meisler posted in her Saturday morning chart show on Twitter. It showed the 50-day MA vs the 200-day MA of the $RUT. I have recreated the chart below and included the full history of the Russell 2000 to provide context.
What is happening here is that the 50ma is more than 25% above the 200ma. And this is the 1st time in history that the 50ma has been this extended vs the 200ma. In fact, never before has it even exceeded 20%. So the current rally in $RUT is remarkable.
I also decided to look at a couple of other indices to see how stretched they were and what their historical charts looked like. First, let’s look at the $NDX.
The $NDX has had some extreme historical readings as well. In 1999 it reached nearly 30, and in 2000, leading right up to the top, it exceeded 35. Then it went to hit its lowest extreme a little over a year later. But the current reading…not all that extreme. And in fact it topped out in September and has drifted lower since.
Lastly, let’s look at the $SPX. For this index I was able to go all the way back to 1929.
The current reading here is about 11. And as you can see, that is not terribly uncommon. There are a lot of peaks on the chart that exceeded 10. But it has not exceeded 20 since the 1930s. So neither $SPX nor $NDX appear to be all that stretched based on this metric.
So what does this all mean? Well, it appears this is some froth in small caps. This does not suggest that the momentum cannot continue or that $RUT or the rest of the market has to decline right away. But I would expect to see a reversion here at some point. There has been a lot of money pumped into the economy, and over the last few months, a big chunk of that has gone into smaller and less established companies.
In fact, it is not only less established companies, it is even shell companies that are seeing massive inflows. Take a look at this Tweet from Charlie Bilello that shows SPAC investment $$ growth:
There is a lot of investment capital out there, and an oversized chunk of it seems to be flowing into the most speculative assets. This is not a trading signal, but it is something that seems worth noting. The music is still playing, but when it stops it may be difficult finding a chair.
Quantifiable Edges has added GLD to our Seasonality Calendars. It is the 10th security (or index) the that we track in the new Seasonality section of the site. We also wrote up a short research paper for GLD. Using spot gold prices we were able to generate historical seasonality data back to 1981. We also showed a comparison of spot gold vs GLD since 2015. The research paper can be downloaded directly from the Seasonality page with a Gold, Silver, Seasonality, or Trial subscription. Additionally, subscribers may download the historical data to do their own research, with the aid of Quantifiable Edges tools that were built in Excel, Tradestation, and Amibroker.
On Saturday at noon EST I will be giving a webinar on “Quantifying Seasonality”. It will cover the following topics:
Difficulties in utilizing traditional approaches to seasonality
Does playing the good and bad days produce an edge?
How to build a seasonal approach that provides an edge
Overview of Quantifiable Edges Seasonality Calendar functionality
In the last few days, Quantifiable Edges released its Seasonality Calendar Backtesting Tool for Amibroker users. We now have tools for traders using Excel, Tradestation, or Amibroker.
The tools allow Quantifiable Edges subscribers to evaluate the Seasonality Calendars and explore ideas on how to utilize them for their own trading. Quantifiable Edges provides subscribers historical calendar data, dating back as for as 1940 for the DJI and SPX. With the tools, subscribers can produce studies like the report I created a few weeks back that looked at only holding indices when the calendars would have shown positive odds based on Win % and Profit Factor readings.
I have not posted many price-action studies to the blog lately, so I thought I would share this one from last night’s subscriber letter. A theme I have seen many times over the years is that persistent uptrends don’t often end abruptly. The study below is an example of this. It considers what happens after the market moves up at least 5 days in a row to a 50-day high, and then pulls back.
We see here a decent edge that becomes stronger and more consistent as you look out over the next several days. Short-term overbought in a downtrend can be a warning sign, but short-term overbought in a strong uptrend often suggests further strength.
Quantifiable Edges recently released its Seasonality Calendars. The Seasonality Calendars utilize multiple inputs that I have found to be helpful over the years. The statistics they produce are based on rolling lookbacks of similar seasonal setups. In a special report I updated over the weekend, I showed how some simple filtering based on the calendars would have worked out over the long-term. The report is free to download for anyone with a Quantifiable Edges login (which you can get with a free trial or a free downloads subscription). In the report, I show results for all 9 of the indices we track with the Seasonality Calendars. Here on the blog I will just show the results for SPX.
The filters used look at the upcoming calendar day and go long or flat based on the stats shown for that day. The approach simply looks to see if there is a better than breakeven chance of seeing gains tomorrow based on 1) Win %, and 2) Profit Factor. So a Win % of at least 50% and a Profit Factor of at least 1 for tomorrow would mean we will be long at the close today. If those requirements are not met, we will exit at the close today. So if today is 12/7/2020, the system will look at the calendar for 12/8/2020 and see that the Win % is above 50 and the Profit Factor is above 1, and it will enter (or continue to hold) a long position. If the calendar on 12/9/2020 shows a sub-par profit factor, then we will exit the long position at the close on 12/8/2020.
When in “cash”, the model will earn interest at a rate equal to the overnight Fed Funds rate. I use this rate because 1) it is often slightly lower than the 3-month T-bill, which many people use, and 2) it has the longest history I could find for a short-term rate (1957). So it is conservative, and has a long history. Below are results for $SPX.
$SPX (along with $DJI) has the longest available history for our analysis.
Only being in the market 62% of the time, it was still able to outperform “buy and hold” by a whopping 4.36% per year.
The strong outperformance was achieved even though there was no interest on cash in the 1st 17 years of the test.
Next, let’s look at a detailed historical drawdown chart.
Drawdown was reduced over “buy and hold”, but was still steep in the 2000-2002 bear market. Part of the reason for this is that the 90s had such a strong bull market that “Baseline” averages were very strong coming into 2000. In other words, reaching 50/1 was easy at the time, so there weren’t many days that got sidestepped early in that particular bear market.
Most other drawdowns saw significant reductions with the seasonality filter vs “Buy & Hold”.
This past week I released some tools for subscribers that allow them to further test the seasonality calendars on their own. This can be done with either Excel or Tradestation, and I also expect to have an Amibroker version available soon. With these tools, subscribers can conduct their own backtests, change parameters, and begin exploring some of their own ideas as well.
The seasonality calendars and associated tools and functionality are included with any Gold, Silver, or Seasonality subscription.
Back in the 1/20/2009 blog I looked at inauguration day returns. I wondered at the time whether a new president brought about new hope and optimism for the market. I have decided to update that study today.
I limited the instances to only those inaugurations where a new president was entering office. I don’t think re-elections carry a sense of “new hope” the way a new president does. I also eliminated inaugurations of Presidents that weren’t elected (Ford in ’74, Johnson in ’63, Truman in ’45, and Coolidge in ‘23). I just don’t believe the same sense of excitement is generated by a replacement as by a newly elected president. The remaining presidents and their inaugurations can be found in the table below.
First, I found it interesting that the wonderful speeches and overall positive vibes surrounding a new president did NOT translate to a strong Inauguration Day performance. (You could throw out Roosevelt and G.W. Bush here, since the market was not open on the days they were inaugurated.) I’ll also note that Donald Trump could (and certainly would) claim he gave the greatest Inauguration Day speech of the last 100 years, since the 0.48% rise on that day was the best of any president on the actual Inauguration Day. It could also be claimed that Obama gave the worst speech in 100 years, since the Dow tumbled 4% on the day of his inauguration.
I’ll also note that looking out over the next 10 and 75 days the market did often seem to embrace the new hope that comes along with a new administration. Harding in 1921 is the only one on the list that saw the market more than 3.5% lower 75 days out. Meanwhile, there were 6 instances where the market was more than 3.5% higher.
As of this writing the Quantifiable Edges blog has published over 185 posts related to market seasonality. Seasonality can be looked at many different ways. You can factor in things like the day of the week/month/quarter/year. Or the week of the month/quarter/year. Or special days, like market holidays or Fed Days. Seasonality can take many forms, and since 2008 I have published a lot of edges related to seasonality.
In the Quantifiable Edges Market Timing Course I looked at longer-term seasonal influences. I featured the Best/Worst 6 Months of the year, and the Presidential Cycle as longer-term seasonal factors.
Over the years I have also privately tested a number of systems that utilized seasonality. And now I have decided to share the Quantifiable Edges Seasonality Calendars. The seasonality calendars utilize multiple inputs that I have found to be helpful over the years. The statistics they produce are based on rolling lookbacks of similar seasonal setups.
Seasonality Calendars are now available as separate subscription at Quantifiable Edges, but we have also included them with all Gold and Silver subscriptions, at no additional cost. (We may increase the price on Gold and Silver subscriptions in the coming months, but all subscribers will be locked in at their original pricing at that time.)
The current list of indices with available Seasonality Calendars can be found below:
The video below provides a short intro to the Calendars, along with the research and resources that come with them:
I’ve posted and updated the “Twas 3 Nights Before Christmas” study on the blog here several times since 2008. The study will kick in at the close today (12/22). This year I will again show the Nasdaq version of the study. While all the major indices have performed well during this period, the Nasdaq Composite has some of the best stats.
The stats in this table are strong across the board. An average year posted a gain of about 2.5% over the next 8 days. The note at the bottom shows the reliability of some kind of bounce has been nothing short of incredible. Traders may want to keep this study in mind over the next several days.
