THE JOY OF CANDLESTICK TRADING - a Learning Experience

[cbarnett808]

Mytwopips you are right. That one would have turned out to be a SWEET trade!!!

-Assuming that we don't just close out after 10 pips that is. Even Tymen hates when that happens.

I really hate the kind of money management strategies that prescribe closing half of your units after a certain target is hit and letting the others ride, but maybe that's the solution- Most of the time we take our 10 pips and walk; the rest of the time we either get stopped out or pile on the pips. Good grief! More backtesting!

Yes, let's see what Tymen has to say.

[cbarnett808]

Another thing we could do is lengthen our take profits based on the distance to our stop losses (if the distance to stop is 20 pips, then our take profit becomes 20 pips instead of 10). This is another fundamental thing traders do. My off-the-cuff guess is that our win % goes down a lot if we do this though. In my backtesting from March 1, there were many trades where we just barely squeaked out 10 pips before the candle raged the other way- These small wins would now have to be classified as losses.

[tymen1]

To Kenneth Lee :

Good to know that you are having success with multiple charts with Oanda. You seem to have come to grips with a good entry with Oanda.

[tonymand]

Gotta say I have been testing dealbook for awhile since you first mentioned it Tymen and I have come to agree with you that it is much more versatile than MT4. Started a conversation with Kinetic as well. Fancy the idea of an Aussie based broker

[tymen1]

To cbarnett808 ;

Your quality posts deserve careful answers.

Risk/ reward would be a wonderful next topic.

I totally agree with this. We need to work on this because it is not right yet.

....... big risk/ small reward breaks a prima facia rule of trading

This is what we are up against at the moment and that is why I wanted to discuss it.

The Daily FX reports also come out with an analysis, but they are based on daily charts only. A plan is proposed with the risk/reward in the correct order but if you do lose, you will lose a lot of pips!
Further a risk/reward of 1:12 is analysed with good results. But the success of the candlestick patterns reduces to only 36%.

I am trying to get frequent trading opportunities, a high success rate and very small losses. So I want everything.

I have some suggestions and will also introduce KC 5 min entries - intermediate method.......... which I believe will get you better entries for a start.
Also, I need to explain my 3 stage entry system which is very powerful for entries.

This will go some of the way in solving the risk/reward problem but not nearly enough.

Either the risk needs to be lowered (lower the stop loss intelligently) or we need to let our trades run the full course - and that we need to do anyway. So minimize one end and maximize the other.

The ratio was something like 27:34, a truly amazing #!

I counted all candlestick formations, not just ones on the Bollinger Band. I know what you're thinking, but when I subtracted the ones that weren't on the band, the ratio didn't significantly change.

That is very interesting. It is also very good news. Possibly explains why other candlestick traders do not consider the BB but they do, of course, trade with the more reliable longer term candlesticks.

But, while this may work with the 5 min KC entry method, I do not think it will work well with my 3 stage entry system. These matters need to be shown.

So much to explain...............not sure where to start.

We have a very solid system here- better than any i've seen yet.

If we can get money management straight,........(....) may be as close as we get to a holy grail!

Emphasis mine.

I absolutely know this to be true. We must assume that candlestick trading is the Rolls Royce of trading and work from there. We must not be distracted by indicator trading by turning to the left or right but rather remain diligently focused on our candlesticks.

[tymen1]

To Tonymand :

Just popped in while I was posting I see !!

Good for you Tonymand. I hope you come to a satisfactory result in your contacts and experimentation!

You are most welcome to post suggestions on dealing with this risk/reward ratio problem we are tackling now.

It seems to work ok in long term candlesticks but I am trying to get it to work well in our short term patterns.

[tymen1]

To cbarnett808 :

Your chart is a 30 minute chart, is it? Not sure why you have a Keltner pattern on this chart - you do not need it there.

Your thumbnail pattern shows an impure evening star pattern with the first candle high greater than the star. I would not have traded this pattern, but even so, it shows excellent profit and, as such, the initial retrace is of little consequence.

Just further along is an engulfing pattern calling to go long. It succeeds on the next candle producing only about 10 pips but subsequent candles send the trade in reverse.
This pattern was not on the BB.

The picture is fine - and thanks for the magnification!!

Again, I'm making a lot of assumptions, but I also have to be realistic in that I will probably never get the best possible entry. As I said, this one worked out, but I've had many go against me, and we would have to successfully scalp three more setups just to get above even.

My KC intermediate method 5 min chart entry will do significantly better but even then a certain amount of skill is required and perfect entries cannot be guarenteed.

In our search for a good risk/reward ratio and, therefore, a better trading tatic, we will need to let our profits run but limit our losses.

That is, we will need to :

1) Revise the stop loss position. We may need to set it earlier and thus lose more of our trades but lose them at less pips. In fact we may need to set a rigid stop loss with a predetermined number of pips.

Research on the longer term trades shows that this tactic dramatically improves profitability.

2) Let our profits run but only to a certain point. Again, research shows that a loose profit target can be improved on greatly by setting a fixed profit target.


My computing knowledge is recent and I do not have the skills to do backtesting.

Also I do not think that the GFT program provides for backtesting. Correct me if I am wrong!

So it would be interesting if a reader with the relevant skills could backtest say EUR/USD with the following parameters.

Pattern=evening star.
Profit target=30 pips. Timeframe=20 minutes. Entry at market or open of 4th candle.

a) stop loss set at 30 pips
b) stop loss set at 20 pips
c) stop loss set at 10 pips
d) stop loss set at 5 pips.

Total number of trades to be tested to be given.
Results in % of winners to losers.

I can provide a computer code/program if you need one.

[tymen1]

To Vulcan Classic ;

Looks like you have finally decided on the GFT platform.

I like 4 charts per screen but will tolerate 5 (no other choice!)

I guess you can set it up two different ways -

1) all pairs on 20min on one screen/workspace.
all pairs on 25min on one screen/workspace.
all pairs on 30 min on one screen/workspace.
etc.....................to 9 screens/workspaces.


or



2) The way I have set it up with each workspace showing a range of timeframes for only one pair.
Then 2 workspaces cover all the timeframes for that pair and...............
9 workspaces cover all the pairs.


I think I prefer the 2nd method because if I spot a pattern then I can see
quickly what is happening in the timeframes on each side of it.

[cbarnett808]

Yes. That was EURUSD 30 minute. Thank you for the guidance on the candlestick formation. I only like to trade EURUSD because the spread is so low and most other currencies move pretty much in tandem or exact reverse of it.
My backtesting (if you can call it that) where I counted the win-loss ratio was done manually. I have some very limited skills in making Expert Advisors in Metatrader, but I don't think these can be set up to recognize candlesticks. The inputs for the advisor are very discrete and usually rely on indicator X reaching measurement Y. Candlestick formations have too much subjectivity in their recognition.
It seems like I read about someone (maybe earlier on this thread) talking about a program or indicator that recognized candlestick formations. I think they said it was crap though. I'll see if I can dig into this. My guess is it's got to be done the old fashioned way.

[mytwopips]

InterbankFX has a free candletick pattern ea alert. I have not used it or know of how well the quality of it is. Just sharing in case someone wanted to give it a look.

One more thing about risking more than making. The way I trade on ranges by building positions I find myself to be forced to do this. I won't complain because it is working just fine for me. The risk is justified by the win/loss ratio. I hardly ever lose. The downside is a loss will blow out 2 or 3 wins. I have never shared this because I just knew if I did that I would get mm flames flying left and right. Maybe having a system with a good win / lose ratio and risk / reward ratio is trying to have your cake and eat it too.