PDA

View Full Version : TA help required, did I make mistake or just unlucky?



wbosher
07-07-2009, 08:29 AM
I put a small amount of money into SDL on the ASX looking for a breakout. I thought I saw a pennant formation and got in waiting for a breakout, hopefully on the upside. Unfortunately it went down :( , but what I want to know is if what I did was valid, and I just got unlucky, or did I misread this completely.

Any TA help appreciated. If I did misread this I don't want to make the same mistake again. :o

Cheers in advance.

STRAT
07-07-2009, 10:08 AM
Hi Wbosher
What day did you buy?

wbosher
07-07-2009, 10:22 AM
June 30th at 16 cents. I wanted to wait until it dropped back to 15c but I was worried that it would keep going up, and I would miss out. It went up to 17c that day but then fell again.

I've added a new attachment which shows what I would have seen when I bought in.

STRAT
07-07-2009, 10:31 AM
I forgot to ask if you were planning a short play or longer.
I would approach each slightly differently in terms of how much historical info I take into account but bottom line I reckon is it aint a breakout till its a breakout. You bought before it broke which depending on other indicators is much like the flip of a coin.
You also bought on the 4th down day in a row. What made you think it would eventually break to the upside?

also at that time.
RSI flat
OBV flat
Down trend from 11th June increasing in speed
Overall market conditions looking a bit grim

wbosher
07-07-2009, 10:43 AM
I read in several places that these patterns usually breakout on the side of the current trend, which until then was up (since April). I was planning on a fairly short/medium time frame, maybe a few weeks.

My thinking was that it may go as high as around 19 or 20c and that there was pretty good support at 15c.

Phaedrus
07-07-2009, 10:46 AM
Let's start by looking at what you did right. You probably selected a suitable stock in that SDL was in a "medium-term" uptrend with a rising OBV.

Where you went wrong :-
(1) The pattern identification. Pennants have sloping lines. This formation had a flat base, so it was in fact a descending triangle. This distinction might seem a little pedantic, but these 2 patterns have very different statistics. Pennants after an uptrend usually break to the upside, but Descending triangles usually break to the downside, so the odds were stacked against you right from the start.

Our perspectives are going to diverge markedly, WB, because I am going to use a line chart vs your bar chart.

(2) The medium-term uptrend was running out of steam. See how momentum was falling? More importantly, note the divergences shown here as the Momentum and RSI made lower highs as the price hit higher highs. These are Bearish divergences.

(3) The immediate situation. SDL was in a short-term downtrend when you bought. It still is. No-one can say how far this will run, but the idea is not to buy into a downtrend, but rather wait for it to reverse before entering.

(4) Medium-term, SDL's uptrend had a nice tidy well confirmed trendline in place. This was broken on 25/6/09 giving a SELL signal. You bought (at the green arrow) just after this.

So, when you bought,
SDL was in a short-term downtrend.
It had broken below its confirmed medium-term trendline.
The RSI was falling.
Momentum was negative.
There had been Bearish divergences.
You bought into a formation that would most likely break down.

http://h1.ripway.com/78963/SDL77.gif

wbosher
07-07-2009, 11:04 AM
Thanks for that P, it looks like I made more than a few mistakes. I have a couple of question though.


(1) The pattern identification. Pennants have sloping lines. This formation had a flat base, so it was in fact a descending triangle. This distinction might seem a little pedantic, but these 2 patterns have very different statistics. Pennants after an uptrend usually break to the upside, but Descending triangles usually break to the downside, so the odds were stacked against you right from the start.

I was under the impression that triangles differed from pennants in that triangles usually formed over a much longer time frame, months rather than days/weeks, where pennants were usually only days to a few weeks.


(2) The medium-term uptrend was running out of steam. See how momentum was falling? More importantly, note the divergences shown here as the Momentum and RSI made lower highs as the price hit higher highs. These are Bearish divergences.

I've never used the momentum indicator, I'll check it out. Trying to keep the amount of indicators I use down to a minimum at this stage, so I don't confuse the hell out of myself. ;)


(4) Medium-term, SDL's uptrend had a nice tidy well confirmed trendline in place. This was broken on 25/6/09 giving a SELL signal. You bought (at the green arrow) just after this.

As you said, we used different chart styles, you used the line and I used the bar. Both gave different results! Looking at the bar chart, it looked to me like the medium term trend was still intact, where as the line chart clearly shows a break below the trend line. This has me a little confused as to what type of chart is more effective, and under what circumstances to use what type of chart.. :confused:

Oh, so much to learn...

STRAT
07-07-2009, 11:10 AM
Hi Phaedrus
Just out of interest. My chart shows a bounce off the trend line on the 25/06 leaving the trend in tact for another day or so. I realize this one thing is not enough to change the analisis but was wondering if our data is different or the plot is different
IC shows
High 16.0
Low 15.0
Close 15.0

STRAT
07-07-2009, 11:20 AM
Thanks for that P,

As you said, we used different chart styles, you used the line and I used the bar. Both gave different results! Looking at the bar chart, it looked to me like the medium term trend was still intact, where as the line chart clearly shows a break below the trend line. This has me a little confused as to what type of chart is more effective. :confused:
.Those results had less to do with the chart type and more to do with the fact that Phaedrus plotted his trend line on closing prices and you plotted yours on the lows of the day

wbosher
07-07-2009, 11:29 AM
That's what has me confused. Almost everything I have read online and in books says to plot an upward trend using the lows of the day and a down trend using the highs. The charts using closing prices seemed to signal a break below the trend line far earlier than the the bar chart. I'm starting to wonder if I should start to use closing prices to plot trend lines rather than the OHLC.

Obviously other indicators screamed stay out but my toolbox at present is pretty limited as I explained earlier.

This trade was based more on what I thought was a good chart pattern than anything else, and I though the support at 15c may hold.

Back to the drawing board. :o

STRAT
07-07-2009, 11:48 AM
That's what has me confused. Almost everything I have read online and in books says to plot an upward trend using the lows of the day and a down trend using the highs. The charts using closing prices seemed to signal a break below the trend line far earlier than the the bar chart. I'm starting to wonder if I should start to use closing prices to plot trend lines rather than the OHLC.

Obviously other indicators screamed stay out but my toolbox at present is pretty limited as I explained earlier.

This trade was based more on what I thought was a good chart pattern than anything else, and I though the support at 15c may hold.

Back to the drawing board. :o Perhaps Phaedrus could comment on the merit of High/Low vs. Close for Trend Lines. I use close but would have wrongly read support into the number of days lows that touched on the horizontal line of your descending triangle but either way I would have waited till it a breakout occured before buying.

STRAT
07-07-2009, 11:54 AM
Hey Dragonz.
Are you able to make the Auckland Share Trader pissup on Saturday?

wbosher
07-07-2009, 11:56 AM
Perhaps Phaedrus could comment on the merrit of High/Low vs Close for Trend Lines but either way I would have waited till it a breakout occurs before buying.

Believe me...next time I will! So to sum up:

1. Don't trade decending triangles
2. Check for Bearish divergences in the preceeding trend
3. Use different types of charts to confirm trend.
4. Wait until breakout confirmed before depositing my money into someone elses pocket!

Looks pretty simple, eh. ;)

dragonz
07-07-2009, 12:06 PM
That's what has me confused. Almost everything I have read online and in books says to plot an upward trend using the lows of the day and a down trend using the highs. The charts using closing prices seemed to signal a break below the trend line far earlier than the the bar chart. I'm starting to wonder if I should start to use closing prices to plot trend lines rather than the OHLC.

Obviously other indicators screamed stay out but my toolbox at present is pretty limited as I explained earlier.

This trade was based more on what I thought was a good chart pattern than anything else, and I though the support at 15c may hold.

Back to the drawing board. :o

This came up on my watchlist as a possiple swing trade. But my overall indicators told me to stay out of the Market at the moment.

IMO TA works best in a Trending Market. Your win/lose ratio sinificantly decreases as the market flattens out. I use 10 SMA / 30 EMA crossover as a visual tool to indentify the strength of the trend. The 10sma crossed below the 30 ema last week (All Ords)and was a red flag for my style of trading.

I have set up a couple of accounts that I can short from if a strong downtrend developes. If not then I'll stay mostly in cash until I see a positive 10/30 crossover.

So in short I say look at the overall market 1st, then individual shares second when using TA.

Oh I might add that I did break my own rules and took smallish positions in 4stocks, only 1 of which is in positive territory at the moment. This was despite having the 10/30 overall market rule pinned next to my computer. :eek:

STRAT
07-07-2009, 12:07 PM
Believe me...next time I will! So to sum up:

1. Don't trade decending triangles
2. Check for Bearish divergences in the preceeding trend
3. Use different types of charts to confirm trend.
4. Wait until breakout confirmed before depositing my money into someone elses pocket!

Looks pretty simple, eh. ;)Yeah mate. Simple as a slot machine at the Casino :D
Just wait till you get it all right and still loose some money

dragonz
07-07-2009, 12:13 PM
Hey Dragonz.
Are you able to make the Auckland Share Trader pissup on Saturday?

I have a mate coming to stay with me over the weekend. I might see if I can meet him in Auckland 1st for a few drinkies. :D

wbosher
07-07-2009, 12:26 PM
Yeah mate. Simple as a slot machine at the Casino :D
Just wait till you get it all right and still loose some money

Can't wait...;)

Would you buy the next day after a close above resistance, or as soon as it is breached on the same day?

Just a wild guess, Fender man? :D

Phaedrus
07-07-2009, 01:04 PM
"I was under the impression that triangles differed from pennants in that triangles usually formed over a much longer time frame, months rather than days/weeks, where pennants were usually only days to a few weeks".
According to Bulkowski, flags, pennants and descending triangles are all short-term (up to 3 months) consolidation patterns. The salient point here is whether the bottom line of the formation is flat or not.

"I've never used the the momentum indicator, I'll check it out. Trying to keep the amount of indicators I use down to a minimum at this stage" It is essential that you have some sort of momentum indicator in your toolbox. There are plenty of others though eg RSI, RMI, Sto etc etc. You don't need a lot of indicators. I have a basic set of just 8 and rarely, if ever, use any others.

"We used different chart styles, you used the line and I used the bar. Both gave different results!" Not as far as your buy decision was concerned. You were looking for an entry, right? SDL was in a short-term downtrend and you were (should have been!) looking for a break of this trend. Your bar trendline and my Close trendline are both still unbroken.

"This has me a little confused as to what type of chart is more effective, and under what circumstances to use what type of chart."
Of the High, Low, Open and Close prices, which is the most important? The Close, no contest. So......which would be best to use for trendlines etc where you can only pick up one data point per day? Answer - a line chart of Closing prices. The plots are much cleaner and simpler as well. My advice would be to use line charts, unless you are interested in intraday data - in which case candlesticks are the way to go in my opinion.

"The charts using closing prices seemed to signal a break below the trend line far earlier than the the bar chart. I'm starting to wonder if I should start to use closing prices to plot trend lines rather than the OHLC."
Stop wondering! Would you really want to act on a buy or sell signal triggered by an intraday spike? Either way you lose.

WB, the bottom line here is actually the top line! Barchart, line chart, triangle, pennant, linear scale, log scale, divergences, oscillators etc notwithstanding, however the short-term downtrend line was drawn, it was not broken. It has still not been broken. The downtrend is still intact.


"My chart shows a bounce off the trend line on the 25/06 leaving the trend in tact for another day or so. I realize this one thing is not enough to change the analisis but was wondering if our data is different or the plot is different"
Strat, I think this is a function of chart size. With very small charts data points are often forced to jump pixels and at the micro level things can look OK but blown up to full size the picture may appear a little different. I can't see any obvious data discrepancies.

http://h1.ripway.com/78963/SDL77b.gif

Phaedrus
07-07-2009, 01:10 PM
Would you buy the next day after a close above resistance, or as soon as it is breached on the same day?Neither. My preference is to buy at (ie just before) the Close, if the price is above the resistance/trendline at that time.

wbosher
07-07-2009, 01:12 PM
Thanks for that, guys. A lot of useful answers for one little question. :D

peat
07-07-2009, 04:28 PM
My advice would be to use line charts, unless you are interested in intraday data - in which case candlesticks are the way to go in my opinion.

I'm a bit surprised by this
Candlesticks can be very informative.
there is a whole encyclopaedia on candlesticks and how to interpret them.

I'm currently playing this bullish hammer from yesterday AUD/USD

STRAT
07-07-2009, 04:38 PM
Can't wait...;)

Would you buy the next day after a close above resistance, or as soon as it is breached on the same day?

Just a wild guess, Fender man? :DProbably not. Im a bit of a snail and like a few more ducks in a row. Usually my impulsive bolt out of the gate buys are based on a good FA story rather than day Trader style TA.

STRAT
07-07-2009, 04:46 PM
Strat, I think this is a function of chart size. With very small charts data points are often forced to jump pixels and at the micro level things can look OK but blown up to full size the picture may appear a little different. I can't see any obvious data discrepancies.

Mmm OK , thanks Phaedrus. Dunno:confused:

Share Trader shrinks my charts. Vicky has had a couple of goes at fixing it to no avail.They are full size on my screen.

wbosher
07-07-2009, 10:16 PM
Mmm OK , thanks Phaedrus. Dunno:confused:

Share Trader shrinks my charts. Vicky has had a couple of goes at fixing it to no avail.They are full size on my screen.

iPhones don't count Strat. ;)