Friday, January 17, 2020

Is A Very Strong Market "Due For A Correction"?

I've been hearing a lot from people who loudly assert that a strong stock market such as we've recently witnessed is "due for a correction".  Those making the assertion have several things in common:  1) a strong conviction that a bearish move is on the horizon; 2) a total absence of any statistical evidence supporting their view; and 3) dramatic underperformance during the recent market period.  

So let's look at a little bit of evidence.  (Eye-opening evidence, by the way, was published overnight by Market Tells, SentimenTrader, and Quantifiable Edges).

We'll go back to 2005 and identify occasions in which more than 80% of stocks in the SPX have been trading above their 3, 5, and 10-day moving averages and in which more than 80% have been trading above their 100-day moving averages.  (Data from the excellent Index Indicators site).  So we're looking at short-term strength in a longer-term strong market.

There have been only 39 such daily occasions out of over 3300 trading days.  That alone tells you that such broad strength is rare, even in a market that has risen over the lookback period.  Out of the 39 occasions, 22 occurred in 2009 and early 2010.  Note that this was a new bull market period following an important bear market.  During such periods, as I noted back in March, we tend to see momentum conditions.  It's easier to see that we have emerged from a bear market when we look at weekly price charts for small caps and overseas stocks, both of which declined from early 2018 through well into 2019.  

After the strength noted above, in 33 of the 39 occasions we posted a lower daily close within the next three trading days.  So, yes, a pause in the rise after unusual strength has been normal.  However, if we look 20 days out, the average market gain has been +1.43% versus an average gain of +.61% for the rest of the sample.  Over that next 20-day period, 27 occasions were up and 12 down.

History is not guaranteed to repeat itself, but formulating strong views in the absence of any knowledge of history is not trading: It is malpractice.  History provides a rich source of hypotheses and an understanding of market flows can tell us if history is, indeed, playing out.  Markets dance to the rhythms of momentum and value; edges occur because so many participants can't hear the music.

Further Reading:


Wednesday, January 15, 2020

A Fourth Dimension of Trading

The recent post tackled the topic of trading based on understanding what markets are doing, rather than trading based upon isolated patterns of variables.  For me, such understanding has always been a function of three variables:  who is in the market, what they are doing, and where they are doing it.  (See the "How to Trade" posts from October to see how I track these variables).  Where solid quantitative studies can be useful is in identifying occasions in history when these variables have behaved similarly to the present and seeing whether there has been any significant directional edge going forward.  It is by reaching a point of understanding and determining if there is an objective edge associated with current market conditions that we can generate genuine confidence in trading.  Conviction without understanding is mere dogma.  One thing you can count on in markets is that your dogma is always likely to get run over by your karma.

A fourth dimension that I have found valuable is how broadly a market behavior is occurring.  In other words, do we see the same behavior across different segments of the market (small caps, large caps); different market indexes (Asian, European, U.S., etc.); and different sectors within the market (technology, energy, consumer, etc.)?  The breadth of market behavior is key to identifying trending markets (ones in which the majority of indexes are behaving similarly) and rotational markets (ones which are dominated by sector and market reallocation).  For instance, suppose we see that the SPX is moving to a new high for the day.  Whether that move is likely to continue depends not only on who is in the market (volume) and where volume is transacting, but also on whether the movement is narrowly or broadly based.  It helps to think of a trend, not only as a directional move in one index, but as a directional move shared by the majority of market components.  

I have found the various TICK indexes (NYSE TICK, TICK measures specific to the SPX stocks, TICK for the Russell stocks, etc.) to be useful in identifying the breadth of market buying and selling.  To go back to the previous example of a market trading within a range and moving to a new high, the uptick/downtick (TICK) readings on that move will be an important measure of the sustainability of that breakout.  It is when we break out to a new distribution of the upticks/downticks that we can more confidently count on a new distribution of forward prices.  An index may move to a new high in a rotational environment because of the impact of one or two sectors.  That is very different from a scenario in which broad-based buying lifts all sectors.

An important takeaway is that the right trading psychology comes from looking at the right market information, assembling it the right way, and generating the right understandings.  As Mike Bellafiore emphasizes, we produce one great trade when we begin with a valid thesis, see that thesis set up in a way that we have studied (play booked) in the past, and then fight for the best price to execute the trade based on our idea.  We don't trade well because we have a good psychology; we cultivate a confident psychology when we learn to trade well.

Further Reading:


Monday, January 13, 2020

Knowledge and Understanding in Trading

I've been teaching myself new approaches to trading and have learned many valuable lessons in the process.

One of the most significant changes I've made is trading from a place of understanding rather than a place of knowledge.  

As the quote suggests, knowing something and understanding it are quite different things.  I know a number of people, but I would not be so presumptuous as to pretend that I understand all of them.  Similarly, I might know a Bible passage or a poem, but that doesn't necessarily mean I understand them.

Much of what is taught in trader education is knowledge.  It might be knowledge about fundamental factors that influence the stock market, such as interest rates.  It might be knowledge about chart patterns, trends, and indicators.  It might be knowledge about potential catalyst events, such as shifts in monetary or fiscal policies.  From their knowledge, traders typically attempt to make predictions, such as whether the market will go up or down.  Sometimes the predictions made from the pieces of knowledge are quantified through backtests.  This is common among many of the services that I recently highlighted.

This knowledge-prediction paradigm of trading is what I have found to be limited.  "X is occurring; therefore the market should do Y" does not necessarily reflect any understanding of why that relationship might hold.  When we look for X-Y patterns in markets, it becomes easy to reach for so many patterns that the relationships we trade are spurious and not meaningful.  That is how overfitting occurs in backtests, for example.  We test so many combinations of variables that eventually we find the 1 in 20 that is significant at the p<.05 level!

In science, we first observe nature and develop theories about what is occurring and why.  Theory building is the hallmark of understanding:  a theory represents causal thinking, not just correlational thought.  "The market is going higher because we've formed a certain candlestick pattern on a chart" does not capture anything of a causal nature.  Conversely, if we look at the expansion of the Federal Reserve's balance sheet and their stance on rates and hypothesize that excess funds in a low rate environment will spur speculative activity, that could represent part of  understanding of why we're in a bull market.  Or if I break down volume that is transacted at market bid and offer prices and notice that institutions are predominantly lifting offers across different time frames, this could represent an understanding of market participant behavior and a theory of why we're seeing a market trend.

To be sure, once the scientist has a promising theory, it's important to put the ideas to the test--and that is where prediction comes in.  Ideally, a trade is a test of a market hypothesis derived from a trader's understanding.  When well-constructed trades are working out, they add confidence to our theory.  When they don't work out, they may lead us to revise our theory.  Ideally, our trading is our way of testing our understanding of the market.

Too often, however, traders assemble knowledge and immediately want to create trades out of what they have learned.  Bypassing the process of understanding leads to a shallow perspective on market behavior--one that does not merit true conviction.  Genuine conviction comes from deep understanding, not simple correlations and patterns.  "There is nothing so practical as a good theory," psychologist Kurt Lewin observed.  What traders need are frameworks for understanding how markets behave and why, not mere "setups" for the next trade.

Further Reading:


Friday, January 10, 2020

Integrating Discretionary Trading Skills With Quantitative Strategies

There is a number of very good research services and platforms that allow traders to identify occasions when there is a directional trading edge in markets.  The common thread among these is that they identify a set of conditions that are present in today's market that are distinctive and meaningful.  They then examine past occasions when these conditions have occurred and determine whether the market has moved a particular way in the next time period with statistically significant odds.  This is known broadly as event research.

For example, if the SPX makes a new 12 month high for the first time in two years, we could explore whether, in the past, this has reliably led to future gains.  Or, if we have a Fed meeting on Day 1 and close that day weak, we can identify the odds of continued weakness over subsequent days.  Or, if the NYSE TICK hits a very strong level of +1000 during the opening 15 minutes of the trading, we can assess the odds of this being a trend day to the upside.

In all these cases, we're using historical research to see if there is a directional edge during the upcoming period in the market.

Seven providers of such research that I have found to be reliable and useful are (in alphabetical order):

We can also conduct our own event research, as my recent post illustrates.  Such studies do not require advanced mathematical or programming methods as one might need for a fully developed trading system.  The goal here is not to become a systematic trader, but rather to identify promising hypotheses for the coming trading period.  When experienced discretionary traders possess one or more valuable historical hypotheses for the coming day or week, they can then track news flow, price action, and overall market behavior to assess whether the hypothesized move projected from market history is actually playing out right now.  In other words, you look in real time to see when there is a reliable "setup" or catalyst that allows you to trade a historical edge with well-structured risk/reward.

History doesn't always repeat itself, but knowledge of history generally beats ignorance.  Understanding how markets have moved in the past under the conditions we see at present can keep us out of bad trades and help us focus on promising ones.  It's an important way that discretionary traders can make use of quantitative strategies without having to do the data collection and coding from scratch.  And it can become an important part of our market preparation.

Further Reading:


Sunday, January 05, 2020

The Fatal Mistake Traders Make

The fatal mistake traders make is that they define themselves narrowly, and this artificially constrains their opportunity set.

For instance, a trader might define himself or herself as a "breakout trader", a "trend trader", or a "bear market trader".  All ensure that the trader will underperform when markets are not breaking out, trending, or moving lower.  The frequent justification for such limitation is that the trader is adapting trading to his or her personality.  But would that work in other performance fields?  Would a quarterback in football last long if he defined himself only as a running QB?  Would a baseball pitcher succeed if he declared himself to be a fastball pitcher?  How about an actress who only played one kind of character?  In every performance field, ongoing and elite levels of success require the ability to adapt to the opportunity set, not expect the game to adapt to the performer.

The trader who does one thing consistently in all situations is not a disciplined trader.  He is a one-trick pony.

Let's  take a practical example:

During 2019, the SPY ETF moved a total of nearly 73 points.  During the NYSE day session, total movement up and down was about 303 points.  Total movement overnight was about 272 points.  During all the day sessions, SPY gained about 39 points.  During all overnight sessions, SPY gained about 34 points.  In short, declaring oneself to be a daytrader effectively cut the opportunity set in half.  

Many daytraders work hard at improving their trading.  Less often do they work at broadening their trading.  As I mentioned in a recent post, there are many markets in which a demonstrated edge *is* present, but not on the day time frame.  Similarly, there are markets in which a demonstrated edge is present, but not directionally.  (For example, volatility may be in a very tradeable declining trend, but the market may not move a lot directionally during that period.  One part of the market may be moving higher, such as large caps, while another is moving lower, such as small caps.  A good long/short trade is present, but perhaps not an overall market trade.)

I consistently find that traders who develop multiple ways to win show the greatest career longevity.  As in the business world, long term success requires flexibility and innovation, not a stubborn insistence that the market adapt to our preferred edge.  In a coming post, I will share how I have been expanding my trading through the integration of quantified models and discretionary pattern recognition.

Further Reading: