Index Charts

Has someone build an indicator that will construct Index charts? I would like to see Index charts for the following major 8 currencies for all Time Frames M1,M5,M15,H1,H4,D1,W1.

AUD, CAD, CHF, EUR, GBP, USD, NZD, JPY.

I want those charts to be based on mathematical calculations used based on price of all the 8 majors. Want to see these 8 charts just like any other currency pairs with each candle having its Open, Close, High and Low. Want to be able to add all and any indicators to it just like we could for any other pairs.

Was wondering if anyone has done any work on this already? Or if there is some tool on the net that does the same?

Yes, done that. However, you can't do it with an indicator. Or maybe you can but I never tried. I create the charts as regular new instruments in the terminal and use them almost as any other symbol. However, creating new symbols and charting them real-time is not easy and seems a bit complicated at first. You need two terminals: (1) One disconnected for charting, (2) another one connected to your broker and playing the price feed.

The challenge here is not calculating or charting a custom instrument. The issue is updating the history once your charting terminal has been disconnected. For an accurate index chart you need ticks. Close prices produce something similar to "line charts with wicks". That may be good enough for last months chart (or older) but for today (or the current minute) you need ticks. So only for having an accurate tick history I run a tick recorder on my VPS. Then the first thing after my charting terminal starts is synchronizing/backfilling missing quotes in the background. I use rsync for that, syncing a 300MB history file of 9 timeframes over 3G takes only seconds. Syncing all timeframes of all currently recorded instruments over 3G takes about 60 seconds (1-2GB data). With regular data transfers those rates are not possible. And speed can still be improved.

With this approach it's not only possible to create index charts but any synthetic instrument or custom currency basket you can express with a formula. At the moment I'm working on a major rewrite to simplify instrument creation and trading of such an instrument. After all I want to use a synthetic symbol like any other instrument and load it in an EA via iCustom(). Having it available only in a disconnected terminal makes trading difficult. I implemented trading of remotely charted instruments but it is too complicated. It's better to have the timeseries of a synthetic instrument in memory in the EA.

An overview for synthetics is here: ICE futures indexes, synthetic currency indexes and custom baskets. At the moment I completely rewrite synthetics, so there is a lot of work in progress. Some details about how to connect a disconnected terminal to a custom feed are here: https://github.com/rosasurfer/mt4-expander/issues/2.

Finally a screenshot how it looks in real life. Bottom left is the recorder terminal feeding ticks to the disconnected charting terminal in the background. In the brackets left-side of the quotes one can see the current synthetic spread. This is in fact the spread you can get if you trade those. As a comparison: The original LiteForex spreads are 10 to 50 times larger than that. They make a lot of money only with the spread.


That is a lot of work and maybe somebody else produces something similar just with an indicator by drawing candles manually. I can hardly imagine whether it may be possible or not.

If my approach is not too scary for you and you are seriously interested then by no means - just ask.

PS: Everything MT4 only, no MT5 crap please :-)

You have indeed put in a lot of work into this. Thanks for sharing your ideas. I will sure check out all details you have shared and get back to you with questions. I appreciate your help on this subject.


I was thinking along those lines. Following was my thought process:
1) Write a script that would download historic minute charts of all 28 pairs.
2) Use these construct the custom 8 Indices that I want. This will be done using Excel formulas.
3) Import the 8 Indices charts into MT4.

All the above steps should display what I am looking for if done correctly. I personally don't care much about tick data. I can live with 1 min accuracy. It does matter that I am able to see all Timeframes for all the 8 Indices.

I am not sure when/if I will get time to work on this. So was wondering if anyone has done all this already and willing to share?

You read my mind. I'm 10 years in FX now, guess I can tell one thing. One of the biggest issues in retail trading (beneath the lack of good information and a too big exposure to bad information) is that people all around the world are constantly re-inventing the wheel. We are just not connected good enough, not connected at all, opposite to the industry. We are easy prey.

1) Write a script that would download historic minute charts of all 28 pairs.

1. done: bin/dukascopy/updateM1Bars.php

2) Use these construct the custom 8 Indices that I want.

1. done: bin/rosatrader/updateSyntheticsM1.php

3) Import the 8 Indices charts into MT4.

1. done: bin/metatrader/createHistory.php

All the above steps should display what I am looking for if done correctly. It does matter that I am able to see all Timeframes for all the 8 Indices.

1. done: see above screenshot

You don't need ticks for producing intra-minute charts. You need ticks for creating M1 charts. If a synthetic instrument is made up of 7 other instruments using OHLC of M1 is not enough. It results in candles with longer than average bodies and even longer wicks. Unusable.
The reason: To calculate the High of M1 you cannot just put all the Highs of the components into the equation. The highs occured at different times and equal each other out. If you calculate M1 by using OHLC of the components you will get long bodies and even longer wicks. That's why: for everything older than a few weeks these distortions can be handled (you may mathematically reduce each candle by 20% and each wick by 30-40%). For current data this is not acceptable, and for current prices you have to record ticks anyway. It's not such a big burden as it seems.

My idea is to put all the calculations up on a public site. Of course average traders arenn't able to run any scripts. Therefore all calculations are done in a central place. The average beginner can download his customized history files. However, the more advanced trader gets remote access via SSH and can synchronize the created history files with full speed. This not only for indexes but for any custom synthetics, you just have to enter your formula. This also covers custom baskets. Basket trading is very similar to index trading and considered to be one of the last holy grails. Why? Because nobody can see it...

So far for my dreams, just to give you an idea :-)

Maybe I am still missing a piece that you are trying to explain. I will do some research myself that way I don't waste your time asking you 1 million questions on the subject. I will just ask you this for now....

What is wrong with pulling the last closed 1M candles of all 28 pairs? Use its open and close values of all 28 pairs to construct 1 candle for each of the 8 currency? This should create 1M index candles for each currency without any wicks or tails (I really don't care about it)

Once we have 1M candles for each of 8 currencies, it should be easy to construct 5M, 15M, 1H, 4H, 1D, 1W candles from this. Ofcourse all these candles will have wicks and tails and these IMO should represent almost true market conditions.

Note: What I am saying here is with an assumption that I do not care for 1 minute movements of markets? I actually don't even care about 5M movements.

You are right. It all depends on your perspective and if you don't care about sub M5 movements you may just ignore the finer details. It's not my job...

From "M1 candles of all 28 pairs" you can only calculate a line chart, not a candle chart. This is because you can only use the close prices, nothing else. If you want to chart an M1 candle chart you need ticks of the sub-components, there is no way around that. I'll try to find the link to the russian thread where this is explained much better than I could ever do (I'm a developer not an author). May be somebody else...?

When the moment comes that you realize recording ticks is not a burden at all, then you understand there's no reason not to capture those details. Usually people think recording ticks requires special efforts but it's not like that.

Why not? Let me give you an example. Lets say we are looking M1 candle of 7 pairs of JPY. We look at each of this candle and find out if bearish or bullish and we also find out the difference between the open and close price of the candles (all 7 pairs) Some candles will be (-) some of them will be (+) We sum all this. If the final number is net (+) this means our resultant index candle for USD will be a bullish green candle and vice versa of whatever the sum we derive from all these candles.

You are right! I am one of those people who believe recording ticks is quite a lot of effort! :-) Guess everything is hard when you don't know...everything is easy when you know! Maybe soon I will agree with you 100% about recording ticks as well.

Yes, and in this way you get close prices. Now if you connect close prices you don't get a candle chart. You get a line chart.

Everything online is easy-peasy. There's only one real challenge: the Tragedy Tester. In Tragedy Tester performance is of utmost importance, so there is the place to discuss performance efforts. Not so in life trading. In life trading (recording ticks is a life encounter) even Visul Basic would be fast enough.

Good morning.

Currently I'm in the process to set something up to spare you/everybody all the script stuff. For now I create a fully synchronized copy of all indexes on the server, so it can be downloaded and charted by everyone. This will not automatically synchronize with new incoming prices but for charting indexes and using indicators it will do.

This is what it looks like when the scripts are at work:

Good work!!! Thank you.

[quote=yalgaar;11828435]Has someone build an indicator that will construct Index charts? Want to see these 8 charts just like any other currency pairs with each candle having its Open, Close, High and Low.

Ok..

I hope I am not off the topic here. If so, let me know pls

I have been delving into "Synthetic Currency" Indices. For example: On Lite forex there are non tradable indices available on in their market watch for all the eight currencies, such as EURLFX, USDLFX, GBPLFX, etc... What I like about it is that they give equal weighing for each currency against others and thus create a more realistic view of the market. This is why the USDLFX (Lite Forex USD currency Index) would be different from the the tradable one offered on other brokers MT4 platforms. The first gives equal weight to all currencies while the second gives more wight to certain ones.

Any way, these Lite Forex currency Indices are almost the same as the TradingView Indices (https://www.tradingview.com/markets/...s/indices-all/).

They have a certain calculation by multiplying pairs and then dividing them by seven.


However and here comes my inquiry and would love to know your comments on the matter please:

In tradingview, there is a feature that allows you to ADD different pairs. So for example: I can add EURUSD + EURGBP + EURAUD + EURCAD + EURJPY + EURNZD + EURCHF. and the result would be one chart with normal candlesticks to be analysed on different time frames.
(Thanks for Degolep at forexfactory for telling me about this feature)

I noticed that if I compare Euro pairs (by adding them) , the chart will be different from the EUR Index available on trading view. I can tell why there is a difference. The index is not calculated by adding pairs but rather by multiplying them.

My observation is that the addition of pairs result in a more realistic view of the currency which is better than the normal formula used for calculating a currency index by multiplying them.

What I like about this feature in Trading view of (adding/subtracting/multiplying and division) is that I can control which currencies to include/exclude and which math I can use (i.e multiplying or adding, etc..).

For example I can compare three currencies against each other: (EUR/USD/GBP). So for EUR I would type the equations as follows : EURUSD+EURGBP to produce a chart of EUR index (relative to the other two currencies) or I can type EURUSD*EURGBP/2.
Of course the options are countless. If I want to give more weight to a to EUR against pound, I can type:
EURUSD+2*EURGBP.


NB: I have attached the Canadian Index (A built in index in Tradingview) (Weekly)
and Attached a Summation of all Canadian pairs. (Weekly)
Both are weekly


Also I attached a EUR Index (a built in instrumnet in Tradingview) . (Daily)
and attached a Summation of all Euro Currencies. (Daily)


It will be great if someone can enlighten me with the math logic here. Which is better? Adding or multiplying and why?

Thank you all in advance

There is no "better"? There is only "correct" or "incorrect".

Please don't take the following statement personally:
Averaging relative values by adding them up is just mathematical non-sense and similar to saying "2 + 2 = 22" and I prove it because if one operand rises the sum rises, too.

Averaging values "by adding" is correct as long as values are absolute and measured against a fix point (that's typically 0, the zero point but it may be anything). Averaging absolute values is done by addition of N elements and dividing the sum by N.

Ratios (like currency rates) are relative and have no fix point to be measured against. Instead they are measured against a (possibly) floating point of some kind of balance and in our case that's 1. Averaging relative values like ratios is done by multiplication of N elements and drawing the Nth root of the product.

Take two imaginary currency rates: 1.5 and 0.5

What is the average? Is it (1.5 + 0.5)/2 = 1.0 ?

No, it's around 0.866025.

Why? Because an average (or a mean) must be in the middle.

1. 1.0 is not in the middle as ratios of 0.5 to 1.0 (ratio = 2) and 1.0 to 1.5 (ratio = 1.5) are not the same.
2. 0.866025 is in the middle. Ratios of 0.5 to 0.866025 (ratio = 1.73) and 0.866025 to 1.5 (ratio = 1.73) are the same.

There is also another aspect many times forgotten. When discussing currency indexes one must not confuse the index itself (representing a valuation of a single currency) with an investment in such an index (representing an investment in money). An investment is always done in real money of some type of account currency and thus always consists of two components. You buy the index (however it is calculated) and pay for the purchase with your account's money. That means an investment always means I am long/short at the same time in some amount of my account currency.

The final result of any investment is a mix of the performance of the purchased instrument (whatever it is) and the performance of the account currency over the same time period.

If you analyse the chart of a USD index investment you will notice that it looks very different than the USD index itself. That is because you overlay two instruments, the USD index and your account currency. Now I don't hold Forex investments for months. In fact retail traders cannot invest in Forex at all, they only can trade Forex. If I want to trade Forex I'm usually not interested in those overlayed composite charts. I want to make trading decisions based on the real instrument. As my "investments" are only for hours or at most for a few days I can ignore the fluctuation of my account currency. Especially because I use a margin account and can limit my exposure to the account currency to only a fraction of my real "investment".

Hello emmzett,

Thank you very much for your detailed response and taking the time to write your post. It is this kind of interaction and sharing of experience is what makes it above par.

In the meantime, I am preparing for a good post to discuss it a bit further with you.

Regards

@emmzett,

I definitely understand your logic with the example you gave and that to extract the average (mean), the addition will give a wrong value. It was a very clear example and clear explanation from yours. It was pretty much straightforward and clear.

And I also agree with you on the difference between us (retailers) and an investor like a bank or market maker. Same logic as with a Coke company hedging sugar commodity for next weeks or months.

------------------------------------

However, I will discuss "What I visually" have noticed recently rather than "arguing that addition is better". It is important you understand my logic of my "observation/visual comments".

----------------------------------------

Here is a trade I took on shorting CHFJPY. It made 200 plus pips and could have netted 500 pips.




With JPY, both the built in index (Lite Forex and Tradingview Index) and "Adding all Yen Pairs" (my approach), have shown a confluence with showing a bullish momentum. So I will not discuss the JPY Index.

What I want to discuss is the CHF Index.

Now notice the built in CHF Index (from either Lite Forex and tradingview), Before the 2nd of January, the daily chart was showing a clear upward momentum with a big possibility of continuation. However, a trade could examine this chart and argue there was a Resistance on smaller time frames H4 and below or any reason to say that CHF bullish upthrust was slowing down and that there was a Daily Resistance and price has almost hit the Resistance.

But, my argument would be that for me, CHF Index was bullish and I do not want to "Force" myself to belive it is bearish by applying too much analysis on LTF. This chart simply will never make me believe that CHF is bearish but it is rather bullish or at the very least it has hit a Resistance and is reacting it. But no bearish trend has been established yet.



But If we check CHF synthetic chart by adding all CHF pairs on tradingview, the story is quite different. CHF is indeed very bearish and there isn't a hint of any bullishness.
Look how smooth the PA of the daily chart is. A break of the TL and a retest and enough room for a more nosedive.

The difference between CHF built in index and adding CHF pairs is very clear in this example. CHF is indeed very bearish.




So again: Take this example as a means of me explaining what I "visually see" rather than arguing it is a correct method.

Please note, I use indices for direction bias and momentum and also to basket trade selected pairs. I can even trade one pair rather than a basket.


Edit: I just wanted to add here is that in "my observation", I found that addition gives an early bias for direction compared to multiplication which is more than often giving late entries. I might be wrong though. I have not tested it for long.

Sometimes, I find adding charts give an overall smoother look with less erratic movements . in other words, the combined chart looks cleaner.

---------------------------------------------------------------------------------------------------------------------------------------------------------------

This is a new question for you please:

Generally speaking, if I choose to trade a basket of selected pairs, and I want to put the stop loss at a level where I expect price would be changing its bias. How can I calculate the amount of money that would be lost? I am not a fan of a using a certain amount of dollars as a stoploss because it does not reflect where the (Technical Analysis) stop loss should be. I want to put the Stoploss on the "Customised or synthetic chart". Let us say that I used your correct way and I multiplied couple of pairs to produce a chart. I have a breakout and a retest. If I am buying and want stoploss to be behind the last swing, How to translate this into (amount of money as a loss ) and hence I can specify my lot size to trade?

------------------------------

Thanks again for your response

Regards,

Measuring Momentum or Acceleration of Currencies

I thought of writing this in a different post as it is different from last posts.


Now what is it about?


For any trader who is analysing currency (Indices) to find bias and momentum with the aim for trading one pair. The trader wants to pick a bullish one vs a bearish one. Right?

That is good and is a good way of analyzing but in many cases it limits the number of trades.

Lets imagine for example that both EUR Index and USD Index (from the indicator we have here in this thread or from Lite Forex or any other) are both bearish. So the trader would think better not to short EURUSD. But soon enough the trader would find that EURUSD did fall significantly.

Later on, the trader would find that EUR was more bearish than USD and thinks to himself "Next time , I will not look at indices", "I should not have missed this setup. I knew EURUSD was going to fall because of my pre-analysis to the pair".


Any way, I thought of sharing this to ask you traders: How would you traders examine the "Rate of Change" or the "Momentum" or what ever you may call it to tell which currency is more bearish than the other and that it is ok to short the first against the other.

The logic here would be : Just avoid those two currencies. Bothe are bearish" . But this comes with a price : Limiting the number of setups".
I myself am between the two mindsets. Or lets say I would place all the odds in front of me and adjust my risk or lot size accordingly.


Any way, I would love to hear your comments please.


--------------------------------

Add to this that sometimes when the trader picks the right bullish Index and the right bearish one, it may be a bit late to trade the pair. It has already moved.
May be if the trader picks the right bullish one vs a "Ranging" one would be a good choice? your views will be great.

Nevertheless, there is also the risk that the trader might have done a wrong analysis for the index and established a wrong bias. (This is part of the job)

Again, would love to hear your views and comments and how have you experienced different scenarios

---------------

Thanks for every one
Regards

Quoting emmzett" Take two imaginary currency rates: 1.5 and 0.5 What is the average? Is it (1.5 + 0.5)/2 = 1.0 ? No, it's around 0.866025. Why? Because an average (or a mean) must be in the middle. 1.0 is not in the middle as ratios of 0.5 to 1.0 (ratio = 2) and 1.0 to 1.5 (ratio = 1.5) are not the same. 0.866025 is in the middle. Ratios of 0.5 to 0.866025 (ratio = 1.73) and 0.866025 to 1.5 (ratio = 1.73) are the same.

@emmzett

I kept reading and thinking

Should the moving average be calculated this way?Oscillators also? Should the data series be multiplied and then divided by its number?

Another thing I have been thinking of is wether we can plot currency Indices by the same logic of using Logarithmic charts rather than Linear ones or by the logic of using PPO versus MACD Oscillator. You get the idea? We just examine cuurencies based on their Percentage Move. This will apply on Low, High, Close, Open. nd we can plot a chart based on this.

The percentage move will be our reference point and hence we can compare the strength of all currencies against each other.

What do you think?

I wish if my math was better but it has been a long time. I wish if I could write detailed equations to explain this recent idea.

BTW: I am still working my head around the arithmetic average of many instruments and try to find why they look smoother and whether they provide a fair reading or not.


@all other traders

I found an interesting article for all traders who have been reading the previous posts and wondering what is it all about.

This article explains the difference between arithmetic average, geometric mean, and the harmonic one. It is purely math and not related to forex. Read it and you will understand the logic.
https://towardsdatascience.com/on-av...is-2a703e21ea0

The author is not favouring arithmetic average and favours the other two. It is the same logic of multiplication method used here.


Thanks again for this thread that made me revisit mathematics. It is interesting and exciting for me to read more on the matter and teach myself.

I can't go into each of your thoughts in detail right now. Just one observation: The "CHF" composite chart from TradingView you posted has nothing to do with the CHF index.

Your above charts represents a synthetic basket with the formula:


If you understand that it is very different from


then you also understand why both charts have completely different shapes. While the first instrument is a mix of 8 currencies and their valuations the second instruments describes only the Swissie. It is up to you what breed of synthetic instrument you create and trade. You can literally do anything and create instruments which trend or swing nicely (Trend-O-Mat and Arb-O-Mat).

Think about how your above synthetic instrument would account for different volatility of the components. CHF is de facto still almost pegged to EUR and moves only very little. What do you really see in your chart? Is it really the CHF? What about the crazy Yen with its much higher volatility compared to the others. To answer the question: In fact you don't see the Swissie, you see a mix. You see the result of just adding up single positions.

Now, can one trade such a mix? Of course one can.

Thanks emmzett,
I understand the second equation that creates the Index,

However the first one should not have 6.42CHF or 109JPY. It is simply adding CHF currencies with equal weight.
The pair that has CHF as the base would be written as it is (for example CHFJPY) while the pair that has CHF as the quote would be written as (1-GBPCHF) for example. so I am simply inverting the chart to get CHF as a base.

Anyway, I know I was very lengthy so take your time and thanks for your reply.

Edit:
I added them I am more inclined the "Added" pairs chart as the break of Resistance is much clearer.


I agree with you that the synthetic charts I am creating can not be named as Indices.
I am trying to create a breed that is as close as possible to real market to give early bias.
BTW, I have posted in the other thread of Math Trader for more interaction. Will be great if we carry on there if possible.

Last thing, I have noticed after reading number of articles on how to calculate averages for any economic data series that the Addition is not totally refuted but somehow thought of as not very accurate. Still digging into it.

Thanks for the interaction and looking forward for more when you have time.

Regards

Edit: I just saw your PS

Yes, for sure the crazy yen has incredible momentum. But I wanted to marry it with a good looking bearish currency that has the best PA without second guess. I decided to go with the CHF.
But your point is clearly right.
It yen was bought against any other currency wouldn't we have made profit? Yes you are right. But when I bought it against CHF, the fall was over the night for 500 pips. CHF was looking great.

Note that I have not included the YEN chart as both the index and the combined chart I created looked the same. That is why I have not added it. We were discussing addition vs the authentic way of creating indices. But I will now.