Auction Market Value Theory & Analytics

Hi all, it's been many years since I tried AMVT Analytics. I believe AMVT really works but the proper tools are required. I stopped pursuing AMVT because the indicator I used to use for it started locking up MT4 and I couldn't find a substitute. I'm seeing screenshots of some new indicators here and I would really like to know where I can download the indicator in the screenshots below from this post and also from this post so I can try it out. Is it for MT4? Thanks.

This was something I worked to get developed to trade futures, not FX. After approaching several programmers about it, one accepted the job. Four years worth of revisions and much time and $ spent, and I am still not completed. My goal is to create an expert system. Much more study on my part is required.

I don't know what your trading goals are but if you simply gave up because you couldn't find an indicator to download online then there is really nothing to say. Rule #76

Wow, you really spent lots of time and funds on that AMVT indicator/EA. Actually, just as you did, I too spent a lots of time improving a freely downloadable indicator and got it to show AMVT analysis similar to yours and it gave excellent signals. Unfortunately, that same indicator started locking up so I abandoned the project and went onto other things. Since my last post I reviewed my code and found an Array out of range error which I corrected and no more errors have occurred since then. However, I stopped researching AMVT 4.5 years ago, so now I need to refresh my memory on all the various AMVT terms. Since you are currently deep into AMVT research can you point me to a document where all the AMVT terms and concepts are summarized? Thanks.

Page #1 of the thread
Post #1 of the thread
Downloads #1, #2, #3

Nothing I know of will give a person an abridged & quick understanding of AMVA concepts. Comprehension comes from reading and thinking hard about it as you look at your screen. It must become a mindset

I agree with you 100%. I already had a good enough understanding 4.5 years ago but all the details are no longer at my finger tips so I need to jog my memory by going over the core material again. Thanks for pointing to the essential info.

Hi, everyone. I'm reading this thread and trying to create my own excel file but i stuck on the begining. How to count attempted direction? I've tride this formula on the screenshot and i also download file from this thread called eurusd1 to comper the result and now i see that there result form the file eurusd1 was diffrent than my. Could someone explain what is attempted direction?

Very interesting, would you kindly mind sharing your excel sheet. Thank you. JP

Anybody else taking a shot on the EUR today? Sell Sell Sell!!!

Got clipped and hit my stop, and the day turned out to be a weak & false break. But I reopened the same trade at a better price after watching the RTH react to the dow, bonds, oil, and especially gold (which had almost identical internals as the Eur/Usd). By the close the brackets widened and there was range extension and completed auctions in both directions. The tick counts show participation but certainly with a lack of direction. TFF is poor. NFP on Friday will give a zap to the downside I'm expectiong, although I have a stop if I'm wrong

The conditions for shorting the Eur/Usd got worse at EOD.
I'm still in my trade but I think I'm an underdog now. I raised my stop and added to my short position because I'm stupid like that.
TFF will improve at the news release, as will volatility. TPOs Abv/Blw favor the longs

Eur/Usd set up for a responsive short.
TFF is poor, TPOs Abv/Blw favor selling, volatility stable, trade ticks and volume declining, 2 Day is balanced

10, 15, and 20 day composites are balancing and below their midpoints (bearish).
TPOs Abv/Blw favor buying and TFF iimproved, however there really is no setup. It's a pass....
The Eur/Usd moved and exceeded the entire 20 day range in a single day! If you must trade thru the ECB rate decision then I would advise using a stop-in to open your orders.

Is anyone still trading these method?

I'm building my own spreadsheet and testing the data against cisco website (https://web.archive.org/web/20110621...planation.html)

I find difference in calculating the POC when there's multiple TPO with the same max length. For example looking at the examples in the url, the meta profile of 2007 01 09 has this profile:

Trade Date 20070109
Number of Prices: 45
1 142500 1 D
2 142475 2 BD
3 142450 2 BD
4 142425 2 BD
5 142400 3 BCD
6 142375 4 BCDN
7 142350 5 BCDEN
8 142325 5 BCDEN
9 142300 5 BCDEN
10 142275 6 BCDEKN
11 142250 6 BCDEKN
12 142225 7 BCDEKMN
13 142200 7 BCDEKMN
14 142175 7 BCEKLMN
15 142150 8 BCEKLMNO
16 142125 9 BCEJKLMNO
17 142100 9 BCEJKLMNO
18 142075 9 BCEJKLMNO
19 142050 9 BCEJKLMNO
20 142025 9 BCEJKLMNO
21 142000 7 CEJKLMO
22 141975 6 CEJKLM
23 141950 6 CEJKLM
24 141925 5 EJKLM
25 141900 4 EFJL
26 141875 5 EFIJL
27 141850 5 EFIJL
28 141825 5 EFIJL
29 141800 5 EFIJL
30 141775 4 EFIJ
31 141750 5 EFHIJ
32 141725 4 EFHI
33 141700 4 FGHI
34 141675 4 FGHI
35 141650 4 FGHI
36 141625 4 FGHI
37 141600 4 FGHI
38 141575 4 FGHI
39 141550 3 FGH
40 141525 3 FGH
41 141500 3 FGH
42 141475 3 FGH
43 141450 2 GH
44 141425 1 G
45 141400 1 G

Has you can see possible POC are:

16 142125 9 BCEJKLMNO
17 142100 9 BCEJKLMNO
18 142075 9 BCEJKLMNO
19 142050 9 BCEJKLMNO
20 142025 9 BCEJKLMNO

From the definition of POC in MOM, see attached picture the POC is the longest TPO line closest to the center of the range, so in this case would be 142025, but on the cisco website the POC is 142075.

So un general what's the rule when we found multiple candidates to become POC?

As you can observe from the CISCO result, where the POC has multiple TPO's of the same length, the middle one is taken, if two are in the middle then it's up to you, which one you decide to use, just be consistent.

Hi, is there anyone who is using Yani's TPOChart2 indicator ? (www.mql5.com/en/market/product/3033?source=Unknown)
Does it work with the new builds of MT4? If not, please suggest an alternative.

After a year of work I managed to take everything to the future, developing an indicator and its respective statistics with my own research improvements.

Very impressive looking.

Seems like this vibrant thread has slowly died out. Though, I don't think I am able to ignite any motivation for other members to post, all I can do is contribute. And I really want to. Firstly, I want to clearly state that I'm no expert in this field. Of course, I'm very much aware in the concepts being discussed in this forum, though, I lack experience. I wanted to make my first post after weeks of going through the posts.

!!THE INDICATOR HAS STOPPED WORKING!!!

I didn't know about this, but yes, it has stopped working. The calculation/data is wrong. And I was thinking how everybody else is doing the analysis. Some people here are financially well to pay other developers and create this indicator. It's rather easy actually. I wanted to say to those people who have just discovered about this method one thing, It can be done by yourself.

The tools, information, articles, blogs, youtube vids and with a little bit of creativity, all this data can be recreated.

For people who are wondering how. It's easy, and I have done it in excel. Though, I am in the process and I want to share my journey using this method for trading in the future aswell.

So, firstly, here's the profile I created using VBA. (See Attachment)

The following is what I did to create the Profile:

Step 1: Get OHLC Data.
Step 2: LEARN VBA

So that was it.

No I'm kidding. Of course I want to help out.

Although, I don't want to hand you everything to a plate. I want you to put some effort yourself. But lemme clarify the things you will need to do in order to actually get the TPO's, like I have from any OHLC data.

Firstly, the code will find where the High and Low data is placed in the sheet.

This is what I have used: periodHigh = ws.Cells(i, 2).Value ' Column B (High)
periodLow = ws.Cells(i, 3).Value ' Column C (Low)
For you, the columns may vary.

Secondly, the code will then create a price ladder. Basically, It Creates an array of prices from lowestPrice to highestPrice in increments of PRICE_INCREMENT.

My code:
|----------------------------------------------------------------------|
| For i = 1 To rangeSize |
| priceArray(i) = lowestPrice + (i - 1) * PRICE_INCREMENT |
| Next i |
|----------------------------------------------------------------------|

Example: If the daily range is 1.0800–1.0850, the ladder would be [1.0800, 1.0801, ..., 1.0850]

Third, it'll assign TPO letters for each 30 minute period.

Code:
|---------------------------------------------------------------------------------|
| For j = 1 To rangeSize |
| If priceArray(j) <= periodHigh And priceArray(j) >= periodLow Then |
| ' Mark this price with the period's letter |
| tpoChars(j, periodCounter) = Mid(LETTERS_TO_USE, periodCounter, 1) |
| End If |
| Next j |
|---------------------------------------------------------------------------------|


Then comes the fun part. Actually building the profile.

Code:
|-------------------------------------------------------|
| For i = 1 To rangeSize |
| tpoString = "" |
| For j = 1 To TIME_PERIODS |
| If tpoChars(i, j) <> "" Then |
| tpoString = tpoString & tpoChars(i, j) |
| End If |
| Next j |
| destWs.Cells(i + 1, 3).Value = tpoString |
| Next i |
|-------------------------------------------------------|

Anyways, all's I'm doing is try to create a kind of "Road-Map" for people who actually want to do this themselves and work along with me. Because I'm also not done yet. I have to do sooooo much more than this. But after creating a market profile, automatically, half of the work is already done. The next half is combining different Profile, and get the "Bracketing" & "Not Bracketing" parts of the indicator. Which will be tricky, but can be done in excel automatically.

I have learnt so much from this thread FOR FREE! And I wanted to provide my little contribution.

I knew about AMVT, but I would've never understood/grasped the potential if mzvega didn't emphasize on it. Hence, I would like to thank claudia and all other members of this thread who have contributed to this little community. Especially DLJONESWORK, who also took somewhat the same initiative that I am doing. There are other members also that did a lot of work. Can't forget Bandung.

Hopefully, with my post, there will be several members posting and making this community come alive again. I have just joined the party. But I feel like, there's still so much that we can do with AMVA. Especially, RESEARCH. Feel like today's market may require us to re-think some of the fundamentals with Auction Market Theory.

So, Yes. There's still some work left undone, on my part, and on AMVA, that has to be done. But I can't do it myself. I need everyone's support. And I want everybody to get involved again. Hopefully, my message is loud and clear.

P.S. it will be amazing to talk to mzvega too! pls respond your highness!! It will be amazing to discuss with you new concepts, talk about AMVT and just in general.

CHEERS.

So, it's been about a month, and I've been very busy with coding and learning along the way. To be completely open about my situation, it has been a very tedious and effortful process. However, everything has now been done and I'm satisfied with the way things are working. To give a bit of a recap, I previously made the post when I completed the Meta-Profile script. However, I have done so much more than that. Even in the Meta-Profile script, I added a bit more sophistication to the calculations. For instance, I added a more robust profile shape symmetry calculation, which is what it sounds like. I have read that Don Jones talked something about a Chii Test or something related when he talked about his Shape Factor calculation. MzVega mentioned about it. However, his approach is somewhat mysterious and not well known. So I just tried on my own. My Shape Factor version does a bunch of TPO calculation, like how well/evenly the TPO's are spread out across the range etc and outputs an average which tells me how much the profile is balanced.

Anyways, the other task at hand, back then, was to create the Composite Profile/ODC. NGL, this has been the biggest challenge so far.

Let me explain. If you're creating Profiles (typically price*time format), then the standard procedure is to get the range and build TPO's. Now, here arises a small issue. If you're always building TPO's like this, then the whole Profile is structured in a high to low format. Meaning, it doesn't represent where the prices actually moved on the chart, rather, to build a two dimensional structure, like TPO's, we HAVE to first re format the OHLC data set to record the FREQUENCIES of prices/TPO's.

The high low sorting, proved to be an issue, mainly because the condition criteria of price closing outside the original distribution seemed to be differing when price was going up as opposed to down.

Not gonna go into too much technical depth here. But, the important point is that, this was mainly the issue for the criteria of Market Condition. The basic criteria for the Bracketing/Not-Bracketing, is not more than 2 distributions (this would've been a 1 line code and much simpler), but it's a bit more complicated. If the market is bracketing, then it doesn't really matter if there 2 or even 7 distributions, if the price closed back into the original limits of the distribution, meaning the 1st distribution.

Now, trust me, I have read Don Jones Book, aswell as articles from Cisco and MzVega's explanation on ODC's, however, I was unable to think of a way, in terms of programming, to code this criteria without chronologically sorting the distributions. Meaning, the 1st distribution, is the oldest and any new distributions are the latest one. By doing it this way, I might've gone out the boundaries of AMVT (which are strict btw), and I'm aware of this, but I also think I might have created a better version (or the original version might include this, as stated, I couldn't find specific info regarding this issue).

Now, I'm in no way saying, that Don Jones approach (who is a literal rocket scientist) is inferior to mine, or mine is better, but I do think, mine is better in the aspect of logical characterization of distributions, which has been never discussed. Hence, albeit I am happy to "invent" something new, I'm also terrified if I have made a huge blunder and I can already hear MzVega typing and schooling me on this fundamental error I might have made.

But anyways, I will continue as is.

Now, to a justify my "invention", I will give an explanation as to why I *might* think this would be a better alternative then the traditional approach and later on provide some evidence on the trades I take with my current approach to further solidify the approach (or prove it's not a good approach). In any case, it's NOT RECCOMMENDEDto follow my approach. I'm doing things my way, and will be posting the results. HOWEVER, I URGE ANYBODY NEW, TO DO THEIR OWN RESEARCH ON THE TOPICS DISCUSSED IN THIS THREAD AND EVEN MY APPROACH, AS YOUR OWN MONEY/TIME/EFFORT IS AT RISK. DON'T GET INSPIRED FROM SOME RANDOM GUY ON THE INTERNET.

Now, that some disclaimers have been made. I would like to give some explanation/justification.

NOTE!: THERE'S NO EVIDENCE POSTED (AS OF YET) THAT SUPPORTS THE CLAIMS I WILL BE MAKING, BUT THEY ARE WORTH TALKING ABOUT AS A "POTENTIAL OF NEW/BETTER APPROACH. IT IS ONLY AFTER EXTREME CRTITICIM/ASSESMENT OF THE EVIDENCE PUT FORTH, ONE CAN DECIDE THAT INDEED THIS APPROACH IS BETTER OR NOT. UNTILL THEN IT IS UPTO DEBATE, NOT A FACT.

Let's first talk about Market Condition. Market Condition, is derived from the Market Structure itself. When we say Market Condition is trending or balanced or choppy/volatile, we're implying that the market has made higher highs and lower lows (trending), market making higher highs with subsequent lower highs and repeating throughout the day within a confined range (balancing) etc. However, this implication/observation is made from a subjective assessment of the market, by literally "looking at the market". There wasn't a quantified approach towards determining Market Condition. Jones approach towards determining Market Condition is sound & logical, and yet simple. It is rooted deep within the fundamentals of AMT. If profile has more than one distributions, it's more than likely, the condition is not bracketing. But why? Why do we need to adopt this method instead of another. Firstly, the concept of distributions itself is very much an excellent concept that supports his claims. If you look at the day types, proposed by Peter, you find, the structure of the Profile, determines what type of day is to occur. The structure, is related to a gaussian distribution, or a bell curve (however, this idea was criticized in Cisco Articles). A bell curve, has three reference points, Mean and two standard deviations. The idea is, if the market deviates away from the mean, the market is suppose to be away from value or not-bracketing.

In Jones' own words: "Comparing the Overlay Demand Curve to the standard “normal,” or gaussian distribution, one can estimate that two-thirds of the trading will occur within one standard deviation of the center line. Overlays are built from observations; they come from experimental data and one does not know exactly know what the “normal” distribution parameters are. The general concept is of use to the trader; value is located around the greatest volume of trading and prices farther out, toward the upper and lower price limits, are away from value." Source: VBPT, PG#77.

So, now we conclude. If prices are in one distribution, the market is staying within it, the distribution's shape is somewhat balance "looking", then the market is said to be balancing, as most of the trading is occurring around the mean. If price starts leading value, breaks the upper limits, price has entered into a something knows as a "Price Discovery Process" and the 4 step process, purposed by PETER, starts:

"This market-generated continuum can be clearly explained using what is called the four Steps of Market Activity. It is the process markets go through to factor out inefficiencies. Inefficiencies are defined as the directional move, money flow, or distribution that propels prices sharply higher or lower within a defined amount of time.
1, Series of prices in one direction
2. Trade to a price to stop the market
3. Develop around that stopping price
4. Move to efficiency (retracement)" Source: Practical Trading Applications of Market Profile -- Tom Alexander, PG#27.

So, now to conclude this part of the discussion. The distribution of TPO's/Profile allows one to discern the condition of the market, as it relates to the fundamental concepts discussed in AMT. Another way to think about this is, if the market is outside of the "value area" the condition is not bracketing, as the market has entered a price discovery process, or the market has made an Inefficiency (vertical development), which it then, must, inevitably, find efficiency to factor it out and let price catch value.

Now that the basics are in place, here comes my own justification. The distributions are logically made from the structure of the profile. Say, a market was in one distribution, on a 5D timeframe, and then, suddenly, right at the close, the market moved up, and closed outside the distribution. Now, the market is indeed in a not-bracketing condition. Because, the market did not close where value was, rather, it closed outside of the so called "value area" or the mean, or the distribution, whatever you may call it. This condition criteria would be true in the traditional ODC approach as-well, as it seeks out the position of the close within or outside distribution.

Now, imagine, the market opened the next day, traded a while above the bracket limits for a few hours (say 5 hours for the sake of the example, as this would make enough TPO's to be considered a distribution), but then, moves back into the previous distribution. What NOW? Is the market still trending? Is the market balancing, even though, it clearly has a deformed profile shape and 2 distributions? Is the market so called, "Multi-Valued"? Indeed, it is multivalued.

"A single distribution with a well-formed bell is the structure of a single-valued market, one in which traders recognize a single value (range). If there are two or more bells somewhat superposed and overlapping, the structure is that of a multivalued market. If the bell or bells are poorly formed, the market structure is telling you that the marketplace does not understand value very well.
-------------
Multiple bells show, first of all, an unbalanced market. That might mean the market is currently trending, or it could just be moving into a bracketing mode, or it could be very volatile. If a transition from trending to bracketing or bracketing to trending is under way, the trader will know from tracking previous days’ behavior." - Source: VBPT, PG#76,77.

The concept of Multi-Value itself is very interesting, and not given much thought, but to explain my point, there still needs a bit more context to be provided.

"1. Series of prices in one direction
2. Trade to a price to stop the market
3. Develop around that stopping price
4. Move to efficiency (retracement)

The above description of how markets transition from Balance to Imbalance is a key to understanding the link between the early market profile literature of Steidlmayer and the evolution of the concept represented by Don Jones' work. Steidlmayer's newer work is very much in line with the conclusions of Don Jones' research that began in the late 1980's and that continues to this day." - Source: Practical Trading Applications of Market Profile -- Tom Alexander, PG#27.

As noted by Tom, this 4 step process is what leaded Don Jones' to in a sense, revolutionize the AMT concepts (I might be wrong with this statement). Though, the reasons he put forth for creating AMVT and ODC are different, when he discusses in his book. But, the 4 step process, still remains, the core ideology for determining Market Condition.

What I want you, the reader, to focus on, is the 3rd step. "Develop around that stopping price". Notice, how it reads "stopping price". This step itself should be enough to justify my approach. However, let me paint a clearer picture by continuing the previous example.

When the market closed above the bracket limits, it was indeed not-bracketing. However, it came back the very next day, and then we were stuck/confused as to if the market is still trending or now coming to balance. The answer I would give is Balance, even if there are 2 distributions. The main reason is price did not develop enough. If the price, say for the next day or any number of days, stayed there and started to develop and develop and develop, the market untill then would be in a state of not bracketing. However, if the market comes back to where the original trend had initially started, that implies two things. The market rejected that price zone, or the market is confused. Either way, the market is definitely not in a trend or "not-bracketing" condition. Mainly because the inefficiency has been now filtered out, the move came back to where it started, it found acceptance in the previous value area, it felt comfortable trading in that zone/distribution. In other words, it's back to square one. What is true, is that the market is now multi-valued. That second distribution, now serves, in this case, a rejection. However, both distributions, are in a sense breakout zones. If that doesn't make sense, don't worry, it's an extension of this topic.

However, one thing is for certain. The market factored out the inefficiency caused by the vertical development. It developed for some hours or whatever amount, but it didn't do it for long enough, for our 5D Timeframe, and came back, so the market came back to the original initiative zone/value area and rejected the Price Discovery Process, in a sense. This could have several meanings, depending on the context. The Flow Analysis may answer some questions. Is the market confused? Is it multi-valued? Is it truly balancing? What's the shape of the profile? Is it symmetrical? Are the distributions overlapping? What will I do if the market again enters the second distribution? What will I do, if the market breaks the 1st distribution down? Do I go short?

Some of these questions, have already been answered by the market condition, being bracketing. Some will require a close up of individual reference points and their subsequent analysis.

The important point is:Chronological Characterization of Distributions, allows for more precise calculations of Market Condition, by allowing inefficiencies to be factored out, if they are.

NOW, there isn't any Supporting EVIDENCE for any of my claims nor are there references I can point out that SPECIFICALLY talk about what I'm saying. All I can reference are materials that hint at this potential. But, the topic is certainly open to debate. Am I wrong? Am I right? This is certainly as of now, an opinion, not a FACT. Things will be much clearer when the market's feedback is involved. In other words, testing is necessary.

Although, I aim to provide a general commentary on my journey with AMVT, I also aim to include my trades, side-projects (experiments) and any thing related to AMVT, like code development.

The next topic of interest is incorporating Game Theory with AMVT. However, this requires a a lot of effort, but still would be an interesting topic nonetheless.
See you soon with the evidence and other interesting topics

Hi!

Currently, I have been working on my own little side projects, like building random forest models/non-linear regression models, data-base etc, I didn't want to put soo much time/effort into providing the "Evidence" I talked about in the previous post. I will give an example, and people can test it for themselves, as I've already provided the main concept. And though, once I start publishing my trades, it'll be much easier for anybody to understand further, but up-till then, make of this what you want.

Kindly, divert your attention to the first attachment/screenshot of mt4 indicator.



You can see the indicator saying "Not-Bracketing", and I'm against this condition. I think it's bracketing.

Let me explain.

The market on 3rd of April broke above but failed to sustain that move, what's important, is that in the process, it created two distributions.
Because of that, the market condition is saying "Not-Bracketing"

However, as we know, for a bracketing condition, we not only seek the number of distribution, but also the closing price position as well as the chronological order of distributions.

This is a very simple example, (I have had market condition being bracketing for 4 distributions aswell), but this is sufficient enough for a new-comer to understand better.

As noted in my previous statement, "failed to sustain that move, what's important, is that in the process, it created two distributions", this is somewhat a vague statement, but it relates to my consistent logic, as written in my previous post.

I said in my previous post, "The main reason is price did not develop enough." --- "Mainly because the inefficiency has been now filtered out, the move came back to where it started, it found acceptance in the previous value area, it felt comfortable trading in that zone/distribution. In other words, it's back to square one."

Now, these statements will become a lot clearer with my examples. As shown in the 1st attachment, for 10, 15 and 20 days, the market is "Not-Bracketing".

I'm saying this is "Bracketing", why? Mainly because the "False-Breakout" on April 3rd has been filtered out. It came back to where the original distribution/initiative began.

I don't want to repeat my words again, but this is the general "gist" of the concept. :Chronological Characterization of Distributions, allows for more precise calculations of Market Condition, by allowing inefficiencies to be factored out, if they are.


Let's see in my database. what was the flow activity for this, see second attachment:




As you can see, a very strong flow. The trajectory/att dir and activity are all positive with impressive numbers.
Hence, this trade would've been what people call a "high probability trade".

Oh, and I forgot the bracketing query, see the the third attachment:




Here, you can see the past condition "history" as well. The alert was also a very good signal for a potential start of a trend.

See the last attachment for results:


The trend has been active/healthy till then.

Anyways, this was a shorter version for the "evidence" I wanted to provide. Initially, in my mind, I was thinking of doing one month of data for 5-6 markets, but I have other projects I can't afford to waste time on.

Hence, make of this what you will, but this is the first step in "Proving" that the Chronologically Sorting Distributions is the next step for improving the Market Condition Criteria (at-least for me).

Happy Trading