Auction Market Value Theory & Analytics Expert Advisor

Hi.

With the Auction Market Value Theory & Analytics EA you can trade in parallel 26 pairs.
This EA is rather unique, as it is capable to "follow the market", this means:
- no optimization is needed
- the same set of input parameters is good for all pairs
- you dont' need to change the set even if market conditions change. The same set will, in principle, be ok "forever"
This 3 unique features mean that the EA is not "manually adapted" to a specific pair in a specific timeframes, as it normally happens when you otpimize a pair in a past period with the Strategy Tester. Having a good backtest result when you use optimization is rather easy...but this doesn't mean that the EA will perform in the same way in the future, simply because market conditions can be very different!
This EA is really playing on another level...it adapts to the market variations and adapts to all the pairs without changing the input parameters!! In my long experience I have never seen such a feature in an EA, did you?
So, how is it possible that this EA "understands the market"?
The EA doesn't use "common" indicators, instead is based on a Theory, that is called Auction Market Value Theory & Analytics. So, let me tell you a bit about this.
It is well described here:https://www.forexfactory.com/showthread.php?t=482744, and is an evolution of Market Profile Theory (see here: https://www.forexfactory.com/showthr...528resources); it's based in understanding the market mood, that is in a sentence "is the present price far from the proper value that the market has in mind?"...it is not a day trader or scalping tecnique, it is rather a position technique. Normally positions are kept open for weeks.
Important to understand is the concept of Overlay in a specific Timeframe, in a few words it is an envelope of prices in the last days, and helps understanding where the value is and if the value is stable or not.
An Overlay is "Bracketing" when the price stays in a channel, almost if the price was "between brackets".
When you have Overlays Bracketing in different timeframes, it means that the price is wandering in the same channel since many days...this is a very interesting situation, as long stability in the market is often a signal for a future breakout.
Here is where my EA trades, looking for "too stable" conditions in specific bracketing patterns and waiting for them to explode! If they do, the EA uses a very smart tecnique for setting stoploss, take profit and trailing stop.

You will not find any pip to set such values in the EA, as the EA understands the best values to set them. If you run the EA, you will see that the stoploss is always hidden under a price support, the same with trailing stop.

I understand that if you are not familiar with Overlay and Market Profile, this explanation may seem a bit confusing, but I am available to try to explain a bit more, just send me a message.
Some hints:
- when testing, ONLY use M1 timeframe, it is mandatory for accuracy. If the testing is too slow, you can use "open prices only". You need at least 2 years of data with M1 and M30 timeframes.
- when running in a demo account, you can adjust the RATE parameter. That is the number of minutes for a refresh of all calculations, normally I keep it at 5. Of course you can also adjust the lot size (recommend 0.6 lot with an initial balance 10000 EUR, if balance is less just reduce the lot in proportion, e.g. balance 1000 EUR lot size 0.06).It is not necessary to change the other parameters, the EA has already a good set, just load the one in the .zip.
About my test results:
I've been using the same set that is included in the .zip, running a 15 months test (in M1, initial balance 10000) on all 26 securityes; as the EA is trading in parallel all 26 pairs, in order to simulate the behaviour you have to SUM the test results:

TOTAL Profit 21020 EUR, 100 trades OK, 34 trades NOK
% win 74.63%
av. profit 156.87
% profit 210%
DD: typically a pair has a DD below 10%, but you have to consider that normally more than on position are opened, so expect an even lower DD.
In one of the attached picture, you will find a graph that sums in time order all the test results. That is my best guess to predict future behaviour of the EA!

Some details more:
the EA looks for conditions where there is bracketing both on a "shorter" Overlay (with NUMERO_BARRE 30 mins length) and in "longer" Overlay (with MULTI*NUMERO_BARRE 30 min bars). In detail, the EA looks (e.g. buy) for conditions where the upperlayer of shorter Overlay is lower than upperlayer of longer Overlay (similar for sell). When Activitiy (see theory for definition) conditions are fulfilled (see below) the EA will open a Virtual Buy Order (it is not a market order!!) at the upperlimit of the longer Overlay(similar for sell). The virtual order opening is shown in the EXPERT tab. Then, if the price goes beyond the Virtual Order level for at least NMIN mins, a real buy order is opened (similar for sell). In a sentence, the EA looks for breakouts of the longer Overlay upper/lower limits. In the following points are described more in detail OV calculation general rules and EA behaviour.
1) the "brick" height, or step, used to calculate all Overlays is variable. Actually the step is equal to the 30 min 3 days average ATR, divided for a integer number (STEP_REF in input parameters). For instance if in a certain moment ATR is 70 points and the STEP_REF is 10, the step will be set at 7 points. You can see the present step in Expert tab, printed together with all other Activity parameters of the 3 days Overlay.
2) The Expert tab prints every RATE minutes all the parameters of Activity for securities in full bracketing condition (means both longer Overlay and shorter Overlay are in bracketing). The Activity is used to "fire" a Virtual stop, if bracketing conditions apply. The input parameter to set the Activity threshold is REF_ATC1 (max value 5), and it refers to the activity calculated comparing Overlay 3 days (last) with Overlay 3 days (GAP days ago, you can set this in input settings)
3) If you set UsaClose at true, an automatic OrderClose is sent when price is Higher than Trailing Level and Activity is higher than ACT_REF2.
4) Stop loss is automatic, no need for any input data. It will be set, for a Buy, at Lower Value Area level of last Overlay 3 days, similar for a Sell.
5) in order to consider the possible effects of volatility, you can use a larger stoploss. This also will be calculated automatically. In detail, if step (automatic calculation described above) is higher than fraz_atr, the stop loss will be moved in proportion. For instance, if in a certain moment step is 15 and fraz_atr is equal to 10, the stop loss will be 1.5 times greater. If you do not want to move the stop loss, just use a very high value for fraz_atr (for instance 100).
6) when stop loss is moved as describe in point 5), lot is reduced. For instance, if you use a LOT 1, in the example above lot will be reduced at 0.5. This is automatic and dynamic, lot size will be different for different orders depending on volatility and the input parameters you use.
7) Take Profit is calculated at fatt_tp times the stop loss width. For instance if in a certain moment stop loss is calculated at 1500 points and fatt_tp is 2.0, the Take Profit will be set at 3000 points.
8) Trailing stop starts when the price is higher than FTRAILING_STOP the Take profit level. In the example above, if FTRAILING_STOP is 0.5, trailing will start at 1500 points from price. At this same level, if you set UsaClose at true, automatic Close (activity dependent) will start. If you do not want trailing stop, jus set FTRAILING_STOP higher than 1 (e.g.2)
9) A Trailig Stop will be set at a level higher that FTRAILING_STEP only if there is an Overlay 3 days in bracketing, and will be set at the Value Area limit. That is, the stop loss wil be "protected" behind a strong support. It also means that if there is no Overlay 3 days bracketing, trailing stop will not be set. If this happens usually it means you are still in trend..that is good...but you can still send an automatic Close with UsaClose=true.
Of course you need a high RATE to monitor properly both Close and Trailing stop. In my real account I am using RATE=5.
10) Sometimes there not a nice breackout, and price just starts wandering in a lateral way for a weeks... If the price is lower than Trailing level and higher than open price, after GIORNI_ATTESA days it will be closed. I believe this is rather good, but if you do not want this just put a very high number for GIORNI_ATTESA (e.g. 1000)
other input settings:
a) ore_wait_bracketing: number of hours in wait before opening an order, after bracketing condition begins (normally set=0)
b) DIST: you can set an offset, in points, to open a virtual order higher or lower than upper/lower limit of the longer Overlay.
c) Usa_Close: is this is false, no OrderClose will be used
d)MIN_TPO, MAX_GAP: parameters for Overlay calculation, no need to change
e) RATE: every rate minutes all data are recalculated. I suggest to set it at 180 during testing (the smaller, the worse is performance) and at 5 in real operation (the higher, the better is accuracy)

Here a sample of how the EA performs using the SAME set on different securities:
http://fabiotest.ddns.net/eurusd_V53demo.htm
http://fabiotest.ddns.net/gbpusd_V53demo.htm
http://fabiotest.ddns.net/usdchf_V53demo.htm
http://fabiotest.ddns.net/usdcad_V53demo.htm
http://fabiotest.ddns.net/nzdjpy_V53demo.htm
http://fabiotest.ddns.net/audjpy_V53demo.htm
http://fabiotest.ddns.net/nzdchf.htm
http://fabiotest.ddns.net/audcad.htm
http://fabiotest.ddns.net/eurchf.htm
http://fabiotest.ddns.net/eurgbp.htm
http://fabiotest.ddns.net/gbpjpy.htm
http://fabiotest.ddns.net/nzdcad.htm
http://fabiotest.ddns.net/nzdchf.htm


Below I attach a -zip with the EA, sets, and detailed results

I also attach the overall results of the testing on 26 pairs, merged with a spreadsheet...here there is no MT4 optimization, so the behaviour in the past is a good guess of what could happen in the future, don't you agree?

cheers

I also developed an EA based on Market Profile, that I believe is much simpler approach respect to Overlay.
Information about this theory can be found here: https://www.forexfactory.com/showthread.php?t=277528

Nevertheless I believe this EA also has also good results in testing.

It uses a Custom Indicator (must be added to MQL4/Indicators).

If somebody is interested, I attach here a demo version to test it

here some test results:

http://fabiotest.ddns.net/usdjpy.htm
http://fabiotest.ddns.net/mp3_newset.htm

Here is the EA, sets and results in detail


cheers

Now,

what would it be running on the same account the 2 EAs?

For instance, you could have one single MT4 instance with 3 charts, each with:

1- Market Profile EA (EURUSD)
2- another Market Profile EA (USDJPY)
3- one Auction Market Value Theory & Analytics EA (26 pairs)

So, what can be the overall behaviour on this account?

For 1), this is the backtest: http://fabiotest.ddns.net/mp3_newset.htm
For 2), this is the backtest: http://fabiotest.ddns.net/usdjpy.htm
For 3), the overall backtest is in the image of the first message


so...I just merged all the trades and ordered since jan 1st 2017 in a spreadsheet..and this is the OVERALL result that account

It's a HUGE profit...with a minimal DD and minimal risk..don't you agree?

Thanks fabiodgt for your thread - realy looks interesting.
Questions:
1. Are there many open positions over the weekend?
2. How does it select the 26 pairs?
3. Is it possible to run the EA with only 6 pairs?

kofix11,

Thanks to you for your interest and for your questions.

Before answering, let me explain a bit more how the Auction Market Value Theory & Analytics EA works:

when conditions are fulfilled (e.g. bracketing + activity + virtual order level reached for some minutes), an order is sent to the market. Here (example: a buy order):
a) the stop loss is set "under a strong support": in detail it is set at the lower level of the value area of the 3 days Overlay. In a sentence, is a very important level in the last three days
b) the take profit distance from order price is is = (stop loss distance)*fatt_tp (for instance2, that is 2 times the stop loss). TP 2 times the stop loss sounds good to me.
c) the trailing stop level is a fraction of the take profit level (for instance: 0.5 time TP), but it only is set if the price "holds" after a trend, in order to avoid to waste a potential good trend
d) last, an open position can also be closed with an automatic orderclose, and this is done: 1) if the price level is beyond the trailing level 2) if the activity is high (infact, tha activity high, like volume or atr, usually means that the trend may change or reverse.

This EA is made in order to follow the trend after a breakout, and usually trends may last for weeks.

So, to answer to your question 1), no, I do not close open positions on friday, because I want to follow trends. I know, there can be news during the weekend, gaps...but usually big gaps happen only on huge news, like elections...In this case I would probably close manually all open positions, there is no damage to the EA in doing this.

About question 2), I selected the 26 pairs simply testing... in principle the EA works fine with every pair, but I made a selection (very wide indeed) by testing. I tested 28 and selected 26 based on test results. This means that "this EA works fine with every pair" is not 100% true, but 93% true...good enough for me

About question 3), yes, if needed I can easily customize the EA to work only with 6 pairs. of course you would have a "less active EA". I believe that with 26 pairs this EA would make about 150-200 operations in one year, with six pairs it would probably be much less. To give you more detail, below is the result of testing on pairs (profit, trades ok, trades ko) , which ones would you select?

cheers

..............profit ok nok
EURUSD 2848 5 1
USDJPY -807 2 2
GBPUSD 2075 3 0
EURGBP 1221 3 1
AUDUSD 796 8 3
AUDJPY 1510 5 2
EURNZD 0 4 2
CADJPY 982 4 3
EURAUD -56 5 4
EURCHF 1181 3 1
CADCHF -16 3 1
GBPJPY 838 3 0
GBPNZD 77 4 1
NZDCAD 955 2 0
NZDCHF 932 3 1
NZDJPY 2097 5 0
NZDUSD 649 5 3
USDCAD 2672 6 1
USDCHF 2176 5 0
GBPCAD -637 3 1
EURJPY -112 6 3
GBPCHF 154 3 1
CHFJPY 289 4 1
AUDCAD 1196 6 2

TOTAL 21020 100 OK 34 NOK
% win 74.63%
av. profit 156.87

Hi,
I have linked to this thread a Trade Explorer.

On this account 3 EA are trading:
1) Auction Market Value Theory & Analytics EA, on 26 pairs, max lot=0.1
2) Market Profile EA, on EURUSD, lot=0.1
3) Market Profile EA, on USDJPY, lot 0.1

let's see what happens in some weeks

Thanks for your answer fabiodgt.
I thought about these: USDJPY; GBPUSD; GBPJPY; EURCHF; EURUSD and USDCHF
But I see it might be boring...
Will try to understand your EAs over weekend.
Thanks anyway and have a good weekend.

You are welcome. Please do not hesitate to ask if my explanation above about the EA is not clear enough, I'll be happy to help.
cheers

Grande Fabio. Non sono fan degli EA, ma mi fa piacere vedere un collega italiano :-)
Nello specifico, ho apprezzato la premessa della non ottimizzazione. Perchè quella, poi -nel tempo- fa invalidare la bontà di un EA.
Devo solo procurarmi una VPS x testarlo.

Very, Very interesting concept, will be following this daily. Subscribed

Grazie!

I switch to english so everybody can understand!
Actually I fully agree with you, optimization is a false friend I believe...
ciao!

@fabiodgt Thank you! for the switch to english was trying to understand it hahahha, green pip for you good sir

Subbed. Will forward test.

Hi,


I made some improvements to the Market Profile EA, now I believe it has better performances.

I attach the result I got with EURUSD, that is higher profit and lower DD compared to previous release:

http://fabiotest.ddns.net/eurusd_314.htm

The set I used is optimized in the last 5 months, but the performance in the previous 12 months is also good, this is a good thing I believe.

I attach a .zip with the EA, the custom indicator and set.
If you run the EA in demo, I suggest to attach to a M30 chart, and also attach to the same chart the indicator (you will see a graphical representation of Market Profile, like in this video: http://fabiotest.ddns.net/mp.webm)

Last, I am testing this new EA also on other pairs, I would really appreciate if somebody can try it too

cheers

For your first EA, what time frame should I attach to chart or does it not matter?
Also must I attach individually to all charts, or will it look at all the charts automatically?

Regards

Hi,

the Auction Market Value Theory & Analytics EA can be attached to any chart in any timeframe, it will trade on 26 pairs anyway.

Anyway, if you use an H4 chart you will be able to see better the choices that the EA makes;

for instance, when you will see in the Expert Tab a message that a Virtual order has been opened for a certain pair, if you check that level on the correspondent chart at H4 you will see that it is an upperlimit/lowerlimit of the longer Overlay timeframe, that is an important resistance/support in the last 15 days; if that level will be retested in the next days (5), and price holds for more than 25 mins, a real order will be sent to the market.

cheers

Sorry maybe I was not clear enough.

Just attach to one chart, that is enough.

cheers

from your post#1 , it should install some indicators ?
could you tell us which indicators need install for the EA , thanks

Hi,

in this thread there are 2 different EAs:

1) Auction Market Value Theory & Analytics EA, post #1-> needs no indicators.
2) Market Profile Strategy EA, post # 15, needs a custom indicator. Such indicator is R-MarketProfile_open.ex4, it is in the .zip file, and must be copied in MQL4/Indicators. You can also attach the indicator to the same M30 chart where the EA is attached.


cheers