Thursday, January 26, 2017

More Signs That a Sell Off Is Looming in SPY - 1-26-17 Stock Market Update

With the break to new highs in SPY, there are now some intermediate term bearish divergences which are prominent.  My perspective here is that stocks have more downside risk over the next couple months in comparison to upside potential.  With that said, the most comparable environments that are showing up in my scans are environments where stocks may spend more TIME in a general uptrend, but there may be a rapid sell off which more than wipes out the lazy gains which occurred preceding it.

Click on Table to Enlarge

The table here shows the results of a scan with the following filters:

  • New 52 week high
  • Close > upper bollinger band
  • VIX high is < 12
  • VIX close is < 11
Then subsequent readings (3 of them) which occurred in clusters after the first reading were removed.  This left us with basically 2 past market periods.  
  1. Very late 2006 into early 2007 - which resulted in the very sharp sell off right at the end of February and into mid March.
  2. Mid 2014 from late June to July - which was followed by a somewhat similar brief, but sharp drop into August 2014.
Looking at the table shows a strong downside skew over the near term, however, with volatility being so low, that does not really portend a big move.  Even looking out to 2-3 months the skew is negative, and still relatively strong at 2 months.

Another notable part of the info in the table, is the very large skew in the MAX gain/loss of the VIX.  There just has not been much history of the VIX staying at 10 or 11.  So if you trade volatility, there may be an indication for action here.

In fact the expected swell in the VIX may be the most significant aspect of this set up.  There is a profitable option play on my model based on this scan.  Buying an ATM put with 2 months until expiration and using a 50% limit order gain after entry worked well on average.  But there are only a couple similar instances.  The point is more that, even a moderate 1-2% decline, in combination with a possible larger % increase in the VIX at the same time, could give the put options a boost to those levels of paper profit.

Part of the question is then WHEN a decline would be anticipated since these lazy markets can go for weeks without much downside.  I am not an independent expert on cycles, but some of the information which I have followed for long enough to feel there is some pragmatic use in, suggests that the upward current could crest somewhere around next week.  Then there may be more downward current (or at least less upward current) into early March.

So my general take here is that there will not likely be much further % gain in SPY over the next couple weeks, and then there may be a sharp but brief sell off in the Feb-March time frame.  Given the low volatility and the couple past comparable time periods, if a decline occurs, it may be expected to be mild, like 2-5%.

I am choosing to make an option play here, but certainly a paired limit/stop order on a SPY leveraged ETF, or even on a VIX tracking ETF could be very sensible given the skews in the data over the next couple months.  

If there are further specifics on trade orders for the ETFs, comment or reply, and I could get something up here.


Pete

Monday, January 9, 2017

Bearish Trade Stats From Recent Set Up

Click to Enlarge Stats

On December 9th and December 13th a trade set up triggered which has had a consistent and marked bearish skew looking ahead a couple months or more.  I did not remark on it at the time because the technical analysis was pretty clear to me that any immediate pullback was likely to be modest and that further bearish divergence would likely develop.

However, I will now make not that this set up is active and still in a range where it could be acted on,

The scan is as follows:
  • daily MACD position is UP
  • daily MACD lines are both above 0
  • SPY closes up for the day
  • there is a total put/call "sell" signal (5 day average is below 1 st dev bollinger band)
I removed subsequent readings from clusters with multiple signals after the first in a 2 week period.

The stats above are for an ATM put option with 2 months until expiration if purchased at the signal, in this case on Dec 9th.  Past instances show peak expected value at 120% gain on the option, but the 90% gain mark puts the win rate at 67% with only a slightly lower expected value.  So that is the stat shown here.  The stats are nice here, and would suggest that a 226 to 228 strike SPY put could be purchased with a standard february expiration, and then set a limit order of 90% gain to exit.  There would be about 2/3 chance of winning based on the stats, and the expected value is around 26%.

Click to Enlarge Stats

In this case, the equity stats have a strong skew and a nice trade set up as well.  Using the simple methodology I have designed for trading these situations, you enter inverse on the set-up and then set and equal percent stop loss and limit gain order.  If neither are hit (which is MOST common) you exit the trade a predetermined time.  In this case it is a 2 month time exit from the initial entry, which would be Feb 9th.  

The stats shown above would be for a short of SPY and then use a 6.25% limit gain and stop loss or exit Feb 9th.  Basis the SPY etf the expected value would be about 2.25%.  But the stats justify 3x leverage.  And so the theoretical stats there would be about 6.75% expected value.  That is pretty nice for an "equity" play when the market is anticipated to fall.

In this case in you buy a 3x inverse etf like SPXU then the stop would have to be 18.75% away from the entry and the limit order the same.


So these stats are certainly actionable by my criteria for trading.  And I also looked at times when there was a put/call "sell" signal at a new 52 week high, which also was the case in December.  And there was a bearish skew looking out 2 months, which also showed profitable stats using a limit order of 140% gain on the ATM put option.  I simply note this to say that even though stocks are at all time highs, the signal here still seems to be legit.  

And now that some time has burned away and some divergence has developed in the technical analysis, I think there is higher chance of stocks moving down with minimal time decay on the options.  In fact the option price is lower now than at that point, and so the value may be better.

That being said, I think it is likely that SPY could move higher for a couple more days to create an hourly time frame bearish divergence.  If so, that could be a fine tuned entry point.  I have a limit order to enter a position based on a move up to the 228 region in SPY.


Pete



Multiple Time Frame MACD Divergence In SP500 - Correction Or Bull Market Top Ahead?




Click on Charts to Enlarge

The charts here are the SP500 cash for the top three (monthly, weekly, daily) and SPY for the bottom chart, which is 60min candlesticks.

The monthly, weekly, and daily time frames are all currently displaying bearish divergence at the recent high on Friday.  Given the confluence and time scale of these divergences, I believe that price may be near a crest.  While tops can be sloppy and drawn out, we have already been seeing that for about 2 years.  

Rather than prognosticate on causes and projections and patterns, what we objectively are seeing is a divergence which truly indicates a "slowing" of the trend.  And the divergences so often occur preceding major changes in direction.  So I imagine that like a ball thrown up into the air will gradually slow before it reaches its peak and then begins to travel back down, the rising market will experience progressive slowing down, until it reaches a peak and turns down.  The smaller time frame divergences, I imagine to help identify the finer timing or reaching of the zenith in price.

So much for theories.  But what has happened in the past under some similar circumstances?

I ran a scan on SPY going back to late 1995 where there was a weekly, and daily time frame MACD bearish divergence concurrent with, or within a month after a put/call ratio sell signal like happened a couple weeks ago in the current market.  I also went through and excludes most days after the first signal day if there were a cluster within a 2 week period.

There result is that there was a negative skew in forward return looking out for a couple months.  It was not real lopsided, but in anycase it does support the idea that stocks may be near a relative top.  From a probabilistic standpoint is the only way we can meaningfully approach it for trading.

The skew is not real strong but is most prominent looking ahead about 3 days and at 1 month.  At 1 month after the past signals, the average closing return has been a mild negative at -0.25%.  And the MAX loss has been about 1.3 times the size of the MAX gain.  The numbers here are not strong enough for me to suggest an inverse trade based on those stats alone.  

So there is not a lot actionable here from the trading front.  But for those who want a larger market context for incorporation into trade selection, hopefully this helps.


Pete