11/15/2009 8 pm update: NQ

• the NQZ09 offer i placed on wed 11/11 for 1797.50, was filled during tonights session...
• this trade: +18.75 NQ points/contract, 1 contract closed...
• on the above chart, tradestation is not showing the trade history correctly...the chart reports i still have 1 open NQZ09 contract when in fact i am flat the NQZ09...

• currently bidding for 1 contract at 1753.50, just above the rising 200ma on the 15M chart and at 1717.50...
• i am also bidding for 1 contract at 1785.00...if this order is not filled during the rest of the overnight session, then i may change it during monday's regular session...

***remember this is an illustration of what i am trading and my thinking...my trading plan may change without notice...this is not a recommendation for you or anyone else, to buy or sell this or any other security...trade at your own risk**

10/8/2009 11:00 am update: NQ

• sold one NQZ09 at 1716.50 during the overnight session...
• this trade: +10.75 NQ points/contract, 1 contract closed...
• this exit made me flat the NQ...
• big gap up this am...i was looking for a gap fill and was bidding 1712.50 for 1 contract...changed the entry price and was filled at 1713.50...
• offered this contract out at 1722.50...when the bounce back up got stuck under 1720, 1 got nervous that the 1722.50 target would not be reached this am, so i exited at market, 1718.50....
• this trade: +5 NQ points/contract, 1 contract closed...
• today: +7.88 NQ points/contract, 2 contracts closed...
• the above NQZ09 15M chart is not showing the correct historical fills again...it is showing two contracts exited on today's first 15M bar, but only 1 should be displayed there...yesterday's exit at 1702.25 is not being shown...
• this is an ongoing issue with tradestation's platform, and hopefully will rectify itself soon...it is usually associated with some bad/missing historical data in the TCache folder... i removed the TCache folder (to no avail) and am currently using the latest version (8.6 build 2674 beta)...
• still a bit off my game today...2 of the 3 fills today were not at my original targets, primarily because of fear...
• at the highs today, NQ was trading about 2% below its last swing high (1753), while the ES was trading about 1% below its last swing high (1076)...ES still showing some rs compared to the NQ, primarily due to XLF out performance and SMH under performance...i prefer rallies that are led by the NQ, however the financials led this market down and are now leading it back up, so periodic out performance by the ES compared to the NQ is less of an issue during this market cycle in my view...
• i am still bullish longer term on both the ES and the NQ, though would not be surprised and would actually welcome, a whoosh down during earnings season...
• currently flat the NQZ09 and bidding for 1 contract at 1712.50...

Disclosure: no positions in the securities mentioned.

***remember this is an illustration of what i am trading and my thinking...it is not a recommendation for you or anyone else to buy or sell this or any other security...trade at your own risk**

The SPY tended to trade bullishly on the day before the FOMC interest rate announcement from 1/1996 - 11/2001, but has been bearish since then.

Trading the Odds posted an interesting study on the behavior of the ES on the day before the FOMC announces their decision on interest rates:

• from their study (S5), the key takeaway for me was the Profit Factor of 1.70 over the course of 156 trades...this peaked my interest, so i wanted to see if i could build a systemic trading model based on this behavior...
• their study uses the ES datastream and the FOMC interest rate release dates from 1990-present...i had somewhat easy access to the SPY datastream and the FOMC interest rate release dates from 1996-present, so i began there...
• my study identified the date which was 2 days prior to the FOMC interest rate announcement, purchased $100,000 worth of the SPY at the close of this day, and then sold it the following day (the day before the FOMC rate announcement) at the close...i did not use any other filters and i did not use any stops...
• my study came up with 106 trades and a Profit Factor of 1.42...there are less trades primarily because my study contains 6 less years of data compared to Trading the Odds study...the FOMC announces a potential change to interest rates 8x/year, which would amount to 48 additional instances over the 1990-1995 period...adding those 48 trades to my studies 106 trades, would make a theoretical set of 154 trades, which is pretty close to the 156 trades in the Trading the Odds study...this is close enough for me for a preliminary review of the SPY behavior...
• once i completed my study, the thing that jumped out at me the most, is the Equity Curve chart:

• the equity curve peaked at trade #46 (11/05/2001) and has been basing, with a recent downward bias since then...
• this equity curve suggests to me that while the SPY had a bullish bias on the day before a FOMC rate announcement from 1/1996-11/2001, there has been a bearish bias since then...
• even though the Profit Factor from my study is a bullish 1.42, it has been on the decline since 11/2001...
• even if i added the performance data for the 1/1990-12/1995 period, it would not change the equity curve from 11/2001-present...
• i will not be building a systemic trading model based on the behavior of the SPY on the day before the FOMC interest rate announcement...
• the purpose of this post is to add some color to the setup that Trading the Odds posted...it is not meant to discredit the great work they do...on a daily basis, i look forward to reading their research...it has been a great addition to my knowledge base and to my trading...
• the following is the list of dates that are two trading days prior to the FOMC interest rate announcement that i used in this study:
  

'Trading the Odds' Blog: Pre-FOMC Day shows enough of a positive bias, that I might be able to build a trading model based on it.

Trading the Odds Blog:

• Pre-FOMC day with strong upward bias on the SPY...
• i should model this to see if i can build a trading system based on this setup...
• i will continue to follow this blog for additional setups that may lead to viable systemic trading models...
• i am also following Quantifiable Edges blog for additional setups that may lead to viable systemic trading models... the SPY Rising While SPY Volume Declines study (selected from the 8/25/2009 - 9/22/2009 QE posts) is another candidate for me to model...

ATR (Average True Range) Volality Bands for Intraday trading: version 2

• here is a 15M chart of the NQU09.D from around 1130 am this morning...
• using an ATR calculation, i want to try to predict where the probable intraday high and low ranges will be for the NQ...
• the subchart displays the ATR volatility in percent terms...the yellow vertical bar histogram shows how volatile each intraday session is....currently the difference between the true high and the true low for the current daily bar in the NQ is 1.31% of the last trade in the NQ....the red horizontal line is an 8 period moving average of the daily ATR volatility...it is currently at 1.84%
• so far for today, the NQ is less volatile than its 8 day average volatility...
• i use the average ATR volatility to help determine where the ultimate high and the ultimate low for the current daily bar might be...there is no guarantee the NQ will trade to these levels, however if the NQ has a typical day, then attempting to project these high and low range areas may have some relevance...
• however, i believe it would not be useful to attempt to predict the exact high and the exact low, so i have opted to try to predict a high range and a low range...
• the calculations for the high range are relatively straight forward: in order to project the high, i use the known true low and then add the current average ATR...this gives me a firm projected high...however, i don't want a firm projected high, but rather a projected high range, so i tweak the firm projected high with a user defined 'band variance factor'...the 'band variance factor' replaces the projected firm high with two new highs, one above the firm high and the other below the firm high...the size of the expansion is dictated by the size of the 'band variance factor'...these two highs are then plotted on the chart of the NQ...
• the dark red band represents the upper portion of the projected high range...the red band represents the lower portion of the projected high range...
• the calculations for the low range are relatively straight forward: in order to project the low, i use the known true high and then subtract the current average ATR...this gives me a firm projected low...however, i don't want a firm projected low, but rather a projected low range, so i tweak the firm projected low with a user defined 'band variance factor'...the 'band variance factor' replaces the projected firm low with two new lows, one above the firm low and the other below the firm low...like mentioned above, the size of the expansion is dictated by the size of the 'band variance factor'...these two lows are then plotted on the upper chart of the NQ...
• the dark green band represents the lower portion of the projected low range...the green band represents the higher portion of the projected low range...
• i specifically used the dark red, red, green and dark green colors...if my bias was to be long, i would look to buy some near the green line, and buy some more near the dark red line....then i would look to sell some near the red line and sell some more near the dark red line...if my bias was short, then i would look to short some near the red line, and short some more near the dark red line...then i would look to cover some near the green line and then cover some more near the dark green line....
• if long and the vehicle traded below the dark green line, then perhaps a stop loss would be in order...if short and the vehicle traded above the dark red line, then perhaps a stop loss would be in order...
• however, this would not be my only analysis tool for determining optimal entries, exits and stops...i would also be using multiple time frame analysis, moving averages and pivots...these projected intraday highs and lows would be another analysis tool to help me optimize my entries, exits and stops...
• the 'band variance factor' is user defined and easily changed...
• i am excited to have added this tool to my trading toolkit...
• currently i am using this indicator for the NQ but it can be used for any symbol...
• if you are interested in taking this indicator out for test run and you use tradestation, then let me know and i will send it over to you free of charge...
• if anyone has any ideas on how to improve this indicator, especially on how to optimize the 'band variance factor' such that it creates true high ranges and true low ranges that get pierced a significantly high amount of the time that then lead to reversals, then please let me know...

Quantifiable Edges: Studies from Rob Hanna that I should do further modeling on

The following links are to studies that Rob Hanna posted on his Quantifiable Edges blog from 5/22/2008-8/24/2009 that I should do some further modeling on:

• CBOE Equity Put/Call finishes more than 25% below its 200 day moving average
• From a High to a Low in 1 Day
• Large Gaps up from a 1 Month High
• NYSE New Highs Contract While SPX Makes a 50-day High
• Slowing Rate of Decline Potentially Short term bullish
• Volume Low and Price High Often Leads to a pullback
• SPY closes in the bottom 10% of its daily range but still above yesterday's close
• S&P is oversold going into a Fed Day

I will continue to review Rob's blog for additional market behavior research.

8/24/2009 post close update

NQU09 discretionary swing trade:

• very narrow day for the NQU09 today...
• still long 1 contract at 1635.50...
• current orders to sell limit 1656.50 and to buy a second contract at limit 1617.50...
• working on building an indicator that plots an average true range (ATR) projection for the High range and the Low range of the current trading day...in addition to the obvious pivot points on a 15 minute chart, i am incorporating these High and Low range projections into my thinking when determining buy and sell levels for the NQU09...

Systemic swing model rebuild:

• the $IDX and the $RUT.X data series on TradeStation incorrectly reports the Daily Bar's opening price...it uses the previous Daily bar's Closing price, and thus there are no Opening gaps...this inaccurate data was effecting the entry filters and the stops for models that use these data series...a different data series is now being used as a proxy for the S&P 400 (MDY) and the Russell 2000 (IWM)...
• the 'BO IDX' model was rebuilt on the MDY data series...it is being discontinued because it is not robust...this discontinuation saved me some money today as the old 'BO IDX' model generated an entry signal for this past friday's close which would have posted a 0.50% stop loss today...i am pretty psyched to have finally had a model rebuild that worked in my favor...
• 'BTBT IDX' and 'BTBT RUT' have been successfully rebuilt using the new data series'...
• there is a lot more testing/building to be done, but progress is being made...

VCM daytrading:

• learning a trading style that gets my position bigger as it goes in my favor, as opposed to getting bigger when it goes out of my favor...very interested in continuing to learn more about this style...
  

Disclosure: Long NQU09.

***remember this is an illustration of what i am trading and my thinking...my trading plan may change without notice...this is not a recommendation for you or anyone else, to buy or sell this or any other security...***

About my Systemic Trading Models

I have spent the past few years attempting to identify and model robust U.S. Stock Market Index behavior that is not just descriptive of the past, but also predictive of the future.

Combining this behavioral analysis with prudent risk management, has enabled me to construct a portfolio of systemic based trading models.

On the right side of the blog, below the 'About Me' section, is a list of my Systemic Trading Models.

'Trading Systems that Work'

'Trading Systems that Work' is the title of a great book by Thomas Stridsman. He outlines how to 'Build and Evaluate Effective Trading Systems'. I have used his concepts extensively when building, testing and optimizing my 35+ Systemic Trading Models.

TradeStation, the platform I use to model and autotrade my Systems, offers online tutorials (for subscribers) on the following:

• Introduction to Strategy Trading
• Strategy Testing and Optimization
• Strategy Automation
• Introduction to EasyLanguage

Learning how to build robust trading models and then learning how to technically execute them takes a lot of time. And this is just the cognitive part. Learning how to emotionally trade them takes time as well. It doesn't matter if we are discretionary traders or system traders, pressing the buy or the sell button when we are supposed to is a skill set in and of itself.
For anyone interested in using TradeStation, Wealth-Lab or any other platform to build Systemic Trading Models, or even to just build an Indicator, please understand that these platforms are like rocket ships in their capabilities. Let us learn how to effectively use their functions before we take off. We don't need to crash and burn to learn a lesson!
Do your research and testing before putting your hard earned capital on the line. In my view, 'Trading Systems that Work' and the TradeStation online Tutorials are a good place to start.
http://www.amazon.com/s/ref=nb_ss_gw?url=search-alias%3Daps&field-keywords=stridsman
http://www.tradestation.com/default_2.shtm

'NYSE TICK Normalized' Indicator for TradeStation

I put together a TradeStation Indicator for a 'NYSE Normalized TICK', such that the High TICK and the Low TICK will be shown as a % of the Total Number Securities Traded on the NYSE that day. This will work for Daily Bars as well as Intraday Bars.

However, the results didn't appear that useful for intraday scalping purposes (at least how I understand how a TICK scalper would use them). My sense is there isn't a meaningful difference between the total number of securities traded day to day, so that normalizing the TICK extremes doesn't yield an edge over using the raw TICK.

If anyone is interested in this Indicator, please let me know and I will send it over. I don't think there is anything useful here though.

'ATR (Average True Range) Volality Bands' Indicator is ready!

I have completed the material for the 'ATR Volatility Bands' Indicator and have sent it out to fellow traders and TradeStation Programmers. Also included is my 'ATR Volatility %' Indicator.

If you would like a copy then let me know.

I hope folks find it useful and can perhaps improve upon its functionality.