Saturday, August 29, 2015

Live Your Life the Way You Want to Trade

A major problem I encounter among professionals is that they pursue business through busy-ness.  They equate doing more and getting items off their to-do lists with being productive.  In the recent Forbes article, I provide an alternate perspective:  one in which unstructured time can be the most productive time of day.

Here is an important principle of trading psychology:  You can only manage your trading capital--and your trading--as effectively as you manage the rest of your life.

Take that one to the bank.

Will we truly trade in a rule-governed, responsible way if we live our lives unconstrained by organization and discipline?

Will we truly develop new and better ideas if we're not regularly feeding our brain with new and better information and experiences?

Will we steadily grow and develop as traders if we're not growing and developing in the rest of our lives?

How we live is how we'll trade: they are both expressions of what we ultimately prioritize.

I can offer no better advice than this:  Live your life the way you want to trade.  Be the person outside of trading hours you want to be when markets are open.

Further Reading:  Managing Your Money and Your Life
.

Monday, August 24, 2015

Trading Notes: Week of 8/24/15

Friday, August 28th

*  Buying weakness yesterday ended up being a useful strategy for much of the day, but a strong selloff and then buying reversal in the last two hours of the trading session made for volatile trading.  I continue to expect these volatility aftershocks to be with us, as noted in Wednesday's post.  That being said, the pure volatility measure has been generally trending down over the past several sessions, though it's still at historically high levels.  

*  The pattern of improving breadth has continued.  Interestingly, we had 131 stocks across all exchanges register fresh new monthly highs and 136 make new lows.  That new low figure compares with 726 and 840 during the prior two sessions and 3553 at what appears to be a momentum low.  As long as we see higher prices, strong NYSE TICK readings, and improving breadth, my game plan is to buy weakness, particularly if it can hold above the overnight lows.

*    Pertinent to the issue of declining, but still high volatility, is the decline in volume in SPY over the past several sessions.  We peaked at 507 million shares traded on Monday, then roughly 369, 339, and 275 million the past three days.  One of my quant routines estimates the likely day's range (very helpful for establishing price targets on trades), with volume as a major input.  By seeing how today's volume unfolds, we can estimate the likely day's volume and make informed estimates as to likely price range for SPY.

*  A measure of the breadth of market strength that I track daily is the number of NYSE shares that give buy vs. sell signals for various technical measures.  Yesterday we had 520 buy signals vs. 20 sells for the Parabolic/SAR measure.  We've had 11 occasions where we've had over 400 buy signals on that measure since June, 2014, when I began collecting the data.  Three days later, SPY has been up 8 times, down 3 for an average gain of +.49%.  It's too small a sample to hang our hats on, but does serve as a nice reminder that strong upside thrust leading to a short-term overbought situation (over 90% of SPX stocks are above their 3- and 5-day moving averages) does not necessarily lead to "mean reversion".

Thursday, August 27th

*  Per yesterday's post, we did indeed see a test of lows in Wednesday's session, followed by significant buying.  The NYSE TICK has provided very good tells for the recent market action, as broad downticking was followed by broad upticking very early in yesterday's rally.  That strength has continued into the overnight, all of which is consistent with the idea of having put in a momentum low per Tuesday's comment.

*  Another good tell in yesterday's trade was the continued drying up of the number of stocks making fresh new lows on weakness.  We had 726 monthly lows across all exchanges yesterday versus 840 on Tuesday and 3553 (!) on Monday.

*  The pure volatility measure continues at high levels, suggesting that volume and volatility might be with us for a while.  We remain above 30 in VIX, off the highs but significantly elevated relative to most of 2015.

*  I find that the psychological issues faced by traders in this recent market are related to difficulty adapting to the recent changes of trend, correlation, and volatility.  Per the recent Forbes article, a big part of emotional upheaval in trading comes from the mismatch between the patterns we've become accustomed to seeing in markets and the new patterns that are playing themselves out in current market action.

*  I note that 89% of SPX stocks are now trading above their three-day moving averages, but only a little over 2% are trading above their 10-day averages.  I expect that latter number to increase significantly, as our correction from the recent weakness plays itself out in time as well as price.  My game plan is to be open to buying weakness as long as we stay above the overnight lows.

Wednesday, August 26th

*  The break of the 1920 level in ES referenced yesterday showed us that the China cut was not a game changer, and we continued the volatile downward market.  Interestingly, fewer stocks made fresh new lows yesterday relative to the day previous.  Specifically, we had 3553 monthly lows across all exchanges two days ago and only 840 yesterday.  This would be consistent with having put in a momentum low, which on major declines can precede ultimate price lows by weeks or even months.  

*  Yesterday we had fewer than 5% of all SPX stocks trading above their 3, 5, 10, 20, and 50-day moving averages.  Going back to 2006, there have only been four occasions in which we've seen similar broad weakness:  10/7 and 10/9 of 2008; 11/20 of 2008; and 8/8 of 2011.  Over the next three trading days, SPY moved -9.74%; 9.68%; 13.95%; and 4.75%, respectively.  Note the very high volatility going forward.

*  Note also that the dates above did not make ultimate price lows for a number of months, but ultimately led to major bull market moves.

*  I am alert for the possibility of our putting in lows near term and am watching that 1830-1850 level in ES closely.  I would expect the combination of short covering and value buying in the high volatility environment to create a major snapback rally from these very oversold conditions.

Tuesday, August 25th

*  The volatility alert proved useful for Monday's trade, as we blew out to a VIX of 40 by the close and traded with a daily true range in excess of 8%.  We traded well off the day's lows, inviting the hypothesis that we've seen a momentum low for this down move.

*  Contributing to the idea of a momentum low is the unusually broad, weak breadth.  Over 400 SPX stocks made fresh 20 day lows yesterday.  We haven't seen that kind of broad weakness since early August, 2011.  That indeed was a momentum low, but note that stocks drifted lower, albeit with fewer stocks registering new lows, for a couple of months after that.

*  I will be watching commodities closely, as these have been a useful alert for EM equity weakness and weakness in U.S. stocks.

*  I will also be watching the ES 1920 area, as that was the most recent low prior to the China announcement of a reserve ratio cut.  If that move by China was a game-changer, that level in ES should hold.  Game plan is to buy weakness that holds above that 1920 level.

Monday, August 24th

*  There's a very key point toward the end of the recent Forbes article:  many times what we interpret as a trading problem actually represents *information*.  For example, the trader who has difficulty pulling the trigger may very well intuit that the market environment has changed and that what had been good risk/reward may not be so at the moment.  Similarly, frustration could be a very good early, emotional signal that regimes have changed.  We spend a lot of time trying to prevent and combat emotions in trading when perhaps we should be focused on learning from them.

*  With respect to the recent article, also note the five things I look for to identify changing regimes.  Very helpful in adapting to market conditions.  These will be relevant once we're ready for the bear to turn.

*  We continue to see price behavior that is wildly different from what we seen through most of 2015.  For example, from midnight to 8 AM EST so far today, we've printed 150 bars based upon 500 ticks of price movement.  By contrast, on the Monday two weeks ago, from midnight to 8 AM we printed 6 such bars.  My pure volatility measure continues to make new highs, meaning that we're getting more movement from more volume, but also more movement per unit of volume.  This makes risk management tricky, as we tend to rely on the same time frames for holding periods without realizing that holding a trade for a hour now is like holding for a day earlier in the year.

*  We closed Friday with fewer than 5% of SPX stocks trading above their 3, 5, and 10-day moving averages.  That kind of broad short-term weakness has been more typical of 2010 and 2011 markets than recent ones.  Very often, when we've had such broad weakness, the first thrust down has been followed by further weakness, which is what we've seen overnight.  I continue to favor selling bounces until there's some evidence that selling pressure cannot continue to yield weakness across the vast majority of sectors.
.

Sunday, August 23, 2015

Adapting to the Market's True Clock

The previous post illustrated why trading is challenging, as markets change their volume, volatility, and trendiness--sometimes quite radically.  On Friday, for example, SPY traded with a true range of over 3%, three times the average daily true range of the previous 20 trading sessions.  Volume expanded to over 345 million shares, also three times the average volume of the prior 20 sessions.  Friday's pure volatility (volatility per unit of volume) rose to twice the level we had averaged since April--meaning that, not only were we getting more volume coming into the market, but the volume was creating twice the previous movement.

Above we see a chart of ES price movement for 8/20/15 and 8/21/15, where each data point is a closing price after 500 price changes (ticks).  We had 139 price change bars for the day Friday.  By contrast, there were 39 bars for the prior Friday.  If we look at time as the market's clock, then every day is like every other one.  Once we view movement as the market's clock (see this excellent paper), then every day can be quite different and offer very different opportunity sets. 

The oscillator (red line) is a five-period rate of change measure that acts as a short-term overbought/oversold measure.  Because we print more bars when the market's clock speeds up, we see more overbought/oversold oscillations.  That could create more opportunity; also more frustration and confusion.  The challenge is that, using time as their gauge, most traders fail to match the market's clock: they are like dancers on a dance floor who keep slow dancing once the tune turns to dubstep.  Conversely, when traders are calibrated faster than the market's pace, they overtrade, expecting more up and down movement than the market affords.

Imagine playing basketball when the shot clock randomly changes from 24 seconds to 48 seconds, then to 12 seconds, and then back to 24.  A team that had one way of running plays--and one time calibration--would inevitably play suboptimally.The market's shot clock changes over time, but so often we don't.

It's just another example of the simple-mindedness behind the advice that traders should "follow their process."  If you have *a* process, you'll get run over when markets change.  It's meta-processes--sets of rules for engaging in different processes--that allow one to adapt to ever-changing markets.  From the perspective of meta-process, creativity--not slavish adherence to one way of doing things--is an essential component of effective discipline.

Further Reading:  Trading in Event Time
.

Saturday, August 22, 2015

Why Trading Financial Markets Is So Difficult

Yesterday was one of my better days of the year trading ES, and a large part of the success consisted of *not* following my process.  Therein lies a tale and why so much common wisdom in trading psychology is horseshit.

First off, there is an underlying message you'll hear from every furu:  trading markets is really easier than you think.  Why is this the meta-message?  Because the furu needs to appeal to the quick money crowd.  There's no sexy message if you tell people that trading is difficult; that even the world class traders have losing periods; that it takes years to master any craft.  So the furu goes after the people who *don't* want to work for a living.  After all, what more alluring message to send than, "it's really easy...all you need to focus on is X."

Well, it's not easy and there's one major reason for why markets are so difficult to trade:  they are always changing.  The market you traded two weeks ago is not the market that traded for the past two days.

Above is a chart where each data point represents 50,000 ES contracts.  It covers the period from 7/1/15 to the present.  Note how the rules that would have made you money very early in the period are different from the ones that succeeded in the middle of the period and extremely different from what has succeeded most recently.

I build quantitative models of the ES market.  The backtests for those short-term models have gone back to 2014, as this has been a relatively stable (stationary) regime (i.e., a stable set of rules has been predictive of forward price movement).  As we went from trading Thursday to Friday, it became clear that the models (which generally put you long in oversold, higher volatility environments and short in overbought, lower volatility periods) were not working.  They were not working because the price action in the market was not typical of the price action from which the models had been derived.  Only the October 2014 period was similar, but even there important differences were present.

(Parenthetically, we closed Friday with fewer than 5% of SPX shares trading above their three- and five-day moving averages and fewer than 10% trading above their 10-day moving average, with VIX having soared to 28.  That price action is much more similar to 2010 and 2011 conditions than 2014-2015 ones).

So I disregarded the model output and instead turned to my research that identifies when we have a likely trend day.  The persistent hitting of bids on expanded volume was indeed present and led to a good afternoon short trade.  If anything, the models would have had me hunting for bottoms.  

I did not follow my process because I have a meta-process for adapting my process.  In other words, I am constantly scanning to see if this time really is different so that I can adapt to changing conditions as quickly as possible.  I don't want to be trading mean-reversion (which worked well through much of the period of the chart above) when we're now in a trend/momentum environment.

This is why trading is so difficult.  In most performance fields, we can count on stability:  the pitching mound does not go from 90 feet from the plate to 60 and then 120.  The basketball hoop does not go from 10 feet high to 12 and then 8-1/2.  The rules of chess are constant throughout tournaments; the rules of markets can change over the course of a week.  

Success can come from faithfully following your trading process.  Ongoing success comes from quickly recognizing when that process needs to adapt.

Further Reading:  Trading Coaches as Whores
.

Monday, August 17, 2015

Trading Notes: Week of 8/17/2015

Friday, August 21st

*  It was a second straight day of very weak volume flows; the failure to sustain bounces was evident relatively early in the day and quite telling, as the range market gave way to a trending trade.  Stocks making fresh new lows expanded and almost 600 NYSE issues closed below their lower Bollinger Bands.  Interestingly, when that has happened in the past year, there has tended to be further downside over the next five trading days.  We are very stretched to the downside short-term, with fewer than 10% of SPX stocks trading above their three and five-day moving averages.  With VIX at 19, I expect more volatile moves and the possibility of stiff short-covering moves.

*  My next day and 3-5 day models have turned moderately bullish, reflecting the short-term oversold condition.  Intermediate term measures are not at levels that we've recently seen at intermediate-term lows.  I will be watching the quality of bounces from the oversold level closely; if weak, I would expect further downside.  My game plan is to wait for those bounces before reinstituting shorts.  I'm also open to short-term buying of weakness that fails to make fresh price lows.

*  Pure volatility is quite high, which means that each unit of volume produces significantly more price movement than it did at recent market peaks.  With expanded volatility, moves can extend further than we would expect--very relevant to risk management.

Thursday, August 20th

*  Volume flows were weak through the day, confirming the bearish expectations of the models and the downside stock trade in the wake of commodity/China weakness.  This weakness has carried forward to premarket trade.  Short term indicators are at levels stretched to the downside, but intermediate-term ones are not.  For example, about 17% of SPX shares are trading above their 3-day moving averages.  That creates an environment ripe for a counter-trend bounce.

*  The possibility of counter-trend bounce is also heightened by the fact that one of my 3-5 day models is flashing a bullish signal, the first in quite a few days.  Should we see weakness in today's session not confirmed by volume flows and breadth, I would be willing to participate on the long side for a short-term trade.  

*  Pure volatility measure is elevated, also highlighting possibility of near-term bounce and VIX closed above 15.  Neither are at levels associated with intermediate-term bottoms, but I do expect to see more volatile trade going forward, which has implications for sizing, stops, and targets.

Wednesday, August 19th

*  Whereas Monday's market could not sustain selling pressure--the number of downticks across stocks was modest--Tuesday was the reverse, with limited buying.  The volume flow measure was solidly negative on the day, and we could not take out the overnight highs.  This tape action fit well with the bearish model signals from yesterday, and we're seeing further price weakness in premarket today.  I continue to doubt a sustained upside for stocks as long as we see continued commodity weakness, which speaks to global economic weakness, especially in EM.

*  We can see breadth weakness in the number of stocks persistently trading under their lower Bollinger Bands versus those trading above (see below; raw data via Stock Charts).  Note how the divergence in the cumulative Bollinger balance preceded the drop in October of 2014.  We are currently seeing quite a massive divergence; there are simply more stocks across the NYSE universe trading with significant weakness than with significant strength.

*  Next day and 3-5 day models are neutral, and we continue to work on the bearish swing signal from yesterday.  Game plan is to continue to sell bounces that cannot exceed the overnight highs and monitor volume flows, particularly should we test support in the ES 2070s area.  Downside action without expansion of downside tape action has been a good short term signal for profit taking on the short side. 

Tuesday, August 18th

*  We had an impressive rally once we held Friday's lows in the ES.  My volume flow measure only got as low as -250 early in the morning, compared with values well below -1000 when we've hit truly oversold levels.  That was a sign that, like Friday, we just were not seeing aggressiveness among sellers.  The ensuing rally has turned my next day model very modestly bearish and the 3-5 day models have turned bearish.  We are at levels of pure volatility that have been associated with weak next 3-5 day returns.  My game plan is to short bounces that stay below the recent highs.  On a swing basis, I am alert for indications of more aggressive selling that would take us below the recent support in the 2070s.  That being said, I'm also watching XLE and XLB carefully for any indications we could be bottoming in the commodity-related sectors.  That would offer important support to stocks on any pull back.

*  My next research project will be to track volume flows in the premarket and also at end of day, including the after market.  It is not clear to me how the flows in these time periods might be related to those in the next period, but there are meaningful flows, especially end of day, that could offer clues as to the next day's trade.

Monday, August 17th

This article goes into depth about what I believe to be the greatest performance problem affecting traders. I see a real mismatch between how traders think about opportunity and how they actually manage their positions. It's too glib to simply attribute this to "lack of discipline" or failure to follow plans/process.  Quite literally, traders become caught between two imperatives and the inability to reconcile those results in the quandary where we plan trades but don't trade those plans.

*  We've been seeing a decline in stocks making fresh new highs for the past several months (see chart above), but we've also seen a drying up of new lows during the rangebound trade.  My next day and 3-5 day models are very modestly bearish and several of my measures are moderately overbought and in ranges where we typically see weakness.  My game plan is selling strength that fails to take out overnight highs.

*  Friday's trade saw a distinct drying up of downside volume flows, followed by consistent though moderate buying flows.  Flows exceeding Friday levels on either side would strike me as significant in establishing direction for early trade this week.  When you see concerted hitting of bids or lifting of offers across the stocks trading the highest volume, you know that major market participants are putting capital to work.  That wasn't happening on Friday, as SPY volume was at very low levels.  Need to see evidence of meaningful volume flows before assuming any breakout from the recent pattern of declining new lows and new highs.
 

Sunday, August 16, 2015

Tracking the Footprints of Supply and Demand in the Stock Market

In some ideal, dispassionate world all participants in financial markets would behave like my friend NRK and only buy at bid prices and sell at offers.  It's a great way to do well in auction markets, and there are optimal execution platforms that will accomplish that for large market participants.  The problem with optimal execution is that you often have to wait to get good prices, and price can get away from you while you're waiting.  That can happen if you're looking for a bargain in a hot real estate market or if you try entering on limit orders in a fast market.  

The truth of the matter is that the world is neither ideal nor dispassionate.  People chase price bubbles and puke out of positions when everyone else is doing the same.  Real world traders--and even those execution systems--recognize this and allow for execution with urgency.  When markets are moving with signs of momentum, it's acceptable to lift offers and hit bids to participate in the movement.  Of course, in this less than ideal world, there's also that small inconvenience called cognitive bias.  We all too often extrapolate straight lines and think that every move higher or lower is a budding trend.  That gets us executing with urgency more often than truly objective signs of momentum would dictate.

The long and short of this foray into the vicissitudes of decision making is that even the largest market participants leave footprints as to their intentions.  Yes, they can chop orders up into pieces and execute in baskets and trade on different exchanges.  At the end of the day, however, their transactions occur within a bid/offer matrix, and they are either executing with urgency or they are not:  they are either willing to pay the seller's price or they're willing to let markets sell down to their desired buying level.

When you analyze markets transaction by transaction, you can actually see the flow of trades and volume hitting bids and lifting offers--and you can track the evolving sentiment/urgency of market participants.  That is highly useful information that is obscured by looking at traditional intraday or daily charts.  A daily chart is a telescope, revealing a big picture.  Viewing markets transaction by transaction to detect the flows of trader behavior requires a microscope.  A major error that traders make is attempting to engage in microscopic analyses by changing the magnifying power of their telescopes.

Above is a chart that tracks a measure of "volume flow" that I've spent the better part of this weekend constructing.  It looks at every trade for every stock in the NYSE universe and identifies where the transacted price occurred within the bid/offer matrix existing at that time.  If the volume of that transaction occurred nearer to the bid price, it is considered "selling volume".  If the volume occurred nearer to the offer price, it is considered "buying volume".  The volume flow measure is a running total of buying vs. selling volume.  If you think of it as a volume-weighted NYSE TICK measure or as a cumulative Market Delta for all stocks, you wouldn't be far off the mark.

Much as I love the TICK measure, it is democratic to the point of promiscuity.  All upticks and downticks are equal, whether they stem from the smallest microcaps or the bluest of blue chips.  That is not so helpful when trading a capitalization-weighted index.  You can have relative strength or weakness among small caps as part of a sector rotation that will skew the TICK measure--sometimes considerably.  A volume-weighted TICK measure, on the other hand, naturally weights larger, more actively traded issues more highly, providing a potentially better supply/demand view for the stocks that matter in the index being traded.

Note above how selling flows were quite moderate in early trade and then accelerated even as price was topping out for SPY.  This is potentially useful information for the intraday trader.  

As I work with the new measure, I hope to share further findings and details.  The important takeaway is not about the specific measure--it may or may not add value over time.  Rather, what is important is looking at markets in fresh ways and sustaining a process that periodically renews and extends your perspectives.  Looking at the same charts and indicators as everyone else is a great way of replicating the returns of everyone else.  If you want distinctive returns, you need to view markets through distinctive lenses that reveal fresh opportunities.  It used to be that discipline meant following a particular set of rules or processes.  In crowded, ever-changing markets, however, creativity is the new discipline.

Further Reading:  Why to Keep a Journal
.

Saturday, August 15, 2015

Reading the Psychology of the Marketplace

How is psychology related to trading?

The usual answer is that the psychological state of the trader helps to determine trading performance.  In other words, traders focus on their own psychology to help them make sound decisions.

That, however, is only half the equation.  Reading the psychology of other market participants provides important clues as to market behavior.  This is a topic that I'll be discussing at next month's Traders4ACause conference event in Las Vegas.  (By the way, that will be an excellent conference lineup, including Jack Schwager, Jim Dalton, Mark Minervini, Bao Nguyen (@modern_rock), Peter Brandt, and much more).    

So how do we read the psychology of other market participants?  Here are three tells I look for:

*  Volume - Are more people participating in the market or fewer?  If volume expands significantly, we know that the expansion comes from institutional players capable of moving the market.  They are perceiving opportunity.  Seeing their flows hit the market helps us enter moves early and place the wind at our backs.  Low volume tells us that large participants do not perceive distinctive opportunity, and we're much less likely to see large trending moves during the day.

*  Upticks/Downticks - Measures such as the NYSE TICK and volume transacted at bid vs. offer (as measured by Market Delta) tell us whether participants are aggressive on the upside (lifting offers) or downside (hitting bids).  These real time sentiment measures are also useful in identifying when broad buying or selling interest hits the market and thus make a useful complement to volume analysis.  A cumulative running total of upticks/downticks provides a very helpful intraday trend measure.

*  % of Stocks Trading Above Their VWAP - This is a measure updated every few seconds on the e-Signal platform.  If there is broad buying interest, the number will stay well above 50%; broad selling interest will show up as persistent readings below 50%.  When we're near 50% and oscillating around that level, we're much more likely to have a rangebound market dominated by sector rotation.     

Basically I want to know:  a) who is in the market; b) whether they're dominantly leaning in a particular direction; and c) how participation and sentiment are evolving over the course of the day.  Psychology is important to our performance, but it's difficult to read the psychology of the marketplace if we're too focused on our own state.

Hope to see you in Vegas!

Further Reading:  %VWAP   Relative Volume  NYSE TICK 
.

Tuesday, August 11, 2015

Trading Notes: Week Of 8/10/2015


Friday, August 14th

*  For me, the standout observation for today's trade is that the yuan stabilized at the overnight fix, we got a small pop in stocks, and then there has been no follow through and, as I write, that small pop has been entirely reversed.  Meanwhile oil and copper continue to languish near their recent lows.  The need for devaluation in China is a response to economic weakness and perhaps the best real-time gauge of that weakness is in commodities pricing.  I have trouble seeing any sustained bull move in stocks until the deflationary pressures reflected by commodities weakness abate.


*  My various breadth measures have been peaking and, recently, bounces in stock market breadth have not been sufficient to lift the index to new highs (see chart of Intermediate Strength above).  I'm content, given the dynamics above, to have a small swing position for a move to oversold levels. The models are neutral, and it would not take a huge amount of weakness to turn them bullish, so at this juncture I'm not anticipating a full-on bear market move.  Too, as the above chart shows, we've seen a drying up of the number of stocks making fresh lows.  Should we get near-term price weakness and fewer stocks making new lows, I will use that as a tactical opportunity to take profits.  If the weakness hypothesis is correct, we should not take out the overnight highs in today's trading session.  With models not lined up, I'm happy to have a relatively short leash for position trades.

*  The breadth volatility measure referenced yesterday is at the lowest levels seen since late February/early March.  Such low levels have been associated with weak returns over a next five-day basis.  Pure volatility has also moved below median levels, also associated with weak forward returns over a swing basis. 


Thursday, August 13th

*  Yesterday's note about caution chasing oversold levels when pure volatility is high turned out to be more right than I expected, as we reversed the recent weakness and rallied strongly in SPX.  We dropped all the way below 2050 in the ES futures early in the day, a two-week low.  Interestingly, we only saw 1058 stocks across all exchanges make fresh monthly lows vs. 1078 on 8/6 and 1862 on 7/27.  Note a number of sectors that held up with relative strength, including the yield-sensitive utilities and consumer staples shares.  The commodity-related stocks have also held up.  This told us that the early drop was more about sector rotation than full-on risk-off--a valuable tell.  Market breadth has not been weakening with the recent China related selloff.

*  Pure volatility remains unusually elevated even with yesterday's rebound; when this occurs, there is usually more upside left in the move, as the combination of volatility and strength leads to near-term momentum.

*  The 3-5 day models are mixed:  one is neutral, the other mildly bearish.  The next day model is neutral.  These are the kinds of signals you get in the middle of trading ranges.  I could be persuaded to buy intraday weakness that holds above the overnight lows for a short-term trade higher based on the strong pure volatility, but otherwise don't perceive a strong edge.

*  My breadth volatility measure is hitting multi-week lows.  That's a measure of the volatility of day to day breadth and it's been a helpful measure.  When breadth volatility has been in the lowest half of its distribution since 2014, the next five days in SPY have averaged a loss of -.01%.  When in the highest half of its distribution, the next five days in SPY have averaged a gain of +.34%.  If we were to get other volatility readings dropping, the models would likely turn bearish.


Wednesday, August 12th

*  China devaluation continues as major driver of stocks globally.  While developed market equities have traded in a range over the past several months, emerging market stocks have been in a consistent downtrend (see EEM chart above).  This suggests that a major engine of recent global economic growth is no longer a contributing factor.  It is this weakness and not the stimulus value of the weaker Asian currencies that is driving stock and commodity markets lower and stimulating a flight to the safety of quality yield.

*  The 3-5 day models for SPX are neutral; the next day model ended Tuesday very modestly bullish.  The models cannot factor in idiosyncratic market factors such as the devaluation, so I am not relying upon them for signals at this time.  Sometimes this time really is different.

*  My measure of "pure volatility"--the average price movement per unit of trading volume--has become elevated, which means that we could see outsized moves (including countertrend ones) as volume picks up.  This has important implications for the setting of stops and targets on trades and makes it particularly dangerous to chase overbought or oversold price levels.

*  General game plan is to continue to sell bounces in SPX that terminate at lower highs.  If the devaluation truly is contributing to an ongoing risk-off trade, we should not trade above the levels seen just prior to the most recent yuan fix.  I continue to harbor doubts about any kind of sustained Fed hiking in the face of what is increasingly looking like a currency war of competitive devaluations.  

Tuesday, August 11th

*  China devaluation key piece of overnight news; USD rises vs. Asia; stocks give back a chunk of Monday's gains.  China devaluation affirms government concern over economic weakness.  Difficult to see much in the way of Fed hiking with Asian goods becoming cheaper in U.S. and dampening inflation.  Also difficult to see Fed hiking in any sustained way in the face of what is increasingly looking like a currency war.  All in all, this is consistent with the macro themes recently outlined and should be supportive of U.S. stocks offering yield.  Economic benefits of lower inflation/lower prices are tempered by headwinds from higher USD.

*  My 3-5 day models for SPX turned modestly bearish at end of day on Monday, but reaction to the China news swamps model effects.  Next day model for SPX turned from modestly bullish to neutral.  

*  General game plan is to sell bounces in SPX that fail to take out Monday highs.  Commodity markets have been a good proxy for the Asia weakness theme, and it is difficult to see stocks sustaining a rally if that theme is dominant.

Monday, August 10, 2015

Expanding Your Views of Markets

*  While the overall SPX index has traded within a multi-month range (blue line above), the cumulative number of stocks in the NYSE universe trading above their upper Bollinger Bands vs. below their lower Bands has been steadily declining.  That tells us that we're seeing more significant weakness than strength among individual stocks, no doubt aided by the relative weakness in commodity-related shares.  I am seeing the same underlying weakness in the cumulative NYSE TICK, suggesting that we're seeing more hitting of bids than lifting of offers among the broad range of NYSE shares  (Raw data via Stock Charts).

*  We can accelerate our growth as traders by mastering change processes and becoming our own psychologists.  It takes a vision and a plan to make changes and sustain them.

*  This week I will begin a new blog feature where I maintain a page of "trading notes" documenting observations from the most recent day's trading.  The notes will update relevant market measures and findings from some of the forecasting models that I've described recently, as well as interesting links.  Readers can feel free to add their own notes to mine via the comment feature of the blog.  The goal is not to pitch trade ideas, but rather to expand the information set that can help traders make better trading decisions.

*  Yet another way to expand your opportunity set is to find one new valuable information source each month that can become a regular part of your research/decision process.  Over the course of 12 months, you can greatly widen your scope and sustain your development as a trader.  When you peruse posts curated by such sites as Abnormal Returns and Quantocracy, focus on the sites offering the posts and not just the posts themselves.  Many times you'll discover great resources that way.

*  Looking forward to the speakers and networking time at the Traders4ACause conference next month.  My program off the program will be that I will bring a one-page write up of an original, valuable trading method and will exchange with all participants who are similarly willing to share their writeups.  It's not just broader networking, but smarter networking, that helps us adapt to changing markets.

Have a great start to the week!

Brett
.

Sunday, August 09, 2015

Forecasting Next Day Price Change in SPY

Every so often, I'll post on topics relevant to my own trading, in hopes of inspiring fresh directions for active traders.  This weekend I undertook an exercise to develop a predictive model for the next day's returns in SPY.

Above are the results for the simple model.  The data are divided into quartiles, with the top quartile representing the most bullish predictions and the bottom quartile representing the most bearish predictions.  The model consists of only three variables, which I would describe as:  1) the amount of buying and selling pressure for the previous trading session; 2) the number of stocks in the SPX universe showing price weakness over a multiday period; and 3) the number of stocks in the SPX universe showing price strength over a multiday period.  The model lookback period was January, 2014 - present.

Here are a few takeaways from the exercise:

1)  As in other models I've built, the predictive value tends to be at the extremes of the distributions of the predictor variables.  This is a way of saying that the relationship between predictors and forward price change is not a simple linear one.  Being able to model those nonlinearities results in superior price forecasts.

2)  The model is statistically significant, but accounts for a modest proportion of overall variance in next day outcomes.  The model provides an edge, but significant uncertainty and randomness are present.  Even the best models require sound money/risk management.

3)  None of the three predictor variables is a traditional measure from technical analysis.  When I add traditional technical indicators to a stepwise regression process, the technical measures are generally not significant predictors in their own right and never add predictive value to the three model variables.  Parsing market data in unique ways promises unique forecasting power.  Measures most commonly followed by traders possess little forecasting power.

4)  Lookback period matters.  The model that is robust in 2015 does not perform well on 2008 data.  Before searching for predictive relationships, it's necessary to identify a stable lookback period that captures strong, flat, and weak markets.  All models assume that the stable regime of the recent past will persist into the immediate future--an assumption that will be incorrect at times.

5)  The hit rate for the top and bottom quartiles is around 60/40.  Again, it's an edge, but it leaves plenty of room for error.  The hit rate for the middle quartiles is closer to 50/50.  

6)  I have other models that forecast SPY price change over 3-5 day horizons.  There is value in lining up the daily forecast with the swing forecasts.

For me, the sweet spot of trading is seeing those occasions in which the action of the tape lines up with the anticipation of the market forecast.  The tape comes first:  I want to see expanding buying interest in order to go long and selling interest if I'm to be a seller.  When the forecast lines up with the tape, those tend to be occasions when it's useful to bump up the risk taking and extend holding periods for potential trend days.  An additional value of these models is that they keep me out of bad trades: I do not fade the market when forecasts are in the top and bottom quadrants.

For those who might be interested, Monday's forecast is the first positive one in several sessions, with a forecasted gain of +.11%.  That is in the second quartile; similar readings have been up 17 times, down 12 times for an average gain of +.12%.  My swing forecasts are neutral.  In all, Monday is not a day with a strong edge from the models.

Further Reading:  Integrating What We Know and What We Feel as Traders
.