NEW!! - Candle Aggregation indicator

nitman,

They are for example purposes, we could produce custom TF with this, you can start with normal AggrFactor, and begin with anytime you want according to specific pairs, you can scan them for specific patterns, again this won't suit everyone, but then we have choices.

But for manual trades, you can choose to what David suggested.

It's easier if MT4 provides debugger, we have to come up with our own log stuff.

Hell Hannover, Mason and company

Fantastic invention and upgrading. You guys are jewels. It's a long time since a great indicator had been made.
Thank you very much.
P.S Could you please post the improved version that had been corrected on the first post each time. Thanks

David,

This is quite stable, hope everyone can do their own custom TF with no hassle.

The only bug known is when the chart is disconnected and reconnected, the indicator will skip bar painting to the last bar, so we need to make sure the chart is always connected.

Sorry, had to delete Right Candle, because I need to design real-time updates for Left Candle.

BTW, if someone put this in a good use, please share how you do it.

Mason, thanks again for all of your time and effort. Sorry that my reply is belated. I’m afraid that the recent posts here somehow passed me by; the FF forum moves so fast, that the thread was onto page 2 before I rediscovered it. (Guess I’ll have to consider “subscribing” to threads again, I’d previously cancelled that feature when I found my e-mail inbox was filling up too quickly ).

Anyway, IMHO you’ve done a brilliant job in fixing the starting point, while maintaining the horizontal shift (by simply starting on a different candle). I’ve added CA v5 to my set of frequently used indicators. It has the advantage over periodcon.mq4 in that you don’t have to open a separate chart just to obtain a synthetic timeframe.

One question: what does the wingbar option do?

Highway, thanks for the feedback. It’s gratifying to know that the idea is proving useful.

Nitman, I believe that you’re correct in saying that synthetic timeframes (H2, H3, etc) won’t be watched by (m)any other traders, hence the “self-fulfilling prophecy” effect of candle patterns on these charts is reduced. Two points, however:

1. This indicator can be used to create (for example) synthetic H4 candles from an H1 chart, but from different starting times (e.g. 00:00, 01:00, 02:00, 03:00), thereby aligning them to other brokers’ MT4 platforms, and hence allow us to see the candles that traders using those platforms are seeing.

2. This is not so much a method, as another way of viewing a chart. For example, traders use indicators (e.g. stochastic) not merely because they believe that others will calibrate and use them as entry signals in exactly the same way, but also simply to highlight phenomena that wouldn’t otherwise be apparent from the chart itself.

David

David,

You're welcome, I should thank you first for bringing this idea to my attention.

wingbar is used to make the candle/bar thicker, so we can see easily see eventhough the chart is zoomed out, I am experimenting many custom TF (synthetic) for daily, it can be used to scan Pin Bar easily.


But using this indicator, we are lacking of using standard indicators such as moving average, so when I have more available time, maybe I will look into that, but if people here are faster, they can contribute as well.

Hanover, Mason,

Great job on this Candlestick Aggregator code. I like having the H4 bar right on my H1 chart.

FYI, although it gets a little busy, I also like adding the indicator 4 times, each with a different offset time. That way I can see what the H4 bars look like on each of the different possible data-feeds. My feed is GMT+3, but I can see GMT+2, GMT+1, and GMT feeds easily by doing so. If any one of them has a pinbar, it is visible. Over time, I might figure out which feeds have pinbars that are more often significant than the others.

(Consider adding an option to repeat the aggregate every X bars. In my case, X=1. In a daily chart, X might equal 8, if based on H1, or 2 if based on H4. You'd get 3 daily bars offset by 8 hours from each other).


I do have a question I hope you can answer because you are using periodcon v1.5 with the offset feature. I would really like to create H4 charts with each of the 4 hour offsets but I can't get periodcon v1.5 to work. (Similarly, I want Daily charts with GMT, NY, London, and maybe others).

In my H1 chart, I attach the indicator, turn on DLL's, etc. and I have no problem getting an H2 chart to work, for example, with a multiple of 2. When I select File->Open Offline, I can see a gray-color icon that says "GBPJPY,H2" and I can open it just fine. The offset works if I use it.

However, if I use a multiple of 4 and a non-zero offset like "-1", when I use the Open Offline command, I can see only the normal gold-color icon for "GBPJPY,H4". This is NOT my synthetic H4 chart with the offset! Rather, it is just the normal H4 chart from my broker. No matter what I do, I can only see one H4 choice in the list. I want to create 3 more, with offsets of -1,-2, and -3. And then when I can open them, I want to attach my full set of indies to each of them and find out which offset(s) work best for my systems!

At present, the only way I know to do this is to open demo accounts on brokers with the different offsets... but of course, I don't want to change brokers just to get a different offset from my current broker's GMT+3 charts.

Have you run into this issue? Do you, or anyone who reads this know of a solution?

Thanks very much for any suggestions you can give me.
Pips4life


P.S. If you like watching for pinbars and other significant chart formations, try using the auto-pattern reconizer by Aha. It's called "PriceAction AHA v0.2" Great indicator! (Lots of alerts though, but that's configurable). Go to:
http://www.forexfactory.com/showthre...t=8491&page=40

See post # 592. Here's the direct link too, but, I suggest checking out the rest of that long thread. LOTS of useful tools and tips!
http://www.forexfactory.com/showpost...&postcount=592

One more related question to ask...

Does anyone know of an indicator which can display moving averages from higher timeframes onto the lower timeframe charts?

Again I'd like to display an H1 chart, but to put the exact equivalent of moving averages as calculated on my H4 chart. And just like this candle aggregate tool, I'd like the ability to declare a time-offset that determines which 4-hours are clustered for the MA calculations.

I need not only SMA, and EMA, but also JMA (Jurik Moving Average). To a lesser extent, I'd like also Smoothed and Linear Weighted MA's.

One approach is to build the entire indicator which aggregates the MA's just as you did for this candle aggregator. However, a more general purpose solution would be to pull the information from the MA's and other indicators off of an H4 chart. Seems like less custom coding would be required.

Any ideas? This programming task is way beyond my present capabilities. I'm hoping some tools already exist which can accomplish this.

Thx.
Pips4life

pips4life,

What a request here, I am glad that someone step in to say something like this, while I do not really understand your needs, but I do have 60+ charts per pair with this indy, it begins with M15x4, H1x2, H1x4 and so on, well the nice thing about this indy is we can easily change the settings unlike periodcon, and someday if we want to go for an EA, it will be more possible and easier.

For PA such as PB, yes I do monitor them, and found that they are related as they are shown to us, just make sure you have your eyes trained to spot them with nice and easy.

About MA, this is where I couldn't help myself, and still figure it out how to put it with this indy, I have started a simple MA, but it's not even ready for beta phase. Once we get over it, we can move to more complicated MA, but IMO I don't think we need Jurik, besides it's very complicated in nature, people are not using them, but you could explain more about this matter if you like.

Pips4life, many thanks for the link to the AHA indicators. I will definitely take a look.

_______________________
I assume you've seen the MTF (multiple time frame) indicators. If not, they're attached in the ZIP below. Re the MTF moving average, you can set the parameters TimeFrame (1, 5, 15, 30, 60=H1, 240=H4, 1440=daily, 10080=weekly, 43200=monthly); MAPeriod; ma_shift; ma_method (0=SMA,1=EMA,2=smoothed,3=linear weighted); applied_price (0=close, 1=open, 2=high, 3=low, 4=median, 5=typical, 6=weighted close), and even Currency (pair) to suit. Plotting the indicator multiple times allows you to have several different timeframe MAs on the same chart.

Re Jurik indicators, I don't see why they should be any more reliable than "conventional" MAs, even if they do reduce lag without compromising smoothness. It doesn't matter what indicators we use, we still have to enter on a price bar, and almost any indicator can be calibrated to provide an earlier or later entry. For whatever my opinion is worth, more detail from me here:
http://www.forexfactory.com/showpost.php?p=1558683
and chart examples here:
http://www.forexfactory.com/showthre...226#post172226


_______________________
I think the problem is that the actual (as opposed to the synthetic) chart already exists, and the periodcon indicator can't overwrite it. However, I will investigate further and report back here. Mason's latest version of the CandleAggregator indicator is a possible workaround.

Best wishes
David

I am working on MA for CA, it will be inside the indicator, and only simple MA, because currently that's what I need.

In the mean time, here is what it looks like.

I check the results are similar to standard MA, so I will discontinue it, just make the settings of MA indicator to be AggrFactor multiplied by MA period you like.

For example : if you set AggrFactor=2 and want to see 3 MA period line, you can apply MA setting with 2x3=6 period.

Incredible work, guys.
I have an idea and I want to bounce it at you without any practical implementations for now, just as a pure dream.
I imagined ideal moneymaking machine. What it does it scans incoming prices for a bunch of currencies; let’s say all 19 currently available in IBFX. I would go for all timeframes above 15 min or 1 hour, for example. Ideally it does all standard and all possible synthetic frames. What means all possible – say, if longer than 30 min U-turn in price (pin) is registered, it compares it with the previous n same periods and if they satisfy some other requirements it does the trade. I am talking about previous n periods because I believe that absence of traffic for several bars in front of the pin is an essential part of its success.
It does not actually print the charts; I don’t want to even look at them. Once an ideal pin (U turn) is identified, the machine trades it. I pretty much visualize the ideal pin, but the details can be worked out. I mean high/close, size of the pin, relevance to the previous bars in the same timeframe etc.
I am not a programmer at all, so I look at it from a naive point of view now. Although if I really decide to make it happen, I can take on it and go far. Or, I can find a professional programmer to do this if I get really serious about it.
How real or unreal the implementation of such trading machine would be from your experience? How much work would be involved? What problems do you see right away?
I would not be surprised if it is not achievable on MT4 at all and it requires a completely different programming environment or feed.
What is real and what is unreal in meta language? Standard timeframes may be enough, so extra frames are not necessary, but if they are possible, why not? More pins, more money.
What do you think?

Nice idea, zzz, here are some of my thoughts, for whatever they're worth.

In MT4 parlance, a robot trader is called an Expert Advisor (EA). I'm not an experienced MT4 programmer, but I believe that there are some limitations with EAs. For starters, my understanding is that you would need to have 19 different EAs running, to simultaneously trade 19 pairs (somebody please correct me if I'm wrong about this). If MT4 imposes too many limitations, you might want to use a different platform.

Even apart from the programming considerations, the concept that you're describing is complex. Simply locating a pin bar, or other candle pattern, that meets a given set of rules, does not in itself necessarily guarantee a high probability trade. If we take a pin bar as an example, there are different levels of quality (e.g. length of "nose"; location relative to preceding bars; whether there is confluence with S/R, Fibos, etc) and also different ways to time entry, depending on how you want to "play" the pin bar, some of which comes down to judgment and experience.

Of course exit must also be considered, which necessarily involves a trade off between managing risk and maximizing profit. To what extent one uses profit targets and/or how one trails one's stoploss might depend on surrounding market conditions. It might also depend on whether the pin bar was located at an extreme, at S/R or a Fibo retracement; to what extent prices are trending; where the surrounding support/resistance occurs; and different strategies might be appropriate for different pairs and/or timeframes.

Another area that likely would require research is whether pin bars on synthetic timeframes work as profitably as those on conventional ones. At face value, a pin bar simply represents a movement of price to a local extreme, and then a reversal back to somewhere near the point where price was when the candle opened. In this sense, a synthetic pin bar works exactly the same as a conventional one. However, it could perhaps be argued that, with markets being driven by collective sentiment, the wider a pin bar's prospective audience, then the more traders that are likely to be influenced by its presence, and hence their actions are more likely to drive price accordingly (the "self-fulfilling prophecy" concept). So a synthetic pin bar might not impact collective sentiment to the same extent.

Of course I'm just using a pin bar as an example. I expect that much the same considerations would be required for other candle patterns.

I'm not saying that your idea couldn't be implemented successfully. However, to get the best possible result, it would be a case of having a master trader define all of his strategies - and their rationale - across a wide range of situations, as the basis for ultimately devising some mathematical rules. It might be necessary to examine hundreds of charts - including different pairs and timeframes - to capture all of the nuances and finesse that are part of his master mind-set. Things that appear obvious graphically don't necessarily resolve themselves easily into a mathematical formula. However, I guess it would eventually be possible to build enough intelligence into the robot trader to have it simulate the decisions of the master.

In summary, a good idea, and definitely possible, but IMHO not a trivial task.

David

Greetings. David you've done a wonderful job with your candle indy. FWIW, there are several EAs out there that can trade multiple pairs from one placement of the EA on one chart. The basic idea is to place the EA on a pair that you may not actually watch, turn the candles to line/none and just view any text the EA may be writing to the chart as indication to entries/exits/stops etc that its logic may have placed.

I still like the old fashioned way of placing an EA on each chart as i can actually watch/verify PA and that particular pair's activity.

Again, excellent work!

Cheers,
Thom

thanks David,

i was thinking only of pinbars. without too much research, i would say that if the size of the bar is twice bigger than the previous 5-6 candles, then i would take it with my closed eyes, knowing that it will go minimum 70% of its size in the opposite direction. Especially if you look at smaller frames, like 15 min to 1 hour. things are more leveled size-wise on higher frames, so it won’t work that nicely, I think. this is my whole point - all the other setups are questionable and require consideration. i think the pin is unique.

i decided to find an example, jumped to the first 15 min chart, and here it is in all its beauty. i normally trade daily charts, just for these several days scanned smaller timeframes and I found quite a few of such pins on 15 min charts. they live there, to my surprise.

if you search for them in real time you won’t be able to scan all charts endlessly hunting for them, it is an exhaustive 24/5 hour job, so it becomes to be a question of being on a right frame and right currency in the right time. i did 5-15 min charts last year, and I won’t get back to it, no way. in case of such a pin, you have a couple of minutes to enter. after that it is too late.

i am sure that if such bar exists on any timeframe, including artificially created, it will work. It reflects the price dynamic, which has nothing to do with perception. don’t forget that whoever trading any currency is looking at totally different picture just because of a timezone shift, so the idea of collective reaction of traders, to me does not make too much sense. Where the majority of traders are? In the US, in China, Russia? Who knows. They all are looking at different time representation of the price even using the same, say, 4 hour bar, so a nice looking IB4 is something different in a different timezone.

several such pins per week can make this concept worthwhile.

what do you think now? i may be wrong, i am just thinking aloud.

forfot to add: i look at such pins as free gifts from the market. i am thankful to jim, who opened our eyes to see this. then, it is sinful not to take such gifts.

how many people where staring at their charts in 5 minutes that this trade was available? i missed it, and i am sure you did, too. that't the whole point of my dream - stop missing such gifts. if you trade only this with 1% of your account per trade, that is all you need to reach anything you want.

Accrete – Thank you for your response. I stand corrected. (Also good to see that the 1KT guys are still available to answer newbie questions outside of the Trading Room).

Zzz – Wow, that’s an impressive pin bar in your example, and one that would have likely produced a nice profit.

I expect you’re right: scanning 19 pairs x possibly several timeframes manually for pin bars, while attempting to simultaneously enter and manage positions, would be a (literally!) mind-blowing exercise.

Interesting that you prefer the shorter timeframe charts, while Jim of the James16 seems to look for pins on the H4 and above. One drawback re the lower timeframes is that the costs (spread) represent a higher percentage of any move, eating into one’s profit.

I think you’re probably right about pin bars representing a price dynamic, no matter what the timeframe. Would be interesting to do some research. I agree that, with H4 charts upward, people in different time zones will see different bars, but with H1 charts and below, that shouldn’t be the case.

I’ve also found that price tends to stall, and occasionally reverse, at points of prior support/resistance. The pin bar in your chart is a good example. If you look at a H1 chart of NZDJPY, 80.10 is a point of prior resistance, and also the 38.2 Fibo between the local high at 82.73 and low at 78.50. So pin bars that represent bounces from these points may have a higher probability of success...... but of course that would make the algorithm significantly more complex.

I guess it ultimately comes down to a simple process:

1. Devise your mechanical entry and exit rules, i.e. the way you will “play” each pin bar.

2. Test and refine these rules until you’re convinced that they are profitable statistically.

3. Ensure that the rules are defined in strict mathematical terms, and find yourself a programmer to implement them.

If MT4 is to be the platform, then I’m afraid I’m not that programmer. I’ve never written an EA before. But I’d be very interested to see the end result!!

Thanks for sharing
David

yeah, i'll think of it. just for the sake of conversation: here is another one, on the same pair few hours before.
i think 15 min charts have their benefit: perfect pin will have outstanding size, while on higher timeframes size difference is less significant. but, i believe there is no way you can trade them manually, i would go crazy.
see my point? if you take only these 2 pins in a week with 1 % risk each, this is 2% of your account per week right here.

i really have to think of it........ thinking already.

zzz,

It's a very good idea, in fact it's a great idea, you are at the same thinking as mine, that's the very first thing why I was interested in this CA indy, and why I am still developing now. I have a high hope that this will break a major wall of current common TF and brokers time (offset), so we can view in freedom whatever view we like, including the odd ones.

Currently I have 60+ charts per pair and going for 100+, they represent each view that I need to see in order to find signal, to confirm or just to monitor.

And yes, it's a very tough job, that's why I am still developing alerts for that, so I can use my quality time to focus on charts that give me the alerts, but it's not an easy thing to do, because we have to find the right setting first for each unique TF and pair, so my first task was to describe what is PB and its quality, and here is your chart now.

As far as I can tell, PB is the one we are looking for, it's easy to spot, the other patterns will follow/confirm the movement, it will never trigger me to develop this unless I joint PF, it's a life-changing experience over there, just make sure the person spend a lot of time of reading and re-playing the videos.

A little explanation:
1. ATRMA50 = ATR(1)=>MA50 (simple)
2. MA4 = (MA(3,0,4) at bar 0) - (MA(3,0,4) at bar 4)
3. MA3 = (MA(3,0,4) at bar 0) - (MA(3,0,4) at bar 3)
4. MA1 = (MA(3,0,4) at bar 0) - (MA(3,0,4) at bar 1)
5. Previous MA1 = (MA(3,0,4) at bar 1) - (MA(3,0,4) at bar 2)
6. MA(3,0,4) = 3 period simple MA with Median price
7. PrevBars = how many bars before
8. DD = drawdown in bar# and pips
7. TP = maximum take profit in bar# and pips

Another thing about smaller TF, I've seen M5 also produces PB, but we need to adjust the visual parameters, they pop up in different way than other higher TF.

thanks for the reply, mason.


i noticed 5 min pins, too, but somehow i feel better about 15 min charts. they are, kind of in between of totally unserious shorter timeframes, and something a little more reliable. if there is a good pin on 5 min it will appear on 15 min, too. i am not talking about just pins, i am talking about exceptional pins, like the one I attached with my first chart.

my problem is that i am not a programmer at all, and i am not sure i want to follow this road. i can if i really want to, but it may be cheaper and definitely faster to hire somebody to do it for me, although i am not sure how much money it may cost. i definitely mean EA which would probably work on 15 min charts. i don't think there is a real need in any other timeframes. if a nice pin appears, it will print itself on all frames, from 5 min to, possibly 1 hour.

i don’t even know how to backtest in mt4, except by using F12. i can use Excel, though, i am much better with it.

generally speaking i am against indicators and all this technological garbage on charts. i tried to trade with them half of the last year, and i did not achieve anything.

it looks like i am doing really well on daily charts now by exercising jim’s magic without any indicators, and i like it so much that I am not sure i want to ever go back to smaller frames. maybe 4 hours in the future.


it’s just seems so reliable when there is a super pin in a right place on a shorter timeframe when i am not around that i feel like a waist of not taking it. i have a full time job, which i am hoping to end with the help of my trading, so i am deadly serious about the whole trading thing. because of the job I am never around to take decent short time frame pins, so the whole idea is my revenge for this.

before i do anything i will play with this concept a bit more. i tried to put algorithm together, it’s about 10 conditions that i want to be met before EA is taking a trade. 10 settings to accommodate or possibly more. how difficult is this?

i did not backtest it yet, it is just based on my observations of the charts. i am thinking…..