Tuesday, January 26, 2010

J200 60min chart


Quick look at Top40 (J200) 60 minutes chart:

Top40 currently is very close to long term support around 24100-24200 (horizontal black line). This support going back all the way to August 2008 and in recent history was tested 5 times as resistance and then as support. 

Conclusion is that Top40 can bounce from here, especially if we take into consideration 2000 points drop from 11. January this year.

Trade with trend!

Top40 (J200) weekly chart



Last week produced biggest weekly drop since early September 2009 and that itself is a valid reason to evaluate charts.

To make it clear, system is still long on weekly chart. Entry was in July 2009 around 21000 and trade is in profit so far with approximately 3500 points accumulated in 6 months.

Market turned its downslide in March 2009 and never looked back. After some 10000 points rally market progressed all the way to 25936, right up to 61.8% retracement level - if drawn from top in June 2008 to low of March 2009. Previous major high (August 2008) is very close to turning point. There is rising wedge on first chart that broke last week, with target around 20000, if it plays out as it should. RSI and Stochastic did not show some massive negative divergences on the way up, but trend line on RSI is broken for now.

Second chart shows bigger picture. Top 40 weekly chart (from mid 2002) is shown. There is long trend line (thick green line) from 2003 that was tested 4 times so far. In recent times this trend line played support in December 2008 and March 2009, and that stopped Top40 from drooping further. What else seems obvious on that chart - 26000 is not only double resistance based on first chart as there is even 23.6% retracement level for the bull market which started in April 2003 and ended in Jun 2008. That makes 26000 triple resistance and Top40 could not breach it first time. Finally, there is clear support just below 23000 (black horizontal line) that was tested 6 or 7 times so far.

Conclusion is that Top40 dropped a bit last week, but nothing yet suggests we are going to see similar drop to one from mid 2008. Healthy correction is very welcomed at this point of time. Will follow weekly perspective and report if something changes.

Trade with trend!

Thursday, January 21, 2010

Top40 end of day chart


Today Top40 closed 1.61% down at 24823 and that was biggest single day drop from 26. November 2009. This drop can easily bring sleepless night to bulls but as it stands tonight up trend started the 3rd March 2009 is still intact.

Today close was exactly on trend line that is at the same time bottom of the rising wedge formed from the 3rd March. If we have red day tomorrow most likely wedge will break and target for that move is down to 21000 or some 3800 points or 15.3%. Off course nobody can expect to be straight line drop.

There is significant support just below 22000 (black horizontal line) but still far.

Indicators are not painting nice picture. Macd is in triangle formation but showing quite bad negative divergences. There are four recent tops in August, October, November 2009 and January 2010 not confirmed with MACD. Stochastic is very low around 20 but as we know stochastic not working well in trending market and can be overbought or oversold for quite some time.

At least we can look forward to increased volatility that will help us with our trading system

Trade with trend!

Average True Range and its effect to various simple systems



This table presents simple concept and it is related to 5 different trend following systems of which one is using 7/21 EMA combo on Top40 (J200) 15min chart; three are using different EMA or SMA penetrations on Top40 (J200) 5min chart and the fifth one is using one of the variations of HH/HL principle on ALSI 30min chart.

Basic idea of this post is to show how, despite having clear daily follow-through to the upside (please consult attached EOD Top40 (J200) chart from 01.July 2009. till 15.January 2010.), various trend-following systems are suffering during periods of low volatility, which was, in this case, measured by 5-day and 10-day Average True Range expressed in percentage terms. Note how volatility droped from October 2008 when was highest from 1998.

In order to clarify attached table I will explain what happened to system named ‘15min 7/21 C’:
  • From 01. December 2008 till 14. January 2010 system made (after brokerage) total of 17088 points ouf of 221 trades or 77.32 points per trade.

  • When 5-day ATR% was bellwo average ATR% value (after each close), sysem made 1720 points out of 75 trades, or 22.93 points per trade. This is leaving us with 15368 poitns that wre made out of remainiing 146 trades, or 105.26 points per trade
Note: ATR% value was calculated on Top40 (J200) EOD chart after each close.


It is clear that we have identified underperforming periods for five systems by using these simple ATR% metrics. Practical application, in terms of switching systems On/Off or position sizing is completely different matter and warrants in-depth analysis, which, for now, exceeds boundaries of this blog.

Trade with trend!

For more reading on ATR please go to:
http://stockcharts.com/help/doku.php?id=chart_school:technical_indicators:average_true_range_a

Tuesday, January 19, 2010

Top40 (J200) 15min chart

J200 15 minutes chart is showing us few possibilities.

1. Falling wedge (black top line (1) & blue bottom line (2))
2. Falling channel (black top line (1) & black bottom line (3))

In both cases downtrend is developing. Only trend line break will negate current down trend.

Please notice there is major trend line around 25250 (marked as 4), some 100 points bellow current level. Trend line is going back to early October 2009 and it can hold market once more.

Trade with trend!

EDIT: Quick look at same chart one day later. Trend line (marked as 1) still holding after being tested 5 times. Major trend line (marked as 4) is still holding as well. Space between them is becoming smaller and we will have to break soon one way or the other.




Two periods in time – follow up

Quick follow up with some different EMA combinations - One was used at one point in time and here I am going to re-visit 10/50 crossover.

Period I (01. July 2008. to 01. April 2009.): 15460pts, 134 trades, Avg.W/Avg.L = 2.99;

Period II (01. April 2009. to 01. January 2010.): 2498pts, 139 trades, Avg.W/Avg.L = 2.39.

If we take brokerage and slippage into the account for first period, we could say that 10/50 was better option to trade than 7/21, as point difference is 735 in favor of 7/21 with 82 trades more, which means that with 2 points brokerage and 7 points slippage per trade that advantage is gone.

Real difference comes in period from 01. April 2009. to 01.January 2010. when faster moving averages made 7250 points. Simply, if we take brokerage and slippage into consideration, system would probably lose money during this period.

It is obvious that shorter (faster) moving averages were performing better in changed conditions, like we had during last 9 months.

Trade with trend!

Monday, January 18, 2010

Two periods in time

This post looks at difference for two different 9-month periods.

Back test was done for exponential moving average (EMA) crossover strategy on J200 (Top40) 15 min chart.

First screen capture shows period from 01. July 2008. to 01. April 2009. System was optimized for values between 6 and 8 for shorter moving average and for values between 20 and 22 for longer moving average. During those 9 months the best system made 17317 points from 204 trades with ratio of average winner and average loser being 2.91. Value for this particular EMA combo was 8 and 21 periods.

Second screen capture shows period from 01. April 2009. to 01. January 2010. System was optimized for identical values as in the above mentioned case. During these 9 months best system made 10187 points from 233 trades with ratio between average winner and average loser being 2.53. Value for best EMA combo was 6 and 22.

If we take same look-back with, let’s say, EMAs 7 and 21 we see the following picture:

Period I (01. July 2008. to 01. April 2009.): 16195pts, 216 trades, Avg.W/Avg.L = 2.77;

Period II (01. April 2009. to 01. January 2010.): 9748pts, 226 trades, Avg.W/Avg.L = 2.66.

Simply, all basic metrics of this system dropped since 1. April 2009. Number of points has decreased by amazing 39.8%. Less points in more trades with worse average winner/average loser ratio is simply not good for trading and this is why so many traders are struggling and looking for new systems.

Will this systems start to perform as good as before again? Are all these systems broken?

If you asked me on 1. January 2010. if I would like to have 1000 points per month guaranteed I would gladly accept that offer - any time, any market conditions and with no regret if chosen system starts to produce 2000 or more points per month.

As we can see, systems made around 1000 per month from 1. April 2009. till end of the year, so if you ask me – systems are still working and there is no need for change. If you want to see what 1000 points per month does in one year, please revisit one of previous posts:

“Realistic goals” --- http://alsi-trader.blogspot.com/2009/07/realistic-goals.html

All what we need to change is our own discipline in following systems, because missing good trades now can cause us not to make money. Please concentrate on execution and discipline and money will come.

“System development and testing never stops.”

Trade with trend!

Friday, January 15, 2010

Two charts




Quick look at two charts:

First chart is Top40 (J200) 15min and is showing head and shoulder with neck line. Same picture can be seen on ALSI 30min chart.

There are some 650-700 points in this formation. Of course, one must wait for break of the neck line to enter trade.

Trade with trend!

EDIT: Neckline broken. Target for ALSI 24800 but don't forget back test


Wednesday, January 13, 2010

ALSI daily range – PART II




In my post from 26. October 2009 I’ve posted chart with ALSI daily range. Today we will take a look at what the changes are and how daily range compares to gap size (on 5min chart) over last 14 months.

As one can see from first chart, daily range dropped even further and now is seating bellow 500 points, but this can be due to December effect with low volatility. It is interesting to see that from around Jun 2009 daily range is bellow mean value and it was only once above 1 standard deviation. Most of the time daily range spent bellow mean, very close to -1 standard deviation.

Our systems are not working as they should because we have 2-3 days in one direction with 450 points move, instead of 2-3 days in one direction with 800 points move in a day. We need about 300-400 points for system to stop us out and give new signal, so we use already 600-800 points on both sides and all in all we are left with 400 points in the middle of the trend.

On the second chart gap size is plotted in both directions. So, gaps down are plotted bellow 0 line and gaps up are plotted above 0 line. Mean value for gaps is around 0, with relatively symmetrical distribution. Gap size dropped in line with daily range even though there were two or three massive gaps on downside.

Also, I find this interesting as there were few gaps in the range of 700-900 points late in year 2008.

Trade with trend!

Tuesday, January 12, 2010

Bit more on discipline

Why discipline in following system is so important for system traders?

Often we read that discipline is the most important characteristic of system trader.

Why is so?

I will give specific example from the system that I follow: It is based on 30 min ALSI chart following Dow Theory. If we go through system’s performance log we will see that in period from 13. January - 03. February 2009 system made 3831 points and from 18. November to 4. December it made 2382 points. So, in about 30 trading days out of 240 days in a year system made 6213 points or in 12% of all trading days system made about 32% of yearly profit. Let’s imagine now that one was not in the system during that period of time. Simply, your profit for the year would be much smaller and on top of that all other metrics for the system would be worse, including average winner/average loser, % of winners, T and F formula, etc.

What I learned from more than 6 years of trading is that points in trend following systems are coming in bursts. So, system will make good profits then it will return portion of profits or will tread water for some time. This is where trading becomes difficult and that is when most of the traders fail to follow system.

Frequently occurring pattern is that trader will not trade good patch and after missing good results system trader will start trading bad patch, then after few losing trades he will stop with trading again - right when good patch for the system happens again.

As we speak, our system is going through one bad patch. Since 4. December it lost 1215 points and if one looks from 15. December it lost 249 points. For some it might look like 1215 points is a lot, but that is one good trade.

I believe that discipline is the most important characteristic for system traders on any time frame – in case when trader has system with positive expectancy and sound money management. If you don’t have patience to follow the system through good and bad patches, then your first task is to work on it and to improve discipline in trading one must improve discipline in his life outside of trading. So, next time when your clock rings get up immediately, don’t skip planed gym session, don’t skip STOP sign on the road, etc. Simply, try to improve your discipline in life in order to improve your discipline in trading.

Another problem derived from lack of discipline is that one can close winning positions early, but will leave that one for the next time.

Trade with trend!