Martin Armstrong Discussion

[Traxo]

He wrote to me:
“I think we already had the capitulation. Do you not see the blood all over the alts? The BTC capitulation was in July. The alts capitulation was this past week. [...]”
you can be damn sure that 9/11 was a false-flag. [...]

Our banned dude asks me to relay to you guys that it seems the Western politics is sliding very rapidly into chaos. [...]

VERY IMPORTANT: our banned dude wrote in great detail that he expects the U.S. Military will likely be lured with a false-flag into attacking Russia, and Russia will be forced to destroy the USA with nuclear ICBMs. The USA ICBMs will never be fired at Russia because the global elite are in control.

He thinks governance is improving in Asia whilst it is collapsing in the West. [...]

He thinks kidnapping and crime will become more rampant in Latin America. [...] As Armstrong’s strong dollar-short vortex proceeds, then developing nations will be squeezed, but expect the Philippines to continue do well economically and ditto pretty much all the nations in Asia after mini-dip into 2020ish.

From 2021 forward, Bitcoin and Asia onwards to world domination. The West is toast. Any non-EU, non-Five Eyes white country might also be viable, such as Russia.
[...]

Armstrong believes the top  in the US hare market should come around 2028. Don't remember exactly.

Martin seemed to have 2 alternative outcomes:

1. Headed into a peak by 2020ish.
2. Bottom in 2020ish and peak as late as 2032.

Seems we already elected #1? Because I doubt Trump will be re-elected in 2020.

[1714160824]

1714160824

[1714160824]

1714160824

[1714160824]

1714160824

[spongegar]

I don't know, it's too complicated for me to understand and besides my understanding of economics in itself is still flimsy. Butni guess we can say that things has a way of turning around and around for us and no one will be the wiser. I just hope we try to understand what is going on in the past fornus to move away from a cyclic nature of things.

[sidhujag]

The Dow has broken out to the upside but it is not yet through technical resistance. Today, a closing below 26406 is not a Reversal, but it will warn that we can still see a temp high with a Direction Change due tomorrow on the daily level. A Friday closing also below last week's high of 26211 will also warn that we are still really just consolidating. The next real target will be the week of 10/01 so exceeding this week's high next week should signal a further rally to Test the January high.
As stated previously, exceeding this year's high next year will most likely signal a Vertical Market into 2020/2021. The politics will play a role, but keep in mind that everything everywhere is just going nuts - not just Trump. Capital will move according to country risk. The volatility is likely to appear for December going into year-end. All we can do is follow the computer and the Reversals. That is the only way to deal with this political-economic uncertainty on absolutely every front globally. We also have a leadership election in Japan tomorrow.

Does anyone know what closing right on that number implies for tomorrow?

in other words make a new high and then come down to 2020 and 2021 got it.

[wingsthegreat21]

Of course there will be a collapse. The FED will make it happen and then they will pocket the money and laugh all the way to their fed reserve banks you ever notice? They are never hurt by any problem in the economy ever.

[sidhujag]

Of course there will be a collapse. The FED will make it happen and then they will pocket the money and laugh all the way to their fed reserve banks you ever notice? They are never hurt by any problem in the economy ever.

never say never

[MA_talk]

I was a long time believer (10+ years) in Martin Armstrong, until recently.  Here is some public news related to his jailing:

https://www.hedgeweek.com/2009/08/25/martin-armstrong-and-two-firms-pay-usd27m-anti-fraud-action

---
In July 2004 the CFTC entered an order against Harold Ludwig, former co-director, with Martin Armstrong, of PGM, which required Ludwig to pay USD4.9m in restitution and a USD2m civil monetary penalty for his role in fraudulently allocating profitable trades to benefit himself rather than the Princeton customers. Also in July 2004, the CFTC entered an order against William Rogers and Maria Toczylowski, the former president and vice president, respectively, of the commodity futures division of Republic New York Securities. The order required them to pay USD6m and USD400,000 in restitution and USD2m and USD240,000 in civil monetary penalties, respectively, for their roles in executing net asset value letters that intentionally misrepresented the true values of the Princeton accounts and for assisting in fraudulently allocating trades to the detriment of Princeton customers.
----
The link didn't mention much about Martin Armstrong, but the co-director Harold Ludwig allocated profitable trades to himself.  I find it extremely hard to believe that Armstrong can claim innocence from this.  Everything he writes on his website related to his jailing, he claims innocence.  But that's what everyone would do.

[minhkhoa]

I've met him, and hung out with him on several occasions.He is rude, arrogant, and basically about the biggest jerk in public you could possibly imagine. Every person in Austin that has had the misfortune to run into him in public will say the same thing.

[MA_talk]

I read EVERY single public articles that he has ever written and available on the internet from year 2000 to about year 2015.  ANd I truly meant EVERY articles.  Never missed a single post.

Here is the biggest issue that I have about "his computer" or "AI" or whatever that exists.  He plays the numbers & dates to fit for his benefits.  Do you know the length of his ECM model?  He often uses 8.6 years.  In the articles from those 16 years, he used 8.6, 8.61, 8.615, 8.6153846615, and 8.6xxxxxx (depending on the day and his mood).

You would think that he is quite "scientific" as he often attacks the stupid economists without any scientific disciplines.  But when I read
https://www.armstrongeconomics.com/world-news/did-world-war-iii-start-on-the-precise-day-of-the-ecm/

I immediately realized that he was making up numbers.

Why?  Because IN HIS OWN WORDS, the date should have been Oct 7th, 2015, instead of Oct 1st, 2015.  Yet, he claimed that ECM is accurate down to the day SO MANY times.  Actually, I recalled that he claimed the bombing started on Sep 30th/Oct 1st, which is the precise day of ECM, but the current content at the above link doesn't show that anymore.  Maybe he has gone back and changed it.

Here is his older article (from the below 3 links) indicating 10/7/15 is the date, and in this article, he claimed
"Both the 1994.25 and the 1998.55 turning points also produced clear events precisely to the day. The probability of coincidence of so many targets being that precise to the day was well into the billions."
https://www.armstrongeconomics.com/writings/1999-2/the-business-cycle-and-the-future/
http://www.contrahour.com/contrahour/2006/06/martin_armstron.html
https://web.archive.org/web/20170419192904/http://www.nowandfutures.com/buscycle.htm

If someone tells you that the cycle model is accurate down to the day, that means that the cycle length must be accurate to within 1/365 = 0.3%.  His 8.6 year for ECM must be accurate to be within 0.3%.  Here is THE NUMBER that he posted numerous times at the headline section on his own website:

http://web.archive.org/web/20141015175106/http:/armstrongeconomics.com/

And the number is 8.6153846615

He has self-certified that the last ECM date was Oct 1st, 2015 on his sites many times.  And accurate past ECM dates that I can find are Feb 27th, 2007, and July 20th, 1998 are accurate past ECM dates down to the day.

Based on his calculation method shown here:
http://www.armstrongeconomics.com/armstrongeconomics101/basic-concepts/how-to-convert-ecm-dates/

I also realized one fatal mistake that he is making.  8.6 x 10 will give you 86, which means that after 10 ECM cycles.  Any decimal parts will be the same, which means that the dates in the year WILL be the same too.  If you keep adding 8.6 to 2015.75, you will only have the following decimal parts:

Decimal part in year (dates calculated using his published method above, for 365/366 days in a year):
0.15:  Feb 24th (54.75 days)  Feb 24th (54.90 days)
0.75:  Oct 1st (273.75 days)  Oct 1st (274.50 days)
0.35:  May 8th (127.75 days)  May 7th (128.10 days)
0.95:  Dec 13th (346.75 days)  Dec 13th (347.70 days)
0.55:  Jul 20th (200.75 days)  Jul 19th (201.30 days)

Well, apparent, that's not enough dates for him to fudge.  So now he adds a PI date also, which is 314 days from the ECM, besides all of the 1/8, 2/8, 3/8, 1/2, 5/8, 6/8, 7/8 dates.

Guess what.  The world is so big, that you can always find a headline EVERYDAY to fit your dates.

And then he starts to use PI, and 10PI as 31.4, or 100*PI as 314, and then he has 224=8.6*26 for political cycle, and since his 8.6 is not exact, sometimes (in the last 16 years) this 224 is 223 years.  Sometimes it's 224.  Sometimes, it comes with a decimal part.  Etc.

And then he not only has 8.6 years, but also 8.6 months, and 8.6 weeks, and 8.6 days.  ANYONE who understands the mathematical definition of fractal will know that the FRACTAL is only based on 1 single RATIO.  You CANNOT keep multiply numbers to year, month, weeks, or days.  Who says 7 days or 7 is a magical number in nature??

Given the power with all of the above numbers, and the magic of multiplying everything by day/week/monh/year, or 10/100/1000/1 million/etc., he can make ANY days to fit.

If I do science like that, I can fill the entire number line with my dates.

[sidhujag]

I read EVERY single public articles that he has ever written and available on the internet from year 2000 to about year 2015.  ANd I truly meant EVERY articles.  Never missed a single post.

Here is the biggest issue that I have about "his computer" or "AI" or whatever that exists.  He plays the numbers & dates to fit for his benefits.  Do you know the length of his ECM model?  He often uses 8.6 years.  In the articles from those 16 years, he used 8.6, 8.61, 8.615, 8.6153846615, and 8.6xxxxxx (depending on the day and his mood).

You would think that he is quite "scientific" as he often attacks the stupid economists without any scientific disciplines.  But when I read
https://www.armstrongeconomics.com/world-news/did-world-war-iii-start-on-the-precise-day-of-the-ecm/

I immediately realized that he was making up numbers.

Why?  Because IN HIS OWN WORDS, the date should have been Oct 7th, 2015, instead of Oct 1st, 2015.  Yet, he claimed that ECM is accurate down to the day SO MANY times.  Actually, I recalled that he claimed the bombing started on Sep 30th/Oct 1st, which is the precise day of ECM, but the current content at the above link doesn't show that anymore.  Maybe he has gone back and changed it.

Here is his older article (from the below 3 links) indicating 10/7/15 is the date, and in this article, he claimed
"Both the 1994.25 and the 1998.55 turning points also produced clear events precisely to the day. The probability of coincidence of so many targets being that precise to the day was well into the billions."
https://www.armstrongeconomics.com/writings/1999-2/the-business-cycle-and-the-future/
http://www.contrahour.com/contrahour/2006/06/martin_armstron.html
https://web.archive.org/web/20170419192904/http://www.nowandfutures.com/buscycle.htm

If someone tells you that the cycle model is accurate down to the day, that means that the cycle length must be accurate to within 1/365 = 0.3%.  His 8.6 year for ECM must be accurate to be within 0.3%.  Here is THE NUMBER that he posted numerous times at the headline section on his own website:

http://web.archive.org/web/20141015175106/http:/armstrongeconomics.com/

And the number is 8.6153846615

He has self-certified that the last ECM date was Oct 1st, 2015 on his sites many times.  And accurate past ECM dates that I can find are Feb 27th, 2007, and July 20th, 1998 are accurate past ECM dates down to the day.

Based on his calculation method shown here:
http://www.armstrongeconomics.com/armstrongeconomics101/basic-concepts/how-to-convert-ecm-dates/

I also realized one fatal mistake that he is making.  8.6 x 10 will give you 86, which means that after 10 ECM cycles.  Any decimal parts will be the same, which means that the dates in the year WILL be the same too.  If you keep adding 8.6 to 2015.75, you will only have the following decimal parts:

Decimal part in year (dates calculated using his published method above, for 365/366 days in a year):
0.15:  Feb 24th (54.75 days)  Feb 24th (54.90 days)
0.75:  Oct 1st (273.75 days)  Oct 1st (274.50 days)
0.35:  May 8th (127.75 days)  May 7th (128.10 days)
0.95:  Dec 13th (346.75 days)  Dec 13th (347.70 days)
0.55:  Jul 20th (200.75 days)  Jul 19th (201.30 days)

Well, apparent, that's not enough dates for him to fudge.  So now he adds a PI date also, which is 314 days from the ECM, besides all of the 1/8, 2/8, 3/8, 1/2, 5/8, 6/8, 7/8 dates.

Guess what.  The world is so big, that you can always find a headline EVERYDAY to fit your dates.

And then he starts to use PI, and 10PI as 31.4, or 100*PI as 314, and then he has 224=8.6*26 for political cycle, and since his 8.6 is not exact, sometimes (in the last 16 years) this 224 is 223 years.  Sometimes it's 224.  Sometimes, it comes with a decimal part.  Etc.

And then he not only has 8.6 years, but also 8.6 months, and 8.6 weeks, and 8.6 days.  ANYONE who understands the mathematical definition of fractal will know that the FRACTAL is only based on 1 single RATIO.  You CANNOT keep multiply numbers to year, month, weeks, or days.  Who says 7 days or 7 is a magical number in nature??

Given the power with all of the above numbers, and the magic of multiplying everything by day/week/monh/year, or 10/100/1000/1 million/etc., he can make ANY days to fit.

If I do science like that, I can fill the entire number line with my dates.

i debunked.him long ago after following him for a week. Basically I do use his analysis for macro fundamentals but pay no attention to single days. Macro turning points are to be taken with a grain of salt. I def wouldnt use him to day trade thats throwing money away or better off doing random walk with some money managment which will get you farther.

[Refcrypto]

I would like to read your ideas, your solution. Because this is a matter of debate.

[MA_talk]

I followed him for 16+ years, and I know all the ins & outs of his arguments.  First of all, we need to recognize that he is a good trader, and a possibly a good programmer, and smart.  So I would give him credits on his good calls due to him being possibly a good trader, but NOT due to his computer, because that most likely doesn't exist.  It is most likely a BIG LIE.

WHY?  He stated that he was able to achieve NLP or natural language processing in the 1980s or even 1970s.  He posted a photo with his PC.  You know what?  That is simply physically and computationally IMPOSSIBLE.  The computer power has increased exponentially faster throughout the decades.  Back in his days that he claimed he achieved NLP, and he also deliberately said that his daughter cannot discern whether she was conversing with a computer or a human (hinting that he passed the so-called Turing's test), it is SIMPLY IMPOSSIBLE to do such thing, because the computer will need to be several buildings big.  Now you can fit such big computer possibly into your cell phone.

He preys on people's ignorance about the ins & outs and status of artificial intelligence and speech recognition.  He claimed that he developed speech recognition many decades ago, but that is most likely impossible again.  I am an engineer who knows exactly how artificial intelligence and speech recognition are done these days.  Martin Armstrong knows nothing about neural network however.

You need to remember that he repeatedly claims that in order to forecast the whole global economics correctly, you canNOT miss any pieces of puzzles.  That is exactly right, and yet, that is also a requirement for his own models & predictions.  So if he missed a high in the stock market here, that means everything else will most likely be a miss AS WELL.

He keeps posting the pictures of traverse and longitudinal waves, and addition of the waveforms.  So if he knows the CYCLES, what the heck is the "cycle inversion".  That is one of the most puzzling parts for me to understand, until I realized that everything was just a scam.  If you know the high and low of different waves, and when you sum up the waves, you will know exactly where the new peaks and valleys are.  There should NEVER be any cycle inversion.  But he uses that cycle inversion a lot, whenever his predictions are off, and then he will say that the cycle is inverted.  What the heck?!  If you cannot even know whether it will be a high or low, what kind of tradings can you do?

The way he keeps you to come back to read his blog is this "gloom-and-doom" method.  Human will always try to look out potential hazards by default.  So he started posting plagues and wars since 2014, and then that timing was off, and then he says it's 2014 to 2017, and then he says it's 2014 to 2017, for the BEGINNING of such events.  Again, what the heck?  If he is off on one thing, that means the whole forecast cannot be correct.

His global watch is simply technical analysis at its most basics.  It requires a closing price for the given period (day/week/month/year).  It has NO prediction values, just like any other stock technical analysis.  To explain this, you can use moving average as the simplest example.  Plot moving average curves along the actual prices, and then it will give you the fantasy of predictive values.  However, if the stock price drops by 25% due to some earning event, the moving average curves will never tell you that.  However, if you plot more beyond that big drop, on the curve, it would give you the imaginary predictive power that the moving average curves track the stock prices.

And when he said that his computer continue to recognize new patterns forming in the prices.  I almost feel like screaming.  If you examine the actual number of the pattern#, it's already into 6 digits if I remember correctly.  Finding new patterns seem to be good, but it's NOT good for trading.  It's really easy to do this piece of software.  Simply continue to record new discovery of patterns of up/down/down/up/etc.  (or by whatever percentage of up/dn).  The PROBLEM is that if you are in pattern #1000, and supposedly given this pattern, you should be earning a 10% profit on your trade, what if this pattern #1000 morphs into pattern #1001, and in pattern #1001, you will actually lose 50% on your trade.

Guess what, he will tell you that his AI computer just found a new pattern of #1001 after the stock drops 50%, and labels the newest pattern as #1001.  Congratulation!  All of his patterns don't come with a probability which he fervently doesn't believe in, because he believes that he can predict everything.  Without a probability, a trader cannot size his trades at all.  Of course, the fact that you lose money has nothing to do with him.

The biggest problem with him continuously making market calls, but without doing the book-keeping of every call or trade is that there will ALWAYS be some people who interpret his posts in the lucky ways and end up earning money, and there will ALWAYS be people losing money based on his wrong calls.  If he is as good as he claims, all he needs to do is to make 20 trading calls that earns 20% each consecutively.  And that can silence all critics.  This is true with ALL stock market newsletters.  They will keep making calls.  The good newsletters or blogs will have maybe 60% hits.  For the 40% miss, they just need to manage it by either shuffling the losing trades off-table, or they can make SO MANY calls that you cannot either keep track of every calls, or that the pages get rolled off, and you don't see them.

So he has been posting on big earthquakes, and plagues, and vertical markets (when big inflation hits), etc.  This is scare tactics for you to come back to read, and also showing off his knowledge, especially through history.

For the first 16 years, I thought this man is really a genius.  Now, I realized he is probably a self-deluded con artist.  If he cannot tell the difference between 8.6 and 8.61 and the significance of it, there is ABSOLUTELY no way for him to predict ANYTHING "down-to-the-day".

I found another forum, where one of the people actually went to his seminar and commented:
https://boards.straightdope.com/sdmb/showthread.php?t=694414

I'm so glad that I never spent the money on seminar, but unfortunately, I was cheated out of several expensive reports which had a few pages of ambiguous predictions, plus several hundred pages of history.  I concur what that reader said about the Armstrong's reports.

[sidhujag]

I read EVERY single public articles that he has ever written and available on the internet from year 2000 to about year 2015.  ANd I truly meant EVERY articles.  Never missed a single post.

Here is the biggest issue that I have about "his computer" or "AI" or whatever that exists.  He plays the numbers & dates to fit for his benefits.  Do you know the length of his ECM model?  He often uses 8.6 years.  In the articles from those 16 years, he used 8.6, 8.61, 8.615, 8.6153846615, and 8.6xxxxxx (depending on the day and his mood).

You would think that he is quite "scientific" as he often attacks the stupid economists without any scientific disciplines.  But when I read
https://www.armstrongeconomics.com/world-news/did-world-war-iii-start-on-the-precise-day-of-the-ecm/

I immediately realized that he was making up numbers.

Why?  Because IN HIS OWN WORDS, the date should have been Oct 7th, 2015, instead of Oct 1st, 2015.  Yet, he claimed that ECM is accurate down to the day SO MANY times.  Actually, I recalled that he claimed the bombing started on Sep 30th/Oct 1st, which is the precise day of ECM, but the current content at the above link doesn't show that anymore.  Maybe he has gone back and changed it.

Here is his older article (from the below 3 links) indicating 10/7/15 is the date, and in this article, he claimed
"Both the 1994.25 and the 1998.55 turning points also produced clear events precisely to the day. The probability of coincidence of so many targets being that precise to the day was well into the billions."
https://www.armstrongeconomics.com/writings/1999-2/the-business-cycle-and-the-future/
http://www.contrahour.com/contrahour/2006/06/martin_armstron.html
https://web.archive.org/web/20170419192904/http://www.nowandfutures.com/buscycle.htm

If someone tells you that the cycle model is accurate down to the day, that means that the cycle length must be accurate to within 1/365 = 0.3%.  His 8.6 year for ECM must be accurate to be within 0.3%.  Here is THE NUMBER that he posted numerous times at the headline section on his own website:

http://web.archive.org/web/20141015175106/http:/armstrongeconomics.com/

And the number is 8.6153846615

He has self-certified that the last ECM date was Oct 1st, 2015 on his sites many times.  And accurate past ECM dates that I can find are Feb 27th, 2007, and July 20th, 1998 are accurate past ECM dates down to the day.

Based on his calculation method shown here:
http://www.armstrongeconomics.com/armstrongeconomics101/basic-concepts/how-to-convert-ecm-dates/

I also realized one fatal mistake that he is making.  8.6 x 10 will give you 86, which means that after 10 ECM cycles.  Any decimal parts will be the same, which means that the dates in the year WILL be the same too.  If you keep adding 8.6 to 2015.75, you will only have the following decimal parts:

Decimal part in year (dates calculated using his published method above, for 365/366 days in a year):
0.15:  Feb 24th (54.75 days)  Feb 24th (54.90 days)
0.75:  Oct 1st (273.75 days)  Oct 1st (274.50 days)
0.35:  May 8th (127.75 days)  May 7th (128.10 days)
0.95:  Dec 13th (346.75 days)  Dec 13th (347.70 days)
0.55:  Jul 20th (200.75 days)  Jul 19th (201.30 days)

Well, apparent, that's not enough dates for him to fudge.  So now he adds a PI date also, which is 314 days from the ECM, besides all of the 1/8, 2/8, 3/8, 1/2, 5/8, 6/8, 7/8 dates.

Guess what.  The world is so big, that you can always find a headline EVERYDAY to fit your dates.

And then he starts to use PI, and 10PI as 31.4, or 100*PI as 314, and then he has 224=8.6*26 for political cycle, and since his 8.6 is not exact, sometimes (in the last 16 years) this 224 is 223 years.  Sometimes it's 224.  Sometimes, it comes with a decimal part.  Etc.

And then he not only has 8.6 years, but also 8.6 months, and 8.6 weeks, and 8.6 days.  ANYONE who understands the mathematical definition of fractal will know that the FRACTAL is only based on 1 single RATIO.  You CANNOT keep multiply numbers to year, month, weeks, or days.  Who says 7 days or 7 is a magical number in nature??

Given the power with all of the above numbers, and the magic of multiplying everything by day/week/monh/year, or 10/100/1000/1 million/etc., he can make ANY days to fit.

If I do science like that, I can fill the entire number line with my dates.

i debunked.him long ago after following him for a week. Basically I do use his analysis for macro fundamentals but pay no attention to single days. Macro turning points are to be taken with a grain of salt. I def wouldnt use him to day trade thats throwing money away or better off doing random walk with some money managment which will get you farther.

Not sure if its fair to say it's possible to debunk someone based off a week's worth. But by daytrading method, I was referring to a particular, and rather strange method a user derived. Taking the numbers Socrates gave on ask-socrates for the index of multiple composites every day and then correlating them in real time. A little bit complicated, and it was never intended to be used that way, but in usage it is quite simple to use. If the Dow, NASDAQ, or S&P 500 hits a support or resistance at the same time one other one does as well (as long as it isn't in a 'cluster'), then that would be a trade entry signal to go long or short, respectively. I saw the charts for September some time ago I'll try to find the link later.

sounds weird I know. Im pretty good at those things.. pattern recognition is easy to me especially data and numbers.Obviously his calls short term were off so I stuck to his macro analysis instead. Its pretty easy for our brains to find meaning in noise. i used gann angles and s&r to find infliction points seems pretty spot on to the pip but as always its eye of the beholder. I suck at money management I bet I can make others rich sharing my knowledge in trading but found I dont have the patience id rather code and build cool things.

[Thekees]

I followed him for 16+ years, and I know all the ins & outs of his arguments.  First of all, we need to recognize that he is a good trader, and a possibly a good programmer, and smart.  So I would give him credits on his good calls due to him being possibly a good trader, but NOT due to his computer, because that most likely doesn't exist.  It is most likely a BIG LIE.

WHY?  He stated that he was able to achieve NLP or natural language processing in the 1980s or even 1970s.  He posted a photo with his PC.  You know what?  That is simply physically and computationally IMPOSSIBLE.  The computer power has increased exponentially faster throughout the decades.  Back in his days that he claimed he achieved NLP, and he also deliberately said that his daughter cannot discern whether she was conversing with a computer or a human (hinting that he passed the so-called Turing's test), it is SIMPLY IMPOSSIBLE to do such thing, because the computer will need to be several buildings big.  Now you can fit such big computer possibly into your cell phone.

He preys on people's ignorance about the ins & outs and status of artificial intelligence and speech recognition.  He claimed that he developed speech recognition many decades ago, but that is most likely impossible again.  I am an engineer who knows exactly how artificial intelligence and speech recognition are done these days.  Martin Armstrong knows nothing about neural network however.

You need to remember that he repeatedly claims that in order to forecast the whole global economics correctly, you canNOT miss any pieces of puzzles.  That is exactly right, and yet, that is also a requirement for his own models & predictions.  So if he missed a high in the stock market here, that means everything else will most likely be a miss AS WELL.

He keeps posting the pictures of traverse and longitudinal waves, and addition of the waveforms.  So if he knows the CYCLES, what the heck is the "cycle inversion".  That is one of the most puzzling parts for me to understand, until I realized that everything was just a scam.  If you know the high and low of different waves, and when you sum up the waves, you will know exactly where the new peaks and valleys are.  There should NEVER be any cycle inversion.  But he uses that cycle inversion a lot, whenever his predictions are off, and then he will say that the cycle is inverted.  What the heck?!  If you cannot even know whether it will be a high or low, what kind of tradings can you do?

The way he keeps you to come back to read his blog is this "gloom-and-doom" method.  Human will always try to look out potential hazards by default.  So he started posting plagues and wars since 2014, and then that timing was off, and then he says it's 2014 to 2017, and then he says it's 2014 to 2017, for the BEGINNING of such events.  Again, what the heck?  If he is off on one thing, that means the whole forecast cannot be correct.

His global watch is simply technical analysis at its most basics.  It requires a closing price for the given period (day/week/month/year).  It has NO prediction values, just like any other stock technical analysis.  To explain this, you can use moving average as the simplest example.  Plot moving average curves along the actual prices, and then it will give you the fantasy of predictive values.  However, if the stock price drops by 25% due to some earning event, the moving average curves will never tell you that.  However, if you plot more beyond that big drop, on the curve, it would give you the imaginary predictive power that the moving average curves track the stock prices.

And when he said that his computer continue to recognize new patterns forming in the prices.  I almost feel like screaming.  If you examine the actual number of the pattern#, it's already into 6 digits if I remember correctly.  Finding new patterns seem to be good, but it's NOT good for trading.  It's really easy to do this piece of software.  Simply continue to record new discovery of patterns of up/down/down/up/etc.  (or by whatever percentage of up/dn).  The PROBLEM is that if you are in pattern #1000, and supposedly given this pattern, you should be earning a 10% profit on your trade, what if this pattern #1000 morphs into pattern #1001, and in pattern #1001, you will actually lose 50% on your trade.

Guess what, he will tell you that his AI computer just found a new pattern of #1001 after the stock drops 50%, and labels the newest pattern as #1001.  Congratulation!  All of his patterns don't come with a probability which he fervently doesn't believe in, because he believes that he can predict everything.  Without a probability, a trader cannot size his trades at all.  Of course, the fact that you lose money has nothing to do with him.

The biggest problem with him continuously making market calls, but without doing the book-keeping of every call or trade is that there will ALWAYS be some people who interpret his posts in the lucky ways and end up earning money, and there will ALWAYS be people losing money based on his wrong calls.  If he is as good as he claims, all he needs to do is to make 20 trading calls that earns 20% each consecutively.  And that can silence all critics.  This is true with ALL stock market newsletters.  They will keep making calls.  The good newsletters or blogs will have maybe 60% hits.  For the 40% miss, they just need to manage it by either shuffling the losing trades off-table, or they can make SO MANY calls that you cannot either keep track of every calls, or that the pages get rolled off, and you don't see them.

So he has been posting on big earthquakes, and plagues, and vertical markets (when big inflation hits), etc.  This is scare tactics for you to come back to read, and also showing off his knowledge, especially through history.

For the first 16 years, I thought this man is really a genius.  Now, I realized he is probably a self-deluded con artist.  If he cannot tell the difference between 8.6 and 8.61 and the significance of it, there is ABSOLUTELY no way for him to predict ANYTHING "down-to-the-day".

I found another forum, where one of the people actually went to his seminar and commented:
https://boards.straightdope.com/sdmb/showthread.php?t=694414

I'm so glad that I never spent the money on seminar, but unfortunately, I was cheated out of several expensive reports which had a few pages of ambiguous predictions, plus several hundred pages of history.  I concur what that reader said about the Armstrong's reports.

I also don´t understand the Cycle inversion part. If there is a high in the array and price and the next turning point is also a high in price then you are in a cycle inversion, what good does that do anyone, it hindsight? Maybe the reversals have something to do with this, I am not sure. The part about him building the first AI is also something that I never really understood, is he really SOOOOO smart that he, and not any of the other programming geniuses, built it? And no one has been able to replicate it? Its a bit to much like a super hero story. I also have an issue with fitting data to the events as mentioned before. When do you use Pi or Pi/2 or 8.6 days, weeks, months or 8.6/2 etc. It´s just like all the other fib and technical analysis tools, check out Harmonic trading, its also all fib ratios combined to find turning points.

That being said, The reversals do seem to work well, even if just as support and resistance. I have kept elected reversal on my trading screen and it is eerie how prices often EXACTLY bounces of them.
I also find it persuading that he has been right on the mark using his reversals for most things he posts on. I can remember the EURUSD riding right up to the 1.6 after it elected the yearly in 2015, he predicted this way in advance. I mean it went up penetrated it by a few frigging pips and then dropped lower and bounced off the bearish yearly at 1.0365. The reversals seem to really be important levels for price. He was right about Oil when he said on Jan 2016 it would closing above 60$ before end of 2017. Price spent 10 months below the 2016 close only to rally in the last 2 months to teach 60$. With the dow he has been right, gold to. Its sometimes just hard to follow because he never gives all the info, its also just bits for the next few days or weeks, we never get all the info.

My point is there is definitely something to him and socrates. The mistakes I have made with regards to the arrays and reversals I am guessing are due to 1) lack of understanding of them and 2) Just a lack of information. We don´t have all the arrays or all the reversals so we are working with incomplete info.

I am definitely going to try the pro trader version when it finally comes out, if it works it will easily pay for its monthly fee and if it doesen´t well, its a very good asymmetric deal. If it does not work a few hundred dollars down the drain if it does work....

I do agree that his reports are ridiculously expensive, I honestly don´t get why they have to priced so high.

[MA_talk]

I got answers about his arrays before.  The bars ACTUALLY do mean the price levels, as he emailed me.  And I forgot where he said this, but he said that the price levels are relative only in THAT timeframe that is being plotted.  So you cannot compare the price levels between two different array plots.

Now, you can pull up ANY arrays from ANY of his posts.  Then you can tell me that whether he even gets 50% of the relative price levels CORRECT in any forecast arrays.  I checked some, and they were just WRONG.  And of course, he won't publish that.  I just happened to get that answer from him.

The reversals would work just like any other technical analysis tools.  Anyone can identify the support and resistance levels.  You don't need anyone else to tell you that.  It's just in any charts, and they do kind of work to a certain extent.  And not everyone would agree on the exact price numbers, but they would all kind of cluster around the actual support and resistance.

Again, I'm not saying that he is not a good trader.  He is probably at least top 10% of the traders.  BUT going from there and then to claim that he has an AI computer predicting the whole world for the long term, is simply ridiculous.

And I also agree that he has VERY GOOD insight to economics, which obviously deals with the very long term forecast.  But again, I don't need a con artist to tell me that stock markets always go up in the long term due to inflation in the long term, etc.  I know that.

The point that I'm trying to make is that when someone smart like him tells the truth 98% of the time, and tells lies just 2% of the time, it will be quite difficult for anyone to figure out the 2% untruths.  In this case, it's the ECM model, the AI computer/Socrates, etc.  Just ask yourself, or him, what is the value of the cycle length of ECM?  If such fundamental question cannot be answered within scientifically acceptable answers, such as mean/variance, then the whole ECM model is simply voodoo numerology.  According to him, it should be 8.6153846615 that he put on the headline section of his own website.  And I know for sure that value is WRONG, since it is not mathematically consistent.

Again, as I stated, it was simply PHYSICALLY impossible to have such computational power back in 1980s to achieve ANY realistic levels of natural language processing nor speech recognition, unless you spent tens of billions of dollars on the computer/network.  You would think that if he could have done any speech recognition, he would have written any published papers.  None at all.  There are decades of research on speech recognition to achieve what can be done today, which is still way far from perfect.

Anyway, he wraps the core untruths with his good trading abilities, and some automated technical analysis, and maybe some good political contacts.  That's all I am trying to say.  But then he would be just like any other good traders, and he will have hits & misses, but NOT an AI that analyzes the entire world objectively without any faults.  The way he approaches AI is the expert-system way, not the neural-network way, and that will NEVER work.

[MA_talk]

What needs to happen is to collect the actual statistics on all of his arrays.  If not, on one single market would be fine too.  If it works this month, and doesn't work next month, it's as good as nothing.  But if it's working 80% or X% of the time, I'm not sure how to trade exactly with that, since you don't know when it will not work.  And it obviously needs to work >50% of the time on a big enough samples, since I can easily get 50% probably by just throwing dice.

On his arrays, did you rank the bar height, sort them in the ascending order, and compare to the weekly closing prices?  Is the relativeness correct?

[MA_talk]

I'm not interested in convincing you either way, because I have looked at enough arrays to convince myself that his forecast arrays are basically garbage.  However, I didn't obviously collect enough statistics to say how bad his forecast arrays are.

Based on that link that you provided (the last link), and I checked the prices of DIA against it, how in the world is his forecast arrays working?

Cut & paste from Yahoo finance, 4th column is the closing price:
Sep 21, 2018   267.55   267.61   266.59   267.12   267.12   3,944,500
Sep 21, 2018   0.546 Dividend
Sep 14, 2018   262.01   262.47   261.02   261.89   261.35   3,077,700
Sep 07, 2018   259.33   260.34   258.46   259.54   259.01   3,796,400
Aug 31, 2018   259.59   260.49   258.98   259.85   259.32   3,081,800
Aug 24, 2018   257.24   258.18   256.99   257.85   257.32   3,631,500
Aug 17, 2018   255.23   257.12   255.03   256.51   255.99   3,902,500

As far as I know, you should look at the composite, the first one, which is supposedly to be the most important, and combines all the information from his different models.  So in the array, he had

Aug 17 very low
Aug 24 a bit higher
Aug 31 Jumped up a lot higher
Sep 07 Dropped quite a lot
Sep 14 Went back to the height of Aug 31
Sep 21 Dropped back to even lower than Aug 24

He has been trying to time the market to call a peak here and there for the last two or three years, but everytime he failed so far.  He would write "peak?", and then he wouldn't say that it's a peak outright, but then he leaves sufficient room in the wording such that if he happens to be right, he can go back and pinpoint at what he wrote, and say "Hey, my AI computer did it again!"

I don't know if you have been reading, but his "long-term" forecast in some 5 years ago is for Euro to collapse ALREADY by now.  And that turns out to be incorrect, but he comes out and use "reversals" to explain away that Euro has bounced up, etc.

I'm not saying that he is no good (since apparently he is a good trader).  I'm just saying that his ECM is a total fake.

[MA_talk]

Also, I really wouldn't make any excuses for different markets.  He claims that he watches global capital market & flow.  And he claims that he can get everything right, and he said that everyone else is stupid because they looked at markets in isolation.  And his "computer" watches for earthquakes, weather, etc. etc.

Well, guess what?  Then he should get ALL markets correct as well.  It's simply not good enough for him to get 80% of different markets correct, because that in theory wouldn't be sufficient, and would indicate big flaws in his computer models.

[MA_talk]

One of the tricks that I realized that he is using is to simply keep making calls, especially the very low probability event, such as earthquake, or war.  Nobody would remember that Dow went down 50 points, when he forecast that Dow should go up by 150 points.  But if he hits the BIG one, and keep calling an earthquake, until it happens, which eventually it will, then he is the HERO, and his AI computer is "impeccable".

The thing is all of those little failed forecasts are still failures.  If you actually do the statistics, the math will tell the numbers.  But we as human just remember the most memorable events, like he got the earthquake event (finally), just like you remember the stocks went down big time the first quarter of 2018.

Just look at that particular array forecast on Dow for that particular post.  How is that accurate in any way?  Again, a failure is still a failure.  And again, he is trying to hit / time that eventual decline, so that you will remember it.

We must give him some credit that he is certainly an above average trader.  But he nor his AI computer is not 100%, and in fact probably not even 80% accurate.

Let's assume that his AI computer can predict something in the ultra-short term with 100% success rate.  Then he can simply go to forex and bet using options ON the futures market, where the market allows you to leverage by up to 400X.  Since he watches worldwide markets, I bet that his AI computer can find a 1% movement in various markets.  Bet on that, and you can get 400% return daily.  So if his AI computer starts with $1000.  Next day, he should have $4000.  After 5 trading days in a week, he should have $320000.  After another weeks, his AI computer should have $1024000, or about 1 million.  And after another week, by the end of 3rd week, his AI computer should have about 1 billion.  And the fourth week, his AI computer would have about 1 trillion dollar.  The liquidity of the markets probably won't support that much trading after the 3rd week, but since he watches global markets, and forex markets are extremely deep, he probably wouldn't have that much problem.

So is he a trillionaire after so many years??  Well, most likely not.  If his AI computer is that good, he should have become a trillionaire in just 1 month.

Now you can go back and re-do my math with just 90% or 80% success rate, and a different leverage factor, and you can similarly reach the same conclusion, except that it will take longer than 1 month.

If he is that good, especially given all the automated analysis by his AI computer, then he can easily open a hedge fund auto-trading for others, and make literally billions out of it.  The fact that he is not doing that, most likely means that his AI computer (if it exists) is nowhere near that success rate, and opening such fund will give away the fact that his AI computer is not good.

Of course, these are my conjectures about his "AI computer".  But I do know that it's essentially impossible to create NLP back in 1970/80s given the computer power.  But ASK yourself, WHY in the world does he want to claim that??  WHY??

[Thekees]

I'm not interested in convincing you either way, because I have looked at enough arrays to convince myself that his forecast arrays are basically garbage.  However, I didn't obviously collect enough statistics to say how bad his forecast arrays are.

Based on that link that you provided (the last link), and I checked the prices of DIA against it, how in the world is his forecast arrays working?

Cut & paste from Yahoo finance, 4th column is the closing price:
Sep 21, 2018   267.55   267.61   266.59   267.12   267.12   3,944,500
Sep 21, 2018   0.546 Dividend
Sep 14, 2018   262.01   262.47   261.02   261.89   261.35   3,077,700
Sep 07, 2018   259.33   260.34   258.46   259.54   259.01   3,796,400
Aug 31, 2018   259.59   260.49   258.98   259.85   259.32   3,081,800
Aug 24, 2018   257.24   258.18   256.99   257.85   257.32   3,631,500
Aug 17, 2018   255.23   257.12   255.03   256.51   255.99   3,902,500

As far as I know, you should look at the composite, the first one, which is supposedly to be the most important, and combines all the information from his different models.  So in the array, he had

Aug 17 very low
Aug 24 a bit higher
Aug 31 Jumped up a lot higher
Sep 07 Dropped quite a lot
Sep 14 Went back to the height of Aug 31
Sep 21 Dropped back to even lower than Aug 24

He has been trying to time the market to call a peak here and there for the last two or three years, but everytime he failed so far.  He would write "peak?", and then he wouldn't say that it's a peak outright, but then he leaves sufficient room in the wording such that if he happens to be right, he can go back and pinpoint at what he wrote, and say "Hey, my AI computer did it again!"

I don't know if you have been reading, but his "long-term" forecast in some 5 years ago is for Euro to collapse ALREADY by now.  And that turns out to be incorrect, but he comes out and use "reversals" to explain away that Euro has bounced up, etc.

I'm not saying that he is no good (since apparently he is a good trader).  I'm just saying that his ECM is a total fake.

Remember that the arrays continue to update. He often gives us just one weekly array for 6 weeks but they are constantly changing. I have saved most of his arrays and when going through them chronologically say starting at week 20, week 24 will have a panic cycle and a high composite, then on week 23 the panic cycle could have moved to week 25 and the high bar could now not be the highest bar.
I have also seen arrays where a panic cycle was predicted 5 weeks in advance and then have a inside bar on the week of the panic cycle. Since there were no more updates it could be that the panic cycle moved to another week and I missed it or it was just cancelled out.

Socrates has been providing turning points and this constantly happens. Here are the posts of a few weeks for the GBPJPY.


--
the most important targets are, 8/20, 9/3, 9/17 and 10/22. The two main targets to pay attention to are Mon. Aug. 20, 2018 and Mon. Sep. 3, 2018. The strongest target in the Weekly array is the week of September 3rd for a turning point ahead. It does appear we have a choppy period starting 8/20 until 9/17 with each target producing the opposite direction for that 5-week period. There are Weekly Directional Change targets from 8/13 to 8/20 warning of a potential choppy swing period for these few weeks.

03-09
Basis the Array, the potential Weekly targets on the Turning Point Model, defined as highs or lows on an intraday or closing basis, are the weeks of 9/3, 9/10, 9/17, 9/24, 10/1, 10/8, 10/22, 10/29 and 11/5. Keep in mind that each target will normally produce the opposite event except during a Cycle Inversion, which is indicated when two targets produce the same type of event. Nevertheless, the primary targets are, 9/3, 9/17, 10/1, 10/22 and 11/5. The two main targets to pay attention to are Mon. Sep.  3, 2018 and Mon. Sep. 17, 2018. The strongest target in the Weekly array is the week of September 3rd for a turning point ahead. It does appear we have a choppy period starting 9/3 until 10/8 with each target producing the opposite direction for that 6-week period. We have a Weekly Directional Change target due the week of 8/27.

10-09
Basis the Array, the potential Weekly targets on the Turning Point Model, defined as highs or lows on an intraday or closing basis, are the weeks of 9/10, 9/17, 9/24, 10/1, 10/8, 10/29, 11/5. Keep in mind that each target will normally produce the opposite event except during a Cycle Inversion, which is indicated when two targets produce the same type of event. Nevertheless, the primary targets are, 9/10, 9/24, 10/8 and 11/5. The two main targets to pay attention to are Mon. Sep. 10, 2018 and Mon. Sep. 24, 2018. The strongest target in the Weekly array is the week of September 24th for a turning point ahead. It does appear we have a choppy period starting 9/10 until 10/8 with each target producing the opposite direction for that 5-week period. We also see a choppy period between 10/29 until 11/5 with each target producing the opposite direction for that 2-week period. We have a Weekly Directional Change target due the week of 8/27. This coincides with a turning point so in this case we can see at least an intraday event or a turning point based on the close.

17-09
Looking at the Array, the key Weekly targets on the Turning Point Model, defined as highs or lows on an intraday or closing basis, are the weeks of 9/10, 9/17, 9/24, 10/1, 10/8, 10/15, 10/29, 11/5. Keep in mind that each target will normally produce the opposite event except during a Cycle Inversion, which is indicated when two targets produce the same type of event. Notwithstanding, the most important targets are, 9/17, 10/1, 10/15, 11/5. The two main targets to pay attention to are Mon. Oct. 1, 2018 and Mon. Oct. 15, 2018. The strongest target in the Weekly array is the week of October 1st for a turning point ahead. It does appear we have a choppy period starting 9/3 until 10/15 with each target producing the opposite direction for that 7-week period. We also see a choppy period between 10/29 until 11/5 with each target producing the opposite direction for that 2-week period.

 
You can see that initially the most important turning points are 03-09 and 17-09 but then the week 10-09 it is 10-09 and 24-09 that are the turning points and then the week of 17-09 the turning point of 24-09 has gone and now its 01-10.
If you look at the chart 03-09 closed just bearish and then price shot up and last week just diden´t manage to close inside the bar of 10-09, which I guess makes the array wrong, but still close. So now we need to see if price will make a intraday or closing low next week.

My point is that the arrays constantly update, so using a 5 week old array may not work, it may not be accurate anymore.

[sidhujag]

I wonder if using the strongest targets exclusively will yield superior results? I also see that, even if the arrays we're wrong, Armstrong's posts update as things go even before hand, which helps avoid some potentially wrong things. I acknowledge that, even without the February volatility call, many of his forecasts were correct this year in the volatility. For example, a Direction change in one of the weeks of June. The market had been sideways for a couple of weeks up till then, so it warned that the breakout would happen that time, which did. There's also a Euro forecast which can be a 'big trade', says Armstrong. I'll try and post that one later today so we can analyze it.

the edge is you. What you are doing is essentially a random walk.. when you start to see price action around s&r and understand what herd is doing and apply some proper money management you can be successful. Most traders arent consistent with their systems. Blindly following numbers is another way to donate money.