Please share with me your opinion

[songcon]

AAPL weakness - for a short term or a bad sign

AAPL 5-day chart shows some weakness:

 

[songcon]

Dec 7, 2009 - A bullish divergence of RSI from price for EEE

Price is in downtrend, RSI is up. Price is about to change the direction to confirm with RSI, MACD and Stochastic

To move up in price now, a double-bottom will be formed also

 

[MTRIG]

There were some good trades on this post. MACD Cross Over, RSI breaking 50... All this looks good but when you back test any of this over the long haul it shows that automated (non discretionary) trades on these factors does not produce long term profits.

Here are some back-tested results on QQQQ using AMIBroker. Forget about those chart scanner web sites just buy a copy of AMIBroker it is only $199 and data can be downloaded for free and you can write and test any type of trading system.

Here is a QQQQ chart with BUY/SELLs on MACD Cross over

Here is the simple forumula in AMIBroker

Here are the back tested results with MAX 8% stop loss since QQQQ inception.

 

[songcon]

Thanks for the information. I don't mind spending $200 for a piece of software but the free BigCharts chart also gives me exactly what AIMBroker generated.

I use other indicators to study their relationship with the price movements. From my learning and my experience I use MACD for confirmation of a trend, MACD is quite often late in generating sell signals, especially for hot stocks.

Life is fairly easy in using very basic chart reading skills in dealing with QQQQ. It's a big lesson for me and it's something that I should follow up more.

Again, thanks for sharing your ideas and experience

 

[MTRIG]

Thanks for the information. I don't mind spending $200 for a piece of software but the free BigCharts chart also gives me exactly what AIMBroker generated.

I use other indicators to study their relationship with the price movements. From my learning and my experience I use MACD for confirmation of a trend, MACD is quite often late in generating sell signals, especially for hot stocks.

Life is fairly easy in using very basic chart reading skills in dealing with QQQQ. It's a big lesson for me and it's something that I should follow up more.

Again, thanks for sharing your ideas and experience

With BigCharts you can plot any indicator, I agree, but you cannot perform a back-test and that is where AMIBroker comes into play.

Using MACD to confirm a trend is a good idea, but code that into a system and then lets see how it worked over historical data. Did it really perform well? Only a test will show the facts. What you describe is "discretionary" trading such as what Dr. Elder of Trading Room or Trading for a Living uses.

No one can back-test any of Dr. Elder's strategies and that is a way out for him to make any claims on how his "Triple Screen" performed because you cannot quantify it since it is up to each trader because it is a "discretionary" system. If we email Dr. Elder and tell him the Triple Screen does not work, he will say "it is your fault, it works for me."

 

[songcon]

MTRIG wrote:

With BigCharts you can plot any indicator, I agree, but you cannot perform a back-test and that is where AMIBroker comes into play.

Using MACD to confirm a trend is a good idea, but code that into a system and then lets see how it worked over historical data. Did it really perform well? Only a test will show the facts.

I used the chart of DPTR from July 01 to Oct 02, 2009 to give an example that MACD is late in giving sell signal for hot stocks.

Please run a back test for the same period to see how AMIBroker program does the job.

The story of DPTR is that I was telling my friends that I did not feel comfortable to see the light trading volume
at that point, there might be a reversion around the corner, and sure enough Stochastic gave the signal the following day.

Please see how you can back test DPTR for this period by using AMIBroker.

Thanks

 

[MTRIG]

MTRIG wrote:

With BigCharts you can plot any indicator, I agree, but you cannot perform a back-test and that is where AMIBroker comes into play.

Using MACD to confirm a trend is a good idea, but code that into a system and then lets see how it worked over historical data. Did it really perform well? Only a test will show the facts.

I used the chart of DPTR from July 01 to Oct 02, 2009 to give an example that MACD is late in giving sell signal for hot stocks.

Please run a back test for the same period to see how AMIBroker program does the job.

The story of DPTR is that I was telling my friends that I did not feel comfortable to see the light trading volume
at that point, there might be a reversion around the corner, and sure enough Stochastic gave the signal the following day.

Please see how you can back test DPTR for this period by using AMIBroker.

songcon, I will run the back-test. It is not so much as how AMIBroker performed, but a mechanical system based on the criteria. Regarding your example what do you want to test? MACD Crossover for the life of stock? Stochastic BUY signal ....then if MACD fires 3 days after that GO LONG (reverse the rules for GO SHORT)? Let me know and I would be happy to run it.

Also, looking at the chart on your post there were many Stochastic and MACD crossovers back to back (LONG/SHORT) when the stock was in a trading range. This is typical in a trading range, what do you do to filter out those signals visually?

 

[songcon]

songcon, I will run the back-test. It is not so much as how AMIBroker performed, but a mechanical system based on the criteria. Regarding your example what do you want to test? MACD Crossover for the life of stock? Stochastic BUY signal ....then if MACD fires 3 days after that GO LONG (reverse the rules for GO SHORT)? Let me know and I would be happy to run it.

Also, looking at the chart on your post there were many Stochastic and MACD crossovers back to back (LONG/SHORT) when the stock was in a trading range. This is typical in a trading range, what do you do to filter out those signals visually?

This is something that I've heard a lot recently but afraid to ask. Your new explanation how to set up the barometer (?) of the program.

I just want to know how do you set up the condition: This is my low tech criteria for selecting a buy: If Stochastic crosses over the slow line on day 1, and MACD cross-over = Day 1, day1+1, day1+2 or day1+3: Buy!

If Stochastic line reversed = Sell.This is something I have to use my eyes to make judgment calls, but I always watch Stochastic closely if the stock is a hot one that jumped up nicely. For example, on Tuesday last week, I'd sold AMFI at 1.47 & 1.59 by watching the 15 min candles. The trades were good since the high of the day was $1.75 but it closed at $1.50.

Thanks in advance, T2W is really a good place to learn.

PS: Buy = Stochastic signal with MACD confirmation
Sell: at Schochastic crossing of 5% trailing stop loss if profit target = 50% since entry point

 

[MTRIG]

Based on your reply the logic will be... (based on your charts it appears you are using Slow vs. Fast Stochastics)

BUY = Stoc Cross (%K Crosses the %D) AND (MACD Cross same day, or +1, +2, +3)
SELL= Would be the revrese of the BUY logic.

Additional Rules
Stop Loss: 5%
Trailing Stop: 5%
Profit Target: 50% (wow that would be nice)

I will post the results with stops and without. Without stops basically says you are holding the position until the reverse signal fires.

Also some other rules to consider.... add to the BUY logic (AND Stoc is Oversold)... for example.

 

[MTRIG]

Songcon,

Overall it performed "not too bad" some improvements may be to add triggers on Moving Averages, or only BUY when oversold and only SELL when overbought.

See the chart, stats, and formula below.

 

[songcon]

xSell = StoSell AND MACDSell

For hot stocks, wait for MACD sell signal is too late. Only Stochastic is enough but the condition is "hot" therefore, a condition such as profit >= 50% and Stochastic crossover = Sell

I'll spend time to learn your stuff, it's very interesting

 

[MTRIG]

I revised the last post. The calc was looking forward for signals but should have looked back.

 

[songcon]

I went back to take a second look and found out that the buy and sell signals generated
by MACD and Stochastic are not accurate for QQQQ. The buy signal from MACD in Nov is late, the sell signal from Stochastic in early Nov was too early, your sell signal in Nov was too late.

I had roughly 100% capital gain for 2009 which includes 200% in one trade which was a totally stupid thing. 2009 was an easy year