What's the rationale?

That those EA that are profitable in wider spread will also be profitable in smaller spread?

First and foremost, what is the key concept that will make an EA profitable?

A matter of spread issues or an issue of accurate entry/good trading rules?

I don't know about searching with spread and then narrow it down.  But no harm testing it out.  I do a lot of experiments myself, cos If I know everything in the 1st place, I won't be here.  So over the years of testing out my ideas, I've gained some experience.  Do test it out and let us know your findings.  I'm curious to know too.

In the 1st place I'm a very rule based users.  In another words, I develop (rather than search) EA based on good trading rules, I find such method yield better results and not a shot in the dark.

Secondly if I were to search for EA, I tend to use rules based concepts.  So to search for wide spread and then narrow it down, is something I won't attempt.

But if I were to attempt a search, I will then search for EA with tight BE instead. 

Why tight Break Even (BE)?

Well I assumed that good entry EA will tend to enter at the most optimal timing whereby the EA will enter just at the "right timing" when price will start to go in the "predicted" direction and not "floating around" and go into negative loss and then start to make money, such EA is profitable but expose trader to greater risk during those "floating" time cos the EA didn't enter high accuracy.

So how to look for high accuracy EA?

Well the characteristics of a high accuracy EA is one that enter will the "least" floating loss.  The ideal one EA is when it enters into a position, it will not have any floating loss but experience only floating profit.  But of cos that's the ideal.

So in order to search of high accurate EA, we use BE to "filter" out bad Entry EA. 

if I set BE to be 150, if any EA don't managed to get at least 150 pips right, it will suffer SL.  And if those entry that got the right direction (got at least 150) but did not identify a strong trend (trend turned around), it will at least stopped at BE.  If the EA got it right and the trend is strong, it will continue to make money.

In another words, I can filter out 3 types of EA

1. Those lousy EA that can't even guess 150 pips right
2. Those EA that got it right but didn't identify a strong trend
3. Those EA that not only enter with high accuracy but also identify a strong trend

How much BE should I set?

Well if we set too tight BE, we would not give enough room for our EA to trade.  There is some level of price fluctuation that is needed for "healthy trading".

My suggestion is somewhere between 150 to 350 that's the ranging levels you often get for a "normal day".

PS: Sorry Rantampla, my earlier post was not a direct reply to your question (though it seem like it), it happened that when I was writing the SL post, yours came in 1st, so it looks like I was replying to your question.  This post is meant to reply your question asked.

Building a 40-Thread Xeon Monster PC for Less Than the Price of a Broadwell-E Core i7

For those who are serious into EA development, remember time is money. 

Imagine you take 1 year to finally develop a good profitable EA but if you have the right computer capacity, you may have reached your goal in half the time needed.  In another words, you would have made profit 6 mths earlier which may be more then the cost of your new computer.  So saving money and postpone your purchase of a good computer may not really be saving afterall if you were to count the opportunity cost (potential profit that you might have earned if you reached your goal earlier).

I've bought (or rather my husband has helped me to buy) the CPU at Ebay and assembled a 16 core dual xeon for me.  That really speed up my developmental process.  Now I'm thinking of buying a 40 Core computer   I saw at Taobao and it only cost about US$2k (unbelieveable but my friend got 2 of it...green with envy).

So here is a link to see how it can be done and you can search and find a lot more that may suits your specific spec

https://www.techspot.com/review/1218-af … onster-pc/

1. CONFLUENCE IN MULTI TIME FRAMES
We use Multi Time Frame (such as Longer time frames settings) as filters to improve our trading entry/exit accuracy

2. CONFLUENCE IN MULTI CURRENCIES

Likewise, we can use Multi Currency as another filtering method to increase our trading success.

If you observed yesterday's market conditions (Sept 20, 2017), you will noticed that beside EURUSD, almost all the other major currencies (including Gold) that paired with USD also experience the big movement.

In a ranging market, USD price may go down, not because USD has gained strength but rather it could be EUR has gained it's overall strength, so relatively speaking, USD prices will go down.  For example, if your friend grew taller, you may looked shorter not because you have grew shorter but that your friend has grew taller and hence, relatively speaking, you have "strung".

So if you observe in the different currencies/markets, you will noticed USD may go up in one market but go down in another market.  There is no confluence.

Hence, if we want to enter into a buy/sell position, we need to examine also the overall USD strength/weakness by checking for multi currency confluence.

REQUEST FOR MULTI CURRENCY EA

I would like to request for MULTI CURRENCY EA FUNCTION

Popov, in FSB software, when we build our strategies, in each opening/closing conditions, we can choose which currencies or the "default" currency.

Can I suggest that FSB enable users build multi currency EA?

For example,

If I input 2nd opening rule (cos 1st rule is Bar opening) that Fast MA is higher than Slow MA for EURUSD
And then add 3rd rule that Fast MA is higher than Slow MA for GBPUSD,
4th rule, Fast MA is lower than Slow MA for USDJPY
5th rule, Fast MA is lower than Slow MA for USDCAD

If all the conditions are TRUE

1) enter into SELL position for currency pair that I attached the EA in that particular MT4/5 time chart.
Eg, If I attached this EA in EURUSD, it will open SELL position or trade in that particular currency pair.

NOTE: BUY (USDCAD, USDJPY) and SELL (EURUSD, GBPUSD)

2) if I attached this EA in MT4/5 charts in 2 currency pairs, it will trade in these 2 currency pairs.   

I believe using Multi Currency Confluence method will greater help us in entering in stronger trend and thus avoiding entering in weak trends that is not worth the risk considering the reward is lower.

Since FSB's setting already has this "infrastructure" whereby we can choose different currency pairs in each opening/closing condition/rule, the only thing left (in my limited understanding) is to enable the coding to be incorporated and be implemented in the EA.  It doesn't matter to me if the back testing results aren't incorporated yet or can't be done cos I won't want to be "deprave" of the ability to trade in multi currencies execution method just because of backtesting's limitations (cos in this world, there are many limitations but we should be creative enough to work around it and not be bound by it).

WHY USE 1MIN TIME CHART WITH LTF (Longer Time Frame) FEATURES?

I have mentioned this many times and long ago but for the benefits of new comers, I'll mention here again in summary.

1. FSB Time Chart, is just a platform you decided where you want to install your MT4 EA. 

If you going to attach your MT4 EA in H4 Time Chart, then in FSB's setting you choose H4 as your time chart.  Likewise if you going to install your MT4 EA in M1 Time Chart, you choose FSB's M1 Time Chart.

2. FSB Time Chart determines how FREQUENT you want to scan the market for your trading rules.

This is the part some FSB's users confused FSB's Time Chart with MT4 Time Chart.

As a manual trader, I mostly look at D1, H4 and H1 MT4 time chart to spot for trend formation etc.  Hence, when I started to use FSB, I mistakenly use FSB's H4 Time Chart as my "platform" and input my rules there.

However, when I noticed the 1st rule I often use is "BAR OPENING/CLOSING" that's where I realised my mistake. 

The H4 EA will only scan the market either at the BAR OPENING or BAR CLOSING to check whether your trading rules are TRUE/FALSE.  In between the time, it does NOTHING. It will scan the market for opportunity at every 4 hours interval, such as 12:00, 16:00, 20:00 etc

So if there is a Breakout at 13:01, the EA will not enter immediately to take opportunity of the Breakout but will wait till the next BAR OPENING which is at 16:00 and do you think at 16:00 you will have captured the most out of a Breakout?

3. ENTRY/EXIT Accuracy

The FSB's Time Chart will determine your ENTRY/EXIT accuracy (besides your trading rules).

If you use FSB's H4 Time Chart, your EA will enter/exit only at Bar Opening/Closing.  That means, you not only missed the Breakout at 13:01 but you also ended up Closing when the trend may have turned around.  So from floating Profit, you may ended up with a losing trade because by the time the EA reacted at Bar Close, the trend may have turned against you.

IN SUMMARY

FSB's Time Chart is to determine how FREQUENTLY do you want your EA to scan the market for opportunities.  And you use Longer Time Frame to input your trading rules such as H4's rules, etc to tell the EA what kind of conditions/rules do you enter.  For Eg, Use M1 Time Chart and use H4's time frame (LTF setting) to BUY when Fast MA in H4 cross Slow MA.

The higher Time Chart you use, the lesser probability that your EA will have high entry/exit accuracy because your EA cannot enter at a "timely" manner if you only trade at every H1, H4 etc interval.  I personally won't recommend any time frame about M5.  In fact, I don't understand why people don't use M1, all your trading rules are the same as you intended to, the only and biggest difference is that you want your EA to scan more frequently at every bar open/close and for M1 time chart, it's every minute interval.  Hence, FSB's Time Chart is to determine HOW FAST you want your EA to react.  Do you want a slow responding EA that missed out all the good opportunity, then choose D1.

That's the reason why I ONLY use M1 Time Chart to trade because I want my EA to react every minute when the opportunity strikes.

Using FSB M1 Time Chart simply mean that it will check your trading rules every 1 minute interval.

How fast your EA will react also depends on your Trading rules.  So even if you use M1 but very laggy indicators or Longer Time Frames setting, your EA will still be "slow"

So your EA's reaction time depends on 2 Factors

1. The platform you use, such as M1 time chart vs D1 time chart, to determine how often you want to check the market conditions (every minute vs every once a day)

2. The Longer Time Frames indicators and parameters you use.  Whether its MA 100 cross MA 200 (which is slow) or MA 14 cross MA 21 (which is faster).

Hi Sleytus

1. Yes, if you use M1 Time Chart and use LTF (Longer Time Frame) setting, you change the "default" to H4 or Longer/Higher Time Frame (i.e not shorter, such as using D1 Time Chart and use shorter time frame such as M1).

2. How FSB does the back testing with LTF, this I don't know but I assume yes it is in M1 data cos when I use M1 Time chart, I assuming it's using all the M1 data to give me the immediate backtesting results

3. When Optimization, how the software handle the data, this I too don't know the "mechanism" behind both backtesting and optimization (though I'm much curious to find out too).

4. What is "Noise Factor"

Some sound is only considered "noise" when it is unwanted sound.  To others, the same "noise" is signal.  So how you "classify" noise may be another person's opinion as "signal".  So it all depends whether the sound is wanted or unwanted.

Thus we filter out "unwanted" sound as noise.  Therefore, whether it's signal or noise depends on what you are looking for/wanted.

As for me, I trade very frequently and I want to get the most out of the same time period.  Hence the ability to handle "noise" and turn it into sound is an art.  If you can't see anything "useful" out of these sounds, it is too noisy for you.  It depends on your trading rules/theory to turn these noise to signals.

Thus, I don't like to use "non repainting" indicators cause they cross out the noise for me instead of leaving it to me to decide how to handle these "noises".  I prefer to use indicators that are constantly change with price changes(repainting indicators), constantly feedback to me the price changes and the "secret" lies in how I handle these "noises" to turn into signals.  But to those who can't handle such "sensitive" indicators, they would go for Non-repainting indicators.

Likewise, to use any higher time chart such as M15 upwards is liken to those who use Non-repainting Indicators.  It cross out all the in between information till Bar close.  So it's like a manual/discretionary trader who don't look at MT4 charts in between the time but only check his trades at every 4 hours interval.  And if there is any sudden market changes, he will do nothing about it because the 4 hours interval is not up yet.  He will ignore all trade information in between the interval.  That's how Non-repainting indicators behave too.

So what's my opinion about Renko chart?

It really depends on what you want to do with those sounds.  If you don't know how to handle it, you will choose indicators that will cross out those noises for you.  I personally love sounds and prefer to handle it myself by using different indicators setting and LTF to mange them to my advantage.  So the more noises I've given, I have greater freedom to determine how to handle it instead of letting someone/indicator handle it for me.  That's the main reason I don't use any time frame above M1, cause I don't want the software to cross out those in between noises but allow me to handle it with my own rules.

In conclusion, your trading theory will shape the way you handle sound/noises.  So if you have a sound trading theory, you will look at the MT4 charts with a certain perspective.  Find or develop a trading plan the suits your expectations.  If you are satisfied as a position trader, then you are less affected by time frames or misled by false signals because you are taking a Macro view of the market.  But if you are an intraday trader, you need the art to handle these noises and turn it into good trade signals.  So how much you will push yourself to learn the art of trading depends on your outcome expectations.  As for me, I want to squeeze the most out of every period, thus I push myself to develop better and sharper EA, otherwise, I believe I would have stopped developing EA long ago if I were just satisfied with finding any kind of profitable EA.  My quest is not just to find profitable EA, my quest is to develop highly accurate and highly versatile EA that is robust and profitable...that's very ambitious to say the very least (and to some, they think I'm being unrealistic but it's because of this relentless determination that I find myself getting better and better EA over the years...so it really all depends on your expectations and commitments).

Lastly, there is no such thing as passive income if you are a trader.  You will never come to a point whereby you just rely on your EA to trade for you as if you don't have to do money/risk management etc.  So if you are thinking of earning some passive income for your old age, you come to the wrong place.  Developing EA is never about passive income.  It's about challenging yourself to make better and better EA.  If you want to enjoy passive income, go and become an investor and rely on trade signals such as Zulu Trade, whereby you don't have to think but just follow and passively collect your money.

Thanks Footon for the clarification.

I use lots of D1 indicators for my M1 EA too, didn't have issues on my MT4 or my brokers, these EA are trading very well.

Hi sorry, my computer is fully utilised with many MT4 terminals and hence, it's not a fair comparison to do speed test.

Mine is a 16 core xeon dual machine, it does really speed up a lot of my work.

I'm planning to buy another 40 core machine.

As far as trading plan or "theory", mine is pretty simple:  "Win more than I lose..."  And I suspect I am not alone in that regard.  Or maybe I am -- I don't know.  I would be curious to know what others' trading plans are.  Hannah -- could you articulate your plan and theory?  I don't understand why it has to be so complicated.

I would disagree.  I mean, it's great to have that additional understanding, but one certainly doesn't need to study the market over time in order to effectively use FSB.  The market and its data is constantly changing -- and that's about all you really need to understand.  You can study it all you want, but unless you have a crystal ball you still don't know the OHLC of that next bar. 

Since we can't tell the future, then the best we can do is react "after the fact".  FSB provides very effective tools that allow you to quickly and easily adapt your strategies to the changing data.

Dear All

This is a software that will help you Rename your files in a folder. 

BULK RENAMING UTILITY http://www.bulkrenameutility.co.uk/Download.php

This is extremely helpful for those who intended to use the Portfolio Maker and since Portfolio Maker has certain file naming requirement that may be very different from your previous naming method, you can therefore use this software to rename all your files in a folder in Bulk.

This software can help you remove, replace, add (in another words "edit" your file names in Bulk).

As I've named my EA in the past with very different system/methods, I found myself ended up having to redo all the file names all over again which is very time consuming.  This software is a time saver. 

Note:

Suggestions:

1. Highlight the list of files (within your folder) first so that you can see the "preview" of New Names and the "Rename" button will become accessible for you to execute the function.

2. Once you have highlight/Selected those files, you can see the effects of the changes you are going to make and therefore, amend accordingly till you achieved your desired goals/outcome and once you have achieve your desired changes, press the "Rename" button for the changes to be executed.

Donchian Channel is excellent for breakouts

What's the implication of Curve Fitting in our EA Selection?

I sometimes test out FSB's optimisation function whenever there is any new update to see whether the software "improved" after each update and in the past, the optimisation results would be worst off compared to the original good selected EA.  This time round, with the latest update (Oct 2017  version), my batch of optimized EA has better results than the original good EA.  That's really a good sign for me.  As I'm desperately looking for ways to "fine tune" my EA to speed up my work.

I noticed that if I were to optimized my good EA, sometimes the very last one in the list of collection is the "best" (often in the past).  It's closest to my original setting.  However, recently I noticed that the optimization function are yielding better results than before.

For illustration.  Let's say the optimized process produced 30 EA in the collection.

Ranked 1 has the best results and the 30th has the lowest results but closest to my original EA setting.  Based on the Optimized EA's demo results, there seem to be a general pattern, from the bottom of the list of EA, the results of EA gets better and better as it moves up the rank and then started to decline in the results till the 1st EA may sometimes have the worst results. 

It's liken a bell curve distribution (starting with the EA ranked in the bottom of the list) whereby the peak performing EA is in the middle of the list (generally speaking, not a uniform distribution pattern) and as one goes closer to the best ranking EA, the demo results decline.

What's the likelihood of the implication.

I believe the answer lies in curve fitting.  As we often lament about curve fitting, we also need to understand the implications.

Hence, from my observation, choosing the cream of crop EA (ie. Top best EA) from a collection may not necessarily be the best practice.  Reason being, these top EA may very well be those that are so curve fitting to the data.  Thus, if anyone is trying to fine-tune and improve their existing good EA, maybe it's good to base our selection from bottom up.  However, if the output is only 50 EA of less.  I suggest test them all and see what patterns (in their performance results) have you observed that will serve as a better guide in your future EA selection. 

Note: Such observation may not always be true but a general observation instead.  However, if one understand the logic of curve fitting, the above explanation make very much sense.

Secondly, I usually only select 5% or less optimization (instead of anything higher).  Reason being, I don't want to "alter" my EA's (good trading framework) Logical Structure too much and thus prevent the software from "over optimized" the EA to curve fit the data.  Using minimal % would then serves to "fine tune" my EA and not change my EA's structure till "spoil my trading framework" and thus reduce the effects of curve fitting in the process.