Saturday, May 31, 2008

Qs Weekly Update

Here's Trading Day's spin on the Qs' prognosis. Good site, good tools and free.
Above, 3 linear regression studies (30,11,3) on weekly and monthly update bars.
For the week, Qs down .45 or 1%, DIA flat, SPY down 1%, IWM down .5%
The Qs have flatlined on the weekly bars for the past month, with a slight positive skew.
On the daily bars, mean reversion on the LR30 would bring the Qs back up to 50.25.
Weekly bars have the Qs riding the upper LR30 channel, and with the current lateral consolidation in progress, the path ahead offers few clues, other than the mantra, "never short a dull market."
The monthly bar closed enthusiastically above the LR30 mean and is currently upslope in a mirror of the pattern that developed mid 06.
The short term path of least resistance for the Qs is currently UP in a repeat of last year's seasonal pattern. Unfortunately, the market will always do the unexpected, and this year's fundamental milieu is decidedly different from last year. As suggested several weeks ago, the rise in the Qs may be a case of the lesser of a few evils - little financial, housing and retail exposure. A50 has turned up again and the A200 has once again broken above the 200 DSMA. Both the A50 and A200 are showing positive momentum without breaching overbought levels.

Friday, May 30, 2008

Qs, TICK and NYAD divergence

An interesting indicator divergence today that's worth a closer look (2 minute bars).
I like to watch the carry over from the previous day's close on the 10 and 20 MAs on 5 minute bars. This is typically a good indicator of whether the short term trend will continue or whether a pivot high, pivot low has transpired the previous day. Today's TICK and NYAD (I always exclude the first 5 minute bar) continued the downslope which began about 14:20 yesterday. The downslope is most pronounced in the NYAD, which I regard as the better momentum indicator than the TICK, which is more of an impulse monitor.
What's interesting is the divergence between the NYAD and the Qs. The IWM chart (not shown) was a mirror image of both the NYAD and TICK today. I consider that type of positive alignment a precursor for high probability day trades and the Qs divergence to the upside looked like a shorting opportunity once upward momentum stalled (first line of resistance at R1). The bullish pop in the Qs and NYAD at 10:24 (on increasing volume) looked like we might get a bullish move in the overall market. . .but, as is often the case with these pops, there was no follow through and the smart money went to the short side. Case in point: watch the NYAD.
And then there's this little gem on WM this morning.
Here are the 2 and 5 minute charts side by side.
You don't need the TICK or the NYAD to show you the way on this one.
As mentioned above, the 10/20 MAs stay wide apart at the open, suggesting more downside to come. But the interesting indicator is the RSI, which goes to the zero line on the 5 minute and stays there. This is real selling pressure, suggesting institutional selling in progress, and provides a relatively low risk daytrading entry (short), with the goal of exiting on the first sign of relative strength. The trick in using the RSI is not to BUY on the cross up, but on the cross down. The reverse applies with shorts - exit on the cross up, enter on the cross down.
In the WM 5 minute instance above, the RSI crosses above the 30 level at 10:15 and gives us a picture perfect exit for our short.
One problem with daytrading stocks in this price range (>$10) is that although we may get dramatic signal confirmation, the profit potential is typically limited. In this case, a good short entry at 9:38 and cover exit at 10:15 yielded .18 for 39 minutes exposure = 2% return. WM is starting to look like a "broken stock" rather than a bottoming stock and if institutions decide to get bearish, WM may be a casualty with the potential for another 50% down. Tread lightly on this one.

Thursday, May 29, 2008

RSI 2 and the Qs

I don't want to overlook the indicative power of the RSI2 for the Qs. Here's the daily study of the Qs looking back 5 years based on the same TS code as the Tuesday IWM RSI post. The settings are optimized to 2,28,80,28,80 possibly reflecting the variance in beta between the IWM and the Qs.
This is a fairly robust system and with an overall performance > 80%, it may be worth taking marginal swing trade positions (IMHO). With 242 trades over the past 64 months, that's about 1 trade a week-- a time frame I find almost comfortable. The number of longs versus shorts maintains a good balance and the win/loss ratio on the long side is particularly attractive. With an average of 2 bars in the winning trades and 8 bars in the losing trades, a stop loss/trailing stop should only improve results and reduce overall exposure time. Ditto for the short side. The intraday drawdown on both the longs and shorts needs to scaled back for more risk adverse traders, and again, stops can help achieve those ends.
Further backtesting (with stops) is suggested and additional confirming signals may also improve performance. Any conditional changes to the system would likely reduce trade frequency.
I utilize the RSI2 daily bars as a fractal momentum indicator to skew my intraday trades. If momentum is positive, I'm inclined to buy dips and upward pivot crosses. If momentum is negative, I'm more inclined to fade the rallies and sell pivot breakdowns.
It's also possible to build a short term options trading model based on RSI2 Qs behavior, buying near term calls (or selling puts) on positive momentum and buying puts (or selling calls) on negative momentum. I pefer to be a seller of options, rather than a buyer, but the risk potential is considerably greater and not every trader is comfortable with this type of exposure.
A couple RSI focused articles in this month's Stocks and Commodities. One tests the RSI to trade a basket of 100 ETFs (capital requirements and commissions make this impractical for most traders), and another expounding the virtues of RSI as an indicator that works. This is a statistical study and helps explain the why of the Qs support and resistance numbers used in today's study (28 and 80).

Wednesday, May 28, 2008

Qs Gap Fade

A classic easy money gap fade on the Qs this morning. (2 minute chart above)
Qs open 49.28, up .23 from previous close and hit R1.
I always watch 9:35--9:45 as a time pivot, looking for signal reversal and possible weakness.
9:38 the Qs penetrate R1 again, but fail to close above that level.
9:40 - a reversal bar after another failed attempt to get through R1.
9:42 - bar opens below previous close, another sign of weakness and I scramble to get short.
Although I enter the SELL at 49.28, I get filled at 48.26 as order flow jumps.
My first downside target is PP at 48.81, which seems likely with the other indices showing weakness (including the IWM).
10:18 - 10:24 The Qs form a little squat bar (not quite a hairy bottom) just a few cents off the PP. I don't like the looks of the long tail doji at 10:22 and at 10:26 with the bar opening above the previous close I cover at 48.92.
Time in trade: 9:42 - 10:26 = 44 minutes
Trade gain: 49.26 - 48.92 = .34
This was a clean gap trade that turned out well.
Looking back at my lower panel trade indicators I could have taken the RSI2 90 cross at 9:32 at 49.32 for an additional .06 gain, but my trading plan restricts me from jumping the 9:38 time threshold and I've learned to stick to the plan.
Leaving .06 on the table is a small price to pay for the SELL signal confirmation.

Tuesday, May 27, 2008

RSI 2 and the IWM

In response to a number of bloggers' interest in the RSI as a system trading tool, I offer the following case study of the IWM, which I have found to be unusually responsive to RSI signals. These results reflect daily bars over the past 3 years and as discussed last week, it will be interesting to check performance correlation over a number of fractal time frames, including the 10 and 60 minute bars. The system is currently on a BUY.
The cyan and magenta dots on the chart are the pivot high, pivot low (4,2) Show Me study.
The results are fairly impressive as I prefer systems that are at least 80% reliable. When it comes time to put real money behind your trades, there are few traders who can stomach a system with 6 or 7 consecutive losses and money management is absolutely essential in such systems. Most traders I know would walk away from any system after 3 or 4 consecutive losses.
This is a simple system with no stops. Drawdowns may be substantially reduced by adding both a stop loss and a trailing stop. The IWM is one of my favorite technical trading ETFs and it responds well to a variety of other technical signals.
This is the code in TS2000i, which you can import into 8.3 to test other time frames and settings. There's a slight edge in performance by executing SELLs at the close, rather than the following open (Market), while Exitlongs are best executed at Market.

Saturday, May 24, 2008

Qs Weekly Update

Above, the 3 linear regression study (30,11,3) of the daily and weekly bars.
For the week the Qs were down 1.81 or 3.6%, almost a perfect reversal of the previous week's gain. The markets are now moving lockstep with both DIA and SPY showing almost identical % declines this week.
My decision at the last update to stand aside any bullish Q moves this week turned out to be a good one. Fred's take on the hanging man formation was a good one, so. . .thanks, Fred. Many of the technicals that were positive last week have now turned menacingly negative on marginally increasing volume. With the Qs poking below the LR30 lower channel on the daily and back to the upper LR30 channel on the weekly, things could get interesting next week. At this point both the DIA and SPY have broken through daily LR30 lower channel support, though the IWM is a mirror image of the Qs. If this is a harbinger of emerging weakness, then the Qs have some catching up to do on the downside.
The A50 has fallen through the 20MA (no surprise) and looks poised to retrace at least back to the 50. The developing A50 pattern looks suspiciously like mid-Oct 2007, so we shouldn't put too much faith in the overbought RSI as you can see how the last turn down through the 20MA got resolved.
The A200 has also turned down, both through the 200 and 20 MAs. What's of note here is the magnitude of the drop relative to the A50. The "weaker" NDX components are clearly falling faster, suggesting that the run up mid-May may have reflected more short-term technical "piling on" than longer term investing. Again, the skew for next week looks negative.
I still have to review the Qs pivot high/pivot low position and the Qs relative trend position on the 60 minute RSI that I discussed yesterday.

Friday, May 23, 2008

Friday updates - short side component

For today's update of the IBD20 we'll look at the short side component of our little RSI study. The RSI value of the trading system is slowed from 3 to 4 based on backtesting studies that reflect a more robust performance with RSI4. Since many of the IBD20 components have suffered in the recent decline, the advantage of including a short side factor is clearly evident by comparing the Buy and Hold P&L versus the Trade Profit $.
As we work towards compilation of our trading system model, we have now added a stop loss on the long side and a short side component.
Now there are several tactics that need to be examined before we can deploy these various trading conditions to help maximize returns and minimize drawdown.
Remember, my goal is to utilize this system for daytrading on 5 and 10 minute bars. Other traders may be focused on different time frames for any variety of reasons. . .but this is where I feel comfortable. The testing framework explored in these posts may or may not fit those longer time frames and results may vary significantly, so be advised.
One approach is to simply deploy the system with a long and short side component with fixed or trailing stops.
Another approach is to assess the market trend based on a longer time frame and then ratio trade size to follow the trend. That is, long trades in an uptrend, short trades in a downtrend. One of my early mentors was Alexander Elder (Trading for a Living, Come Into My Trading Room, Entries and Exits) and he frequently suggested using a 1:6 ratio of time to bias trading direction in order to catch the trend. Based on that concept the trend position of the 60 minute bars will bias the type and size of our trades on the 10 minute bars. Following this bias should help reduce drawdown.
I'll explore this concept further in next week's updates.
I've added the short side to the XLE component update study as well and, as expected the Buy and Hold versus Trade Profit shows why the shorts can make a big difference. Of interest is the outstanding performance of RIG XLE DVN and OXY on the short side. These were top performers on the long side yesterday and today's results support my working theory that certain stocks will outperform others in both up and down markets.
Remember, my goal is to identify stocks/ETFs that best comply with my technical signals and trading system parameters. My goal is consistency of returns, not necessarily size of returns. I really don't care if the stock is going up or going down. . .they all do eventually. The goal of the trading system is to extract a low risk revenue stream from a stable of stocks/ETFs that can be ranked according to their performance reliability.

Wednesday, May 21, 2008

Thursday - RIG and IBD updates

The XLE component update shows some interesting developments. Most obviously, there's a lot of red in the open P/L column reflecting yesterday's decline in the sector. Remember, we're only looking back 500 bars, so in the case of the 5 minute bars, that's a little more than 6 days. RIG continues to hold top slot on the 5 minute, but has fallen to last place on the 10 minute, clearly a change in trend. (I'm not trying to mimic AdvancedGet dashboard).
OXY and DVN continue to show strength and XLE also keeps a top slot. If you think I'm working towards favoring the XLE over the component stocks from a R/R perspective, you might be right.
Just a footnote to yesterday's Qs fade trade. After I closed my position shortly after 11, the Qs rallied nicely and then plunged. It's been my experience that drops like that from 13:30
to 14:15 are very hard to catch, especially since the drop blew through S2 so quickly with a continued weakness down to almost S4 (not shown). You've got to have the reactions of a F16 pilot or nerves of steel to play those falls and I have neither. My hats off to anyone who managed to catch the afternoon drop. . I was happy to be slapping my new driver at Eagle Crest on a wonderfully cool day here in SoCal.
(My belief is that the best way to play those falls is with a system approach and autotrading, hence another reason for my current attention to the possibilities of system trading.)
For the IBD20 update BIDU has suddenly collapsed to last place, while HES TNH DRYS DNR hold strong. The idea of trading the top 10 as a basket with a culling factor to eliminate those that fall off the list (with no replacement) is gaining some traction. Again, this is a short term trading model with week to week adjustment as the IBD20 changes composition. At the end of the week, there may only be a few of the top 10 left, or there may be 10. It's like that box of chocolates. Next week I'll start following this concept in a systematic manner and post the results.
There is also the dark side which I haven't delved into yet and that is the idea of shorting the IBD20 as overextended, overbought candidates. This too is a work in progress, with the premise that once IBD20 components start to fall off the list, there may be significant downside potential. Whether such retracements can be captured on a short term basis remains to be explored.
So many ideas, such little time.

An Open Fade Lesson

Despite my recent spate of posts on systems, I have not given up trading the Qs. The emerging "system" trades component of the blog is intended to augment the revenue stream produced by the Qs. The markets are unpredictable and you can never project when trades will setup or how much will be gained (or lost) with each incremental trade. Since diversification is always a good survival tactic, the system trades will reflect a totally different approach to trading. . .one that eliminates many of the irksome discretionary issues. In addition, as mentioned last week, my spin on system trading is focused on a non-directional performance, and I hope this concept will become clearer in future posts as I move towards deployment of the system models. In the meantime (and for ongoing system trading education) see Rob Hanna's post on other system blogs and of course, check out Futures Truth.

Above is a picture perfect opening fade trade. Those who have followed me for a while know I watch the open on 2 minute bars watching the 9:35-9:45 window for a reversal or a continuation.
Today's action was classic.
After surging out of the gate and hitting R1 at 9:38 several things happened that foretold the reversal.
The TICK formed a slightly bearish doji at 9:38, followed by a bearish bar at 9:40
The NYAD formed a spinning top at 9:40 while the NYAD signal line turned down at 9:42
The failure of the Qs to follow through at 9:40 and a retreat back through R1 to close below R1 was bearish.
This was followed by another Qs close below R1 - bearish.
My stort term trading indicators( RSI, STO and STORSI) all fire SELL signals at 9:40.
I enter short at 9:42 at 49.51 with a retracement back to PP as an initial target.
The Qs hit PP at 10:14 and now the 10/20 MA have confirmed my SELL and are running wide apart. I close half the position at 10:18 on 49.21 and put in a stop at 49.38 for the remainder.
With the 10/20 still on a SELL I watch S1 for the next downside target.
The Qs finally close below S1 at 11:08 and I close the rest of the position at 48.96
Trade results:
Time in trades: 9:42-10:14 = 32 minutes; 9:42-11:08 = 86 minutes
Trade gains: 49.51-49.21 = .30; 49.51-48.96 = .55
A good trade with no regrets and I'm off for the rest of the day.

Wednesday - RIG and IBD system updates

You've probably noticed the new "look" of the blog. I found the skinny margins a waste of space and the color scheme was getting boring. Since many of the charts I post have a black background the black blog background should ease the contrast. Let's see how this new format works out.

Above is a daily update of our XLE component system. These results reflect an incremental move of 80 bars and show RIG holding strong while the XLE itself continues to show attractive returns.
Above is the daily update of the IBD 20. For this week IBD has dropped SQM FTI KWK and added MOS GEOI TNH.
As I mentioned yesterday, the negative P/L open positions might be reduced by use of stops.
The above scan reflects the same data set as the first IBD20, but with a 4% stop loss. The use of the stop clearly reduces the overall drawdown, while in multiple cases increases the net gain. The relative performance ranking of the list is effected to a minor degree and the working premise of trading the top 10 as a basket gains some support.

Tuesday, May 20, 2008

Tuesday - RIG and IBD system updates

The above reflects XLE component stocks performance for Monday. RIG continues to hold the top slot while XLE itself shows respectable results.
Above is the update of the IBD20 system test begun last. Issued last Monday, the IBD weekly top 20 ranked as follows: CLR GHM POT BUCY MA TITN BIDU RIMM ARD DRYS DNR CF SWN PCLN SOL SQM TTES FTI HES KWK. The original thought was to trade this group (or the top 10) as a basket to pull in some low risk short term returns. These results are based on a 500 bar backtest. On 5 minute bars, this equals 78 bars a day or 6.4 days total span. Although 16 of 20 stocks returned positve, its clear that money management stops could prevent open P/L situations like DRYS RIMM POT and PCLN. In addition, a companion short side strategy should improve overall returns, although in an uptrending market, these results tend to be muted.
With the markets currently showing some weakness, the results of these scans next week should prove informative.

Monday, May 19, 2008

Monday - RIG

Based on the Mr. Goodbar study last week XLE emerged as a potentially superior daytrading ETF while XLE component stocks also produced consistent returns. This morning's scan of the XLE components suggests RIG as a promising candidate using our RSI model.
For today's study I've also added a stochastics trading model (10,2,2,85,25) to demonstrate the high correlation between the 2 indicator based trading systems. The Schwab platform limits my ability to display more than one trading signal set at any given time, so for display purposes I have arbitrarily chosen the stochastics model. You can see the lead/lag nature of the RSI and stochastics as trading triggers on the middle chart pane below. I'll continue to monitor the behavior of RIG with daily updates for the remainder of the week. Since this is a very short term trading model I'll rate performance based on both a 100 and 500 bar look back. I'll also look at 5 versus 10 minute trading bar risk/reward.
I'm not putting cash behind any of these trades at the present time. This is a work in progress and until parameters of the system are confirmed, made more robust and refined (money management stops, short component, longer range backtest) these posts should be regarded as exploratory only. Reader suggestions are always welcome.
As mentioned in previous posts, http://www.futurestruth.com is an excellent resource for those interested in learning more about algorithmic system trading.

Saturday, May 17, 2008

Qs Weekly Update

With oil at $127, gold at $1000 and hookers at $5000 the Qs followed suit, finishing up $1.80 or 3.7% for the week.
Above is the daily and weekly 3 linear regression (30,11,3) study. The interesting chart to me is the weekly, displaying the solid penetration of the LR30 upper band. While the weekly RSI is clearly in overbought territory, the other technicals are showing a distinct bullish bias. Both the LR11 and LR3 shorter term channels continue to show positive breakout patterns. Further supporting the case for the bulls is the cross of the 10 and 20 MAs precisely on the LR30 mean.The daily chart is still showing positive momentum and that long tail hammer on Friday is not a sign of weakness. The technicals are all in overbought territory, but a basic maxim of the markets is that overbought,oversold conditions can prevail well beyond the limits of probability theory before reverting to the mean (looked at the VIX lately?). Both the A50 and A200 got a pop this week and the A200, in particular, is in overbought territory. Hard to understand the new highs in the A50, now at lofty level last seen in Oct 2007 before the big slide. Perhaps the NDX offers safety relative to the other indices which are hampered by housing, retail, banking, consumer discretionary and other worrisome components. We are , of course, for those of you relatively new to trading, entering the traditional summer doldrums of low volume and haphazard technical alignments, as all the big dogs are off in their private jets to the south of France, Corfu, Monaco, the Hamptons and other playgrounds of the guys who really control and move the markets, while us mortals are left to try and figure out if that channel cross is really support or resistance. (sorry for the attitude).
Based on the A50 and the A200 the "look" for the NDX is clearly upbeat, with little indication of impeding weakness (other than the overbought RSI).
HOWEVER, from a seasonal standpoint I do not consider this a good time to get bullishly aggressive, preferring instead to stick to the daytrading setups that have yielded a steady income stream with very little drawdown so far this year.

Within the next few weeks I will be switching over a new blog format which will be focused on algorithmic trading since, as I have argued in previous posts, I believe this to be the future for my trading. My intent to is feature 2 simple scans and daytrading systems with daily updates. For those of you with compatible technology in your platforms, you're welcome to follow along, although I make no warranty for your success based on the interpretation of my signals (see disclaimer on right panel of blog). No. . .this won't be a "chat" room. Trading signals will be dictated strictly by the systems as my eventual goal is to evolve the systems into an auto-trade mode on my favorite computer "Wilson", who is my otherwise constant companion during the day (think Tom Hanks in "Castaway").

Friday, May 16, 2008

Qs review and a quick TICK

A strong technical day with the parabolics showing the way. I was scared out of the first parabolic entry just off the open as the 10/20 MA was not confirming. 11:15 to 12:00 saw the Qs bumping against the PP pivot, with the TICK in a modest upslope. Finally, just before noon the Qs broke above the PP. The NYAD 10/20 MA crossed right at noon and the TICK also closed above the PP - all bullish signals. With the parabolic BUY confirmed with a second signal at noon, the odds for a surge long improved dramatically with R1 as a likely target. The 10/20 MA remained open on the Qs and the NYAD until 13:55, which I took as a signal to exit.
Time in trade 12:00-13:55 = 105 minutes (21 bars)
Trade gain - 49.42-49.70 = .28
I regret missing the first trade entry at 10:10 which reflected an opporunity lost of about .25.
Nevertheless, I stuck to the trading plan, followed the signals and am flat at the end of the day.

Arun asked about the TICK a few days ago and Wednesday I noticed this picture perfect example of TICK tell behavior. This is just a partial chart on 2 minute bars, but clearly displays the dramatic change in the 10/20 MAs on the TICK, NYAD and Qs. When the signals line up in sync like this example its time to bet the short side aggressively. The more tactical approach would be to go short at 14:00 and then add to the short at 15:00. I like to close these swoons at the end of the day, but with the 10/20 MA showing air into the close, there's also a good odds that weakness will carry through in the subsequent open. I prefer to play it safe and go flat. Another tactic is to close 1/2 the position at the close, letting the remainder ride overnight and using your accrued gain serve as a cushion against a possible pop.

Thursday, May 15, 2008

Looking for Mr. Goodbar - Part 4

Continuing with yesterday's theme and the emergence of XLE as the top performing ETF of our test sample, let's now break down the XLE into its top 10 component stocks and examine returns relative to the ETF itself. An initial assumption would be that the individual components perform better than the ETF since they have greater volatility than an index. We should also expect the individual stocks to entail greater risk. Whether any enhanced performance is worth a risk premium is the focus of this backtest. For the sake of comparison I included both 5 and 10 minute bars for the test, again to see whether volatility in the component stocks skews performance relative to bar size.
Our results are signicantly different on 5 and 10 minute bars. On the 5 minute, XLE essentially breaks even for our 500 bar backtest and places 9th out of 11 in terms of short term gains. Only the top 3 stocks, SLB, CVX and OXY show relatively attractive returns. For 5 minute bars our best strategy is to find the highest correlation performers and trade them in lieu of the XLE.
For the 10 minute bars, the situation changes. Here XLE places 5th out of 11 for our 500 bar test. And, while SLB and CVX hold top positions, APA and COP join their ranks, closely followed by the XLE.
These results support our initial conclusions from yesterday.
Bar size does clearly effect ROI and win/loss ratios. Different stocks are subject to different accumulation/distribution forces that one can only guess at. An important point with this type of analysis is that we don't have to try and figure out why this change in sentiment is occurring, we just have to know when it is happening.
This information will provide a useful risk management tool to trade not necessarily the top relative strength stocks, but the top technical alignment stocks (and ETFs).
Above are the IBD top 20 stocks, backtested on 5 and 10 minute bars with our RSI (3,90,15) long only, no stop model. I include this test as another example of how to select a technically reliable trading vehicle (stock or ETF). Note that in both 5 and 10 minute bars, 14 of the 20 stocks make money over the 500 bars test. Of the 6 stocks that fall below the zero line, only 2 are the same in the 2 time frames, CLR and SOM. That's not to say that these are weak stocks, the point to emphasize is that it's tougher to make money trading them using this system. Of course, you could find another system that would have completely different results. . and that's a valid approach. The goal is to find close technical alignment of the stocks (or ETFS) that you trade in order to maximize your returns and minimize your risk. Another approach is to monitor scans like the IBD20, which is a dynamic list of high momentum, high relative strength stocks, then cull the technical aberrants and trade the top 5 or 10 as a basket, depending on your capital, and risk management plan.
This is technical approach to trading that looks at "relative strength" in a different way. As presented over the last few days, this approach has been examined as a daytrading system using 5 or 10 minute bars (other time frames may also prove viable, but have not been evaluated). Looking at the charts is barely necessary if auto-traded. The need to monitor the NYAD, TICK and VIX, the number of stocks making new high, new lows, etc, etc. becomes unnecessary. Adding a short side component would increase the numbers of daily trades and, hypothetically, total ROI. I'll be looking at further refinements of this approach in upcoming posts as it provides the direction for my current trading evolution.

Wednesday, May 14, 2008

Looking for Mr. Goodbar - Part 3

First, an important reminder about the new US currency for those of you who still have any of the old stuff left.
For today's analysis we look at relative performance of a basket of high liquidity ETFs reflecting a variety of sectors.
The basket was tested using the same simple RSI long only stratgey (3,90,15) with no stops and looking at 5 and 30 minute bars.
The results are interesting on a number of points.
Most importanly, the best profits from this ETF basket are produced on 30 minute bars, a departure from yesterday's study where the 5 minute bars showed the best risk/reward. The sole exception is the XLE, whch showed a modest profit bias for the 5 minute bars. . . a situation which I'll explore further tomorrow.
Of the top 5 performers on 30 minute bars, 3 are the worst performers on 5 minute bars (DIA, SOY, QQQQ). This may tell us something about the effects of order flow management algorithms as well as the tempering effect of indices versus single stocks. Since one of our main goals is to develop some good risk/rewards candidates for daytrading, we might have to adjust our thinking about the relative saftey of the ETFs when we consider the tradeoff in terms of potential short term profits.
There are only 4 ETFs from this list that show positive 5 and 30 minute bar alignment: IWM, XLE, SMH and XLB, and strictly from a risk management point of view, these are trading candidates that may warrant a closer look.

Tuesday, May 13, 2008

Looking for Mr. Goodbar - Part 2

Today we look at the some of implications of bar size, again using Dr. Brett's watchlist.Attached are the backtest results of the same simple RSI system (with no stops) for 5, 10, 15 and 30 minute bars. Since I'm really only interested in finding the best technical performers, I'll cull results of the top five stocks in each time frame to see if any pattern emerge.
All tests range over 500 bars, so the 5 minute bars extend 6.4 days, while the 30 minute bars extend 38.4 days.
Top 5 stocks cumulative trade profit results are as follows:
5 minute: $2095
10 minute: $3417
15 minute: $4516
30 minute: $7473
Interestingly, in each of the four time frames 4 out of 5 stocks are the same: SLB,OXY,COP and CVX (all energy stocks). Also of note: the last place performer in each time frame was AIG.
The results of this little (and I emphasize little) test sample support several concepts I utilize in my daytrading.
1. 5 minute bars provide the biggest risk reward payoff for short term traders. Had I added trailing stops and/or break even stops, results would be even more impressive, but time is limited here in blogdom and I am deferring discussion of stops for several weeks until I get rolling with the new blog format.
2. Certain sectors display technical analysis correlations better than others. When I first started using TradeStation about 10 years ago, INTC was the poster child for technical analysis correlation. Simple systems based on RSI, stochastics, MACDs, linear regressions, Darvas Box- it didn't matter- they all made money with INTC. The situation has changed today with the dominance of program trading, order flow management (masking), backchannel trading (Pipeline) and any number of other factors that make the old workhorse algorithms ineffective. (also see The Game)
3. Finding a trading edge to improve your ROI is not always where you expect to find it. It should could as no surprise that some of the most successful prop trading firms seek out new traders with backgrounds in mathematical modelling, game theory and forensic statistics. Understanding probability analysis is crucial to a trader's success, and whether done on the fly or with the aid of sophisticated real time algorithms, every successful retail trader I know is very adept in pattern recognition and statistics.
Tomorrow, we'll further explore the technical alignment possibilities of various sectors as reflected in ETF performance. Keep in mind that this current series is a work in progress and, just as in trading, it's best not to have too many expectations about what's going to happen next.

Monday, May 12, 2008

Looking for Mr. Goodbar - Part 1

This is the first in a series that will examine some of the parameters of mechanical trading systems. My goal is to aid understanding of market dynamics and, in practical terms, making the greatest return with the least risk exposure.
This series is NOT about providing you a list of recommended trading stocks. . .everybody looks at the market differently and with varying degrees of risk management controls. I'm just going to present a trading perspective that may help in the formulation of your own trading plan.
Today, some thoughts on picking the right trading vehicle. Now this could take years to explain as there literally hundreds of scanning programs, books, DVDs, training courses, etc. to help you achieve those ends.
My spin is a little different.
I prefer to reverse engineer scans to fit my trading parameters using the components I'm most comfortable with: support, resistance and time.
My decision some years ago to focus on the Qs derives from my desire to simplify my trading and the belief that a basket of stocks like the Qs offers superior risk management and technical reliability as opposed to single stock trading. Many traders make a good living trading only the SP500 and Russell minis. . .I like the Qs because of the liquidity, the options, the QID and the penny spreads all round.
But let's set those preconceptions aside and begin from scratch in search of a viable trading model.
To jump start my little research effort, I'll borrow Dr. Brett's watchlist of 40 stocks. The list reflects mostly NYSE stocks with a few of the NAZ big dogs. It's designed to reflect relative strength and weakness of the various market sectors, while simultaneously providing a proxy for wider market dynamics. Dr. Brett makes no arguments for the alignment of the individual stocks and technical analysis, preferring (like myself) to favor the TICK and NYAD to gauge short term momentum.
The scan above is the Schwab SSPro strategy tester, loaded with a simple RSI system on 10 minute bars.
This is a look back of 500 bars = 5000 minutes = 12.8 days
The scan is based on a 100 share trades
Commissions are $6 per side
System only trades long
Of the 9 stocks with positive results, 5 are energy related
Tomorrow, we look at the impact of bar size.

Saturday, May 10, 2008

Qs Weekly Update

Daily and weekly 3 linear regression (30,11,3) study above.
For the week the Qs were down .56 or 1.15%. Not bad considering the fallout in some of the other indices.(DIA -2.5%, IWM -1.5%, SPY - 1.85%)
On the daily chart the Qs are dead on the LR30 mean and downslope on the LR3. Once again traders are faced with the question is whether the Qs are at support or resistance relative to the channel. The mid panel technicals are uniformly negative and from that perspective we can expect more retracement with the lower LR30 channel (46.50) as a likely support target.
The weekly chart is a bit more optimistic as the Qs have not yet breached the upper LR30 channel while the technicals are neutral to modestly negative. Interestingly, a retracement back to the weekly LR30 mean would take the Qs to 46.50.
My bet is that we'll see that number again, but being a daytrader (for now) allows me the leisure to speculate about it without having to pay up if I'm wrong.
The A50 is showing slight weakness, having come off the highs last week, as expected. Still above the 20 DMSA . . a break below that level would be confirmation of more short term momentum down, with the 50 DMSA as intermediate support.
The A200 actually showed relative strength this week. . a bit surprising. Like the A50, the 20 DSMA has provided support since April. A break down through that level in conjunction with a break in the A50 should provide a good swing trading opportunity.
I had intended to post some TICK/NYAD correlation charts today, but have delayed that until next week while I run some backtests over the weekend.

Friday, May 09, 2008

Friday - Something New

Over the next few weeks I'll be changing the format of the blog. To those ends I am posting 1 of the 4 screens that I watch in real time all day. (click to enlarge). I have another screen with identical components with the exception that the far left and far right panels display stocks other than the Qs and the middle pane trading windows reflect those other charts.
The far left and far right charts are formatted the same, with the exception that the Left is tailed by the TICK at the bottom and the Right is tailed by the IWM at the bottom.
The technical indicators are defined for in each of the chart panes.
Left is set to 2 minute bars, right is set to 5 minute bars. . .although I can instantly toggle to other time frames for confirmation purposes, and frequently do.
What the indicators mean is what I've spent 20 years learning.
The middle of the screen is composed of an upper and lower trading window - upper displays level 2 options and lower displays actual Qs bid/ask prices.
The NYAD sits front and center above the lower trading window as I consider it critical to my trading triggers.
The two indicators in the center of the middle pane are really pulse meters for the market.
The left vertical scrolling indicator shows new intraday highs and lows for the NYSE and NAZ, and the daily high or low count. When the scroller is running fast and all green. . .it's time to BUY. When the scroller is running fast and red. . .it's time to SELL.
The bigger scrolling indicator on the right with the horizontal bars showing AAPL to XLF is a user programmable algorithm that Schwab calls the Dynamic Ticker, and basically displays the net asset value of of 11 items that I have chosen to stick in the watch list. The horizontal bars pivot off a zero value in the center and display the amplitude of the buying or selling in each item on the watch list on a TICK basis. When these all turn red and hit 80-100%, you'd better be short.
These 2 indicators are absolutely invaluable for determining short term changes in buying and selling pressure.
I'll make additional screen posts on subsequent Fridays in an effort to enable my gentle readers to understand how I trade and my attitude about what to trade.

Finally, a link to Dr. Brett that's worth a close look. Post #10 by your trading screen. It's just a few words but a truly invaluable insight (what I call a nugget).

Thursday, May 08, 2008

Thursday Qs recap

A narrow range day with moderate volume. The Qs hung to the PP like glue and retraced to the pivot on every rise and fall. The pop in the NYAD at 13:20 got my attention, but the TICK was not supportive. 14:00 saw a fade in the NYAD and TICK and with the parabolics on a SELL in the Qs and the NYAD, I went short the Qs at 48.62. This was late in the day for my typical entry since, as those who have followed me for a while know, I'm not currently holding positions overnight. This was almost a mechanical trade as my sole focus was the 5 minute parabolics. I probably stuck with the trade too long, as the dithering around the PP at 15:00 was a clue of more downside to come. Nevertheless, the parabolics fired a exitshort at 15:00 and I exited at 15:05 at 48.33.
Net time in trade - 65 minutes
Trade gain - .29
Not a great trade, but I took what the market gave me, followed my trading plan and maintained my defensive posture.

Wednesday, May 07, 2008

Deconstructing the EWC - Canada

For longer term investors, calendar spread traders and buy/writers, EWC might be something to examine a bit closer. Longer term support at 28 has proven to be very resilient and the long term (9 month) daily pivot at 31 (not shown) looks sustainable. Currently the EWC looks mildly overbought and this may resolve itself with the current pullback in the basic materials sector which comprises a good proportion of EWC assets (see below).

The weekly chart reflects the unbroken 20/50 MA rise for the past 3 years and the sudden rise in volume that corresponded with the cratering of US equities and the search for investing (not trading) safe havens. Cleary not overbought technicals on the weekly chart With over 50% of the EWC net assets in energy and industrial materials, the dynamics in the commodities and energy need to be monitored. When $5 gets you a gallon of gas or a box of corn flakes (coming soon), EWC net asset value could show some legs. RIMM also accounts for 5% of the assets, currently at a new high on the NDX, and on a roll.. Gold is currently on a cyclical low and with over 7% of net assets in that sector a 15% rise would also significantly enhance the EWC. The Risk Bar above is a nifty little free tool courtesy of the NASDAQ which I have mentioned in previous posts. Interestingly, it gives the EWC a safer reading than XLE, perhaps because of the diversified character of the EWC. . .the reading on XLF is not surprising.
Open interest in EWC options is just plain dismal. The largest strike position is the September 33 calls (2,903) (June chart shown above), which is probably a buy/write situation as they are currently $2.35. With the EWC at 33.75, the Sept return is 4.7% with a break even at 31.40. That's not a bad bet for the summer doldrums months (which some other folks have apparently already figured out). These options are not for active traders as spreads are typically .25 and it's not unusual to have zero daily contract volume in all but a few strikes. Limited participation of the option exchanges also puts a damper on liquidity and the willingness to fill limit orders.
Finally, short interest is showing little negative sentiment, having maintained an envelop of support and resistance since 2004, with the exception of a few squirrely months. The ratio is the number to keep a eye on here as the volume increase over the past 7 months otherwise distorts the absolute numbers.