Monday, November 30, 2009

Monday VIXology

This week should provide a clearer view of the Dubai World default fallout. Based on the results of the latest VXX PDQ update the technicals are poised for further market weakness.
Note that most of the Long signals are 2 to 3 days old and the median N days value for the current longs is about 5, which means signal re-evaluation on Wednesday will be required in order to confirm VXX momentum.
The VIX has jumped ahead of the VXX since July, although the Stockcharts.com ratio study below suggests that, at least short term, the VIX may be due for a little pause. With a confluence of the charts now sitting near or dead on the (LR30) linear regression channel 30 day mean, mean reversion traders are looking for a bounce.
In the face of these mixed signals, Monday's momentum catalyst is likely to be news driven and risk adverse technical traders are best served by waiting for the trend to be more clearly defined.
Finally, I note without comment a little item in the Financial Times that traders at Goldman Sachs suffered only one losing day during the 65 business days of the third quarter. On 36 separate days during the quarter, the firm's trades netted more than $100 million each day.

Friday, November 27, 2009

VIXEN Meets Old Crow

It's always gratifying to know that somebody is actually trading the VIXEN setup and one of my loyal readers, Old Crow, has forwarded 2 of his recent forays using his own version of the setup.
As with all my VIXEN posts, Old Crow has shown these trades on the Schwab Street Smart Pro platform.
He uses a few more MAs than I prefer and has turned off the pivot points, but this is what works for him. It's always interesting (to me) to see how different traders adjust my default settings and indicators to create their own unique trigger perspective and risk control signals.
Old Crow uses a number of smoothed MAs, a resident feature of Schwab and has applied the signal line (the moving average component of the MACD) directly on top of the chart in lieu of placing it in a study window.
He retains the parabolics and in the examples posted it's easy to see that the SAR provides a consistently reliable trend reversal signals.

Wednesday, November 25, 2009

The VIXEN Four Faces

With yesterday's volume at about 60% of the daily norm I thought it might be a good time to review some of the basic VIXEN concepts and at the same time show the disparity between some of the large asset class ETFs relative to the VIX crosses.
Shown above are the NYAD, Qs, IWM and SPY. I could have shown the DIA, the VTI, the EEM, etc., but I leave that to you, gentle readers, in explore over your holiday.
What's of interest here is the relative performance of these 4 charts on the VIXEN cross.
The duration of each concurrent trade is shown with the red and green horizontal lines and the vertical white line. . . stats are as follows:
Qs . . +.3 %
SPY . . +.2%
IWM . . +.5%
NYAD . . +54%
Yes, a little breathtaking on the NYAD, but that's typical short term behavior characteristic of a rally. . in this case 60 minutes. Too bad you can't actually trade the NYAD. . but, like the VIX, it's a statistic, not an equity or derivative.
One of the reasons I never trade the SPY, or any of it derivatives is reflected in the relative performance of the SPY vs the IWM vs the Qs. There's just less momo, less beta, less volatility, less risk, but also less risk/reward.
This particular example is a bit peculiar as the ranking of volatility is typically Qs, IWM and SPY.
I suspect that recent higher volatility in the IWM can be traced to it's current lagging the other indices in new highs, with the result that bullish surges tend to favor the IWM as it tries to play catch up.

Tuesday, November 24, 2009

Facing the Music

Monday's PDQ forecast for a bullish VXX turned out to be a bit off. Actually the VXX did rise, (temporarily at mid-day) but that was after it had fallen almost 5% at the open.
OK . . , no excuses . . the VXX long signals were bad. In an effort to determine why, I went back and looked at each of the pair trade signals. Since the PDQ produced a mixed bag of signals, clearly there were some attributes of the correct signal pairs that need to be integrated into the incorrect signal pairs.


My investigation led to a continuation of my earlier post on the zero line rejection behavior of the z-score. While the Z-score rejection signal does offer a useful stop, the salient discovery that will first be integrated into the PDQ is a fixed time stop based on the optimized N day value for each pair.
I spent a few hours manually toggling through each of the PDQ Dashboard performance studies and virtually every one could be enhanced (some significantly) by applying this N day stop. Now these results are a bit surprising (to me) because it suggests that there really is a unique underlying timing cycle for each pair.
While this sounds deterministic, we need to factor in the all-important caveat that the PDQ model is designed to be adaptive and the N day cycle is really only valid as long as the equity curve of the pair trades remains positive.

Monday, November 23, 2009

Monday VIXology

Just off the cuff for Monday's VIXology the charrts look long the VIX and short the Qs. The VXX, the tradeable derivative of the VIX is a bit more ambiguous, which is reflected in the signal(s) currently being generated by the new PDQ Dashboard (Lite version).
The dollar is also looking bullish via the short term technicals, although we had a close call with the last UUP Dashboard BUY, so caveat emptor on this one and the wise course of action is to either 1. Stand back until an uptrend is actually confirmed or 2. Trade small and be ready to bail on a tight stop.
As mentioned above, my continued noodling with the PDQ has led me to evolve it's format into a Dashboard with 5 pairs rather than the original 8. Through on-going backtesting I've found that using the 5 highest linearity correlation pairs produces signals just as reliable as the 8 pair model with reduced noise.
I got a bit sidetracked last week on some new refinements of the PDQ including the zero line rejection signal and never got around to the VXX/UUP study I promised earlier. I'll rectify that delay this week and also show some graphical studies of a pairs trading model that will, I venture to say, give credence to the old adage that "a picture is worth a thousand words."

Friday, November 20, 2009

Qs VIXEN

Here's another example of the VIXEN setup, this time using the Qs as the underlying. With the NYAD in a VERY slow ascent (.09 total rise) as of one hour pre close, there were still a few bucks to be pulled out of the Qs using the VIXEN.
I'm not showing the NYAD chart today as the % change was so small as to be barely noticeable and effectively untradeable.
Although these 5 trades only netted $ .11, .16, .12, .11 and .06 respectively, the total gain over a 4 hour span was $.56 (- commissions), a respectable return IMHO considering the total range of the Qs is currently $.75.
The VIXEN produced a very nice ride from 11:00 to 12:30, and when that trade reversed it looked like the 11:00 lows might be revisited. The turn at 13:30 demonstrates the utility of confirming signals as the parabolics fired a BUY (or cover) and the 3 MAs turned upslope.
Readers who clued in on the SMA8 high and low channel that I mentioned earlier can also see how the channel crosses helped signal trend change.

Thursday, November 19, 2009

More on NYAD / VIXEN

This is a little variation of a setup I previously profiled using the VIX and NYAD to signal trade entries and exits in GE. The setup works for a variety of underlying stocks and/or ETFs, although the large caps and the indices seem to work most consistently.
I like GE because it mimics the overall market so well while at the same time displaying a beta that makes it trade more like a Proshares ultra than a stock. I also like the penny option spreads and the huge open interest across a long option chain given the fact that this is a $16 stock.
What's different about this setup is that instead of looking at parallel GE charts, we looking at the VIX crosses on GE (or your choice) and on the NYAD. You would intuitively think it was important to keep the underlying target a vehicle that's convergent with the major indices and not something like the UUP, which is divergent, but the simple remedy for that problem is to make the VIX the underlying on right chart with the NYAD as the overlay.
These setups can provide great little intraday scalps with the caveat that you always need to have one foot out the door in case things go awry.
There was some ambiguity in the underlying technicals on the NYAD/VIX chart (red circle) which will now prompt me to review and retest those indicators prior to future trades.
*
Finally, after over 3 years of daily posting on this blog I've decided on a new visual clue to reflect my attitude towards the markets. Gone is Homer's Scream, now replaced by a photo of my humble support team. . . Lopaka, Wilson and Tonto.

Wednesday, November 18, 2009

Zero Line Rejection Stop

In my ongoing refinement of the PDQ Dashboard I've been studying each pair chart individually to see how these trades actually play out.
One phenomena that keeps popping up is what I'll call z-band zero (zbz) line rejection and it can be seen clearly on the EWC/VXX performance chart above.
Note how each of the winning trades (green lines) move smoothly from one band to the other with only minor chatter and a clean break of the zero line.
Now note how the 2 losing trades had a completely different profile, characterized by multiple bounces off the zero line, which also produced open trades of 43 and 26 days respectively, whereas the average duration of the winning trades was only 9 days.
Examining each of the losing trades a bit further shows that if the trades had been exited when the z-score retraced 10% off the zero line both trades would have actually produced marginally profitable results and closed after only 7 and 12 days respectively.
Lesson learned. A new risk management stop (BZB zbz) will be added to our current stop menu that will incorporate a fixed time stop and/or a 10% zero line bounce hit.

Tuesday, November 17, 2009

The Simple Life

I'd be lying if I didn't confess to being a mite perturbed about the dismally incorrect signals that the technical indicators have broadcast recently including the VXX PDQ model posted Monday. But HEY!, it's a learning experience and I'm always striving to learn from my mistakes.
Two years ago I had the opportunity to actually talk privately with the Donald for about 15 minutes and, despite his public image as a bit of an arrogant arse, I found him quite personable and open. I was lamenting a recent trading slump at the time and he simply said, "Don't beat yourself up, just promise yourself that you'll learn from this down period, apply what you learn and strive to to do better next year." Nothing profound there, but the words stuck with me and have served me well as a sort of self-improvement mantra.
Which brings me to today's post. While not seeking to sound like a broken record, I always go back to the tried and true if the technicals are confusing.
And, in my toolbox of high probability daytrade setups, my favorites are the oft posted VIXEN and NYAD crosses.
Shown above on 5 minute bars is today's picture perfect launch of the GE/NYAD cross at 10:45.
As of Friday's close I was prepared defensively for a Monday down day. When it became apparent that we were much more likely to see a bullish trend day, I wanted to gooble up a few scraps and, for my money and risk tolerance, these are the easy money trades.
With the 3 MAs in the lower window continuing to track higher this was a multi-hour trade and I was expecting a fade late in the afternoon as the NYAD flattened and then ultimately turned down . . . so I was looking for a signal to exit
I haven't mentioned this before but one of the very short term technical trackers I use is a channel of the SMA8 highs and SMA8 lows (2 blue lines). Through the course of a few thousand trades I've found violations of the channel extremely helpful as both entry and exit triggers, and in today's example (white circle) the violation of the channel corresponded with a parabolic SELL, the cross of the 3 lower SMAs and the beginning of the afternoon downtrend.

Monday, November 16, 2009

Monday VIXology

This is a little different spin on the VXX PDQ Dashboard. Last week I profiled a VXX basket using the currencies while today we're looking at a basket of ultra 2x ETF pairs. The net effect is that the ultra ETFs have much higher inherent betas and so the pair betas average out to a little over -1, considerably more consistent and muted than the VXX currency pairs.
And, proceeding with the weekly update, here's a little different VIXology mix, which includes the VIX, VXX, UUP and, of course, the Qs.
Thursday's UUP basket update post with a strong BUY signal turned out to be a one day wonder (actually 2 day since the PDQ fired a consensus BUY on Wednesday's open.
Our 5-6 N day target now looks seriously in jeopardy and the initial BUY was stopped out on Friday's open for a measly .05 gain based on the ELE stop which closes the position if it has retraced 50% of it's previous gains.
Of interest here is the mirror image of the VXX and the UUP charts with the VIX chart understandably showing an exaggerated relative trading range no doubt due to the fact that VIX is not a price, but a statistical reflection of volatility in the SPX.
I'll run some comparative PDQ pair studies to assess the relative risk/ reward of trading either the VXX or the UUP against the same pairs.
To be honest, I should have seen the striking similarities in these charts earlier, but HEY!, I'm old and there's lots of distractions down here in beautiful Oceanside, CA so I'll try and redeem myself later in the week.

Friday, November 13, 2009

GE PDQ Dashboard

Readers who have followed me for a while know that the Qs and GE make up a major portion of my portfolio. I trade each using a variety of strategies including the VIXEN and NYAD crosses for daytrading, premium decay on longer term positions using a number of option setups.
As part of my ongoing research into non-correlated ways to mine a few more bucks out of GE I've added a GE dashboard to my expanding repertoire of PDQs, with the eventual goal of having a robust PDQ Dirty Dozen.
The PDQ above reflects the current state of that research. As with many other of the PDQs, this GE version is a mix of stocks and ETFs and I can make a pretty good case for the logic behind each of the pair components. They all share a high linearity correlation with GE and, although a wide range of N days and trade frequency is evident, I was actually looking for this type of robustness given GE's relative volatility.
Current signals are moderately short.
Volume was running about 35% sub par Wednesday and Thursday, which typically causes short term volatility to rise, and that caveat has be factored into the current signals.
Next week we'll look at a graphical interface for the PDQ that dramatically displays the robustness of these forecasting models.

Thursday, November 12, 2009

My New Dollar Basket

This is an update of my previous dollar basket post. The UUP is, of course, the Powershares double US dollar (bullish) and this time around the focus is on other currencies and short term bonds (SHY) as the wild card.
The PDQ Dashboard has gone through quite a few refinements since the initial post and this basket should be considered an alternate to those earlier PDQ components.
The relative beta of the CYB (yuan) is obviously out of sync with the rest of the basket so this signal bears a closer look.
Also out of sync is the relative SHY (short term bonds) beta.
Both CYB and SHY display consistently high linearity and an identical 19% P&L over the past 6 months, and both have a similar trade frequency. And, as in the case of the FXY basket weaving post, in both cases an examination of the actual returns for each side of the pair trade reveals that UUP is the side to trade (the CYB and SHY trades only generate 1% and 2% of the 19% return respectively. For the above reasons I regard these beta skewed pair partners as valuable forecasters of UUP momentum.
The UUP fired a number of long signals on Monday's close and I'll be keep a close watch for exit signals. Since virtually all the PDQ components have 5-7 N day values (Z-score band cycle side to side) I would expect reversals sometime in the next 3-4 days.
Here's a peek at the current UUP call option string. Virtually all open interest is at the ATM 23 calls and despite $.15 net gain in UUP the 23s fell a nickel. A lot of interest here and the early action today was focused on buying the Ask.

Wednesday, November 11, 2009

Dynamic Ticker

As a follow-up to yesterday post here's the Schwab dynamic ticker in action. Also shown on the left is one of the 4 setup menus. The set up menu time frame indicates 1 minute, but the ticker ebbs and flows on a tick by tick basis and as a result any turns in market momentum are immediately obvious.
You can set up a number of watch lists for the ticker to monitor or, if you've got spare screen space, you can set up several of these tickers to run concurrently.
Some traders are data stream oriented but my brain is hard-wired to work off pattern recognition so keeping a couple of these little nuggets tucked away on one of my monitors and setting them to track the GE PDQ Dashboard components or the Qs PDQ Dashboard components is like having a good buddy looking over my shoulder while I trade.
Other platforms may have features similar to the dynamic ticker, you should check on yours.
BTW, after a few exchanges with my trading buddy The Clueless One over the weekend I've modified the 3 FreeStockCharts on the right blog panel to be in sync on 2 minute bars. Also, keep in mind that the Qs (top) chart is not a VIXEN setup while the lower (GE) chart is.

Tuesday, November 10, 2009

Schwab VIXEN / NYAD Setup

This post is in response to a number of reader queries about the GE VIXEN. To help clarify how I use this setup in real time I'll show the one of the screens I watch during the day to trade GE.
Frequent readers know that my daytrading platform of choice is Schwab Street Smart Pro, which I favor for a number of reasons including the real time new daily hi -lo count scrollers (shown here tucked between the 2 larger GE charts) and the ability to drop comparison tickers on top of primary stock or ETF tickers. In this case I've got the VIX laid on top of GE on the left and the NYAD laid on top of QLD on the right. I've set the VIX and NYAD to display as mini-candles, although I can just as easily display them as a line or hi-lo bars.
I really like the Schwab NYAD, which is displayed as a ratio of advances/declines.
Although I posted the GE/NYAD studies a while back, it's still a killer application and I use it in conjunction with the VIXEN to catch as many setups as possible (and also to provide signal confirmation) I watch both the GE/VIX and GE/NYAD setups in parallel. There's over a dozen posts each on the GE/VIXEN and the GE/NYAD in the archives so , if interested, scroll back and read a few of them for more details on how the dynamics of these trades typically play out.
I'm not showing the actual trading window here, that's on another screen along with the scrolling 3 finger lead, the 7 finger lead, a dynamic momentum signal (proprietary to Schwab) and a few other windows (news, portfolio management, and a dynamic watch list that let's me replace GE with any other stock or ETF that I choose).
These setups work well for me because I trade only a handful of stocks and ETFs (actually less than 10) and I prefer to carefully study the dynamics and idiosyncrasies of my little basket rather than jump all over the market chasing the fast money.
Not for everyone, but it works for me.

Monday, November 09, 2009

Monday VIXology

Here's a little variation of the VXX Dashboard that I've been fooling with that involves pairing the VXX with a basket of currencies. While the model doesn't produce those 90% + linearity returns seen in some of the other PDQs the P&L is certainty impressive given that this model reflects only the last 6 months and that these are currencies.
Of particular note is the green status of the P&L status for each and every one of the pairs. I'll be tracking this little nugget in the background for the near term as the returns so far have been very impressive.
We did get a mean reversion move last week as the Qs worked off oversold levels and the VIX settled back to the LR30 upper channel band. With most of the volatility surprises prompted by earnings reports now in the past the technicals are strongly suggesting a continuation of last week's bump.
And, just to put things in perspective I've included a comparison of the VIX and VXX below. Peaks and valleys coincide but notice the LR30 down slope of the VXX indicative of a net decline of volatility in the VXX relative to the VIX.
Finally, a little side note here on the NYAD proxy gadget added to the blog on Friday. I neglected to mention that I set the gadget to display at 2x the amplitude of the NYAD in order to provide clearer indications of impending trend changes. There are two smoothing algorithms on top of the underlying detrend indicator and a doubling of the initial raw signal value to accomplish this goal. Keep in mind this chart does not necessarily reflect prices, but rather the momentum of the NYSE. With this template in place I can now create a NYAD-like signal for virtually any index and I'll add one for the Qs and the IWM in the future, just to test their forecasting value.

Friday, November 06, 2009

NYAD Proxy Gadget

I've added another new gadget to the right side of the blog. This time it's a real time proxy for the NYAD, the NYSE advance decline line that I use to confirm 100% of my trades.
I've mentioned before that I could probably daytrade quite successfully using just the NYAD and the pivots but I like the conformations offered by the MAs and the parabolics to take the edge off.
Just as a real time feed for the VIX is unavailable in FreeStockCharts, it's the same story for the NYAD. I've managed to solve the VIX problem recently by using the VXX, but the NYAD proxy required quite a bit more work. I've placed it between the Qs and GE VIXEN feeds on the blog so you can watch the NYAD dynamics on all 3 charts simultaneously.
The violet line is a composite of 6 different indicators and I regard the coding as proprietary, just because its development required so much time.
I've added a little momentum shadow behind the violet line to show relative strength and topped off the indicator with 7 (yellow) and 14 (blue) period linear regression lines that clearly indicate the direction of the short term trend.
Many thanks to Worden Brothers for making this product available. 3 years ago this level of technical refinement, user programming options and a real time feed would have cost you at least $60 per month.
Also thanks to readers for feedback on the GE VIXEN Trader, which seems to have generated a fan club of it's own.
I failed to mention a reliable entry/exit setup in the initial post so I'll quickly describe it here as shown by the green and red vertical arrows below.
Although I use the GE/VXX cross as an initial entry I'll add to the position when green arrow conditions present. On the other hand, when in the position, I'll exit when red arrow conditions present.
The red, green and yellow lines in the lower technical panel represent a 3 (green) and 7 (red) Moving average, while the yellow line is a FSC smoothed detrended price oscillator (DPO) that has been individually "tuned" for the Qs and GE charts shown here.
As with the NYAD gadget, I've also added a LR7 (white line) to the yellow line DPO to clearly display short term momentum.

Thursday, November 05, 2009

VXX PDQ Dashboard

This PDQ Dashboard is the logical extension of my explorations into the volatile world of the VXX, an ETN based on S&P VIX futures. If only there were VXX options, then we could trade a derivative of a derivative of a derivative and wouldn't that be fun!
For this study I've turned off the risk buffers (ESP, ARMM and ELE) which would have otherwise put the RTH pair in OUT status.
Two other factors to recognize when looking at this Dashboard are . .
#1 . . the inherent beta of VXX relative to the rest of the pack and
#2 . . the fact that these are all divergent pairs (as opposed to convergent)
As such, the amplitude of the daily MoMo is going to get exaggerated much more so than the Qs or the FXY dashboards and therefore requires its own unique risk threshold baselines. While we're fiddling around with refining that algorithm the raw signals for the VXX generated by the dashboard have been consistently reliable over the past two weeks and will become a regular feature of future Monday VIXology updates.

Wednesday, November 04, 2009

GE VIXEN Gadget

I've added a new element to the right side of the blog panel to keep my daytrading posse amused. It's based on a FreeStockCharts (FSC) chart and uses GE and the VXX to signal VIXEN trades. The chart is set on 2 minute bars and tracks 45 bars or 90 minutes. Keep in mind that the VXX crosses have to be executed relatively close to the actual cross since scaling factors will literally move the cross point forward as time elapses. No looking back here.
One of the problems with FSC is that there is no live feed for the VIX or the NYAD, just EOD.
We can resolve that problem for the VIXEN trades by using VXX as a proxy for the VIX. I've mentioned the differences between the VIX and the VXX previously so won't rehash old ground here. Suffice to say that VXX actually trades at a market price, while VIX is a statistical value.
From a technical viewpoint, the VXX tends to display a bit more short term volatility than the VIX, but viewed on our little gadget, it's barely noticeable. The VXX is shown as a white line.

Here's a little snippet VIXEN trade from Monday's session, and displays how simple (and effective) these scalps can be. If you undertake this strategy you must pay attention to the trades or have fixed stops in place to protect your capital. Things can happen fast on these trades. My typical expectation is a trade duration that may last 10 minutes to 30 minutes, and getting caught on the wrong side of these trades can be costly.

Tuesday, November 03, 2009

PDQ Dashboard Update

The Qs Month End Tickler, which bought Friday's close at 41.00 did manage to eke out a gain today, but, being the nervous Nellie that I am, I bailed on the position at 10:55 as the NYAD parabolics fired a SELL at $ 41.40 and booked a nice little weekend gain. According to the system parameters I should have held the position 2 more days but I wasn't comfortable with the deteriorating technicals and one of my guiding mantras is "When in doubt, get out", and I did.
.
Here's a peek at the status of the PDQ Dashboard per Friday's close as we've made some changes and (hopefully) improvements to the signal line.
#1 . . The pairs sets have revised - NEM has been deleted and UNG will likely follow as the N day cycle is too long and the number of trades is too small. VIX has been replaced with the VXX. The VXX is actually tradeable, with a daily volume around 1.5M and a typical .02-.05 spread. I had been noodling around with an algorithm to convert the VIX into a more in scale contrary indicator but, after a number of metric studies, I'm going with the VXX for the time being.
#2 . . the current Z score column has been augmented with a signal showing the current direction of the z score trend, which should help us determine if the trade triggers are getting more probable or more remote.
#3 . . The Final Signal column now is overseen by 3 programmable risk buffers which can be turned on or off independent of each other. These include a % loss stop, a % trailing stop and the position of the current equity line relative to the R2 slope based on a variable lookback period that coincides with the optimized N lookback period for that unique pair.
#4 . . El Jeffe', our programming guru, has added 2 little heat bars below the data columns to display an original signal momentum and a Final Signal momentum.
.
Next to come . . . 2 PDQ models - one short term ( N less than 10), and the other longer term.
Also, a breakdown of net P&L for each side of each trade, which will help us see the relative attraction of each pair setup. New PDQs will be limited to 7 pairs including the VXX, which will enable me to run trade date sequencing on an EXCEL tracker, as with the case of the FXY study.

Monday, November 02, 2009

Monday VIXology

A strong change of character this week as the VIX displayed some razzle dazzle and a net 31% gain. Now riding well above the LR 30 upper channel, the VIX has retraced back to July 9th levels. Looking technically overextended on both the RSI2 and MAs, the VIX can and has played some nasty tricks on pure technical traders before so caution is advised here with my personal bias being to just focus on intraday patterns until we see how the next trend develops.
Conversely, the Qs look ready for a bounce at the current -.43 levels although we could as easily get a repeat of the late September, early October double bottom pattern so, once again, caution is advised before getting too frisky on the long side.
The PDQ dashboard is short the Qs in one position . .the rest of the portfolio is flat.
.
Below is an update of the VIX pivot bands.
The emerging scenario suggested by this indicator of pivot volatility is bullish (for the VIX).