Australian (ASX) Stock Market Forum

Joined
Mar 3, 2020
Posts
33
Reactions
35
Tobe honest in the short term it is
4pc up, then 4pc down, today is 4pc down and i would bet it will be up by tomorrow morning
Suckers buy the dip, big offload from big money, then when this atops and crash again Feds pumps or media move into a : vaccine found....
Soon overpositive news
One of the trouble of neural
networks is that learning is done on the expectations than future behaviour is an extension of the past
Great to recognise car plates or faces less for market movement where it acts when you think about it as a master Chartist.
nothing wrong with that as this is what i attempt to do in my system, but do not expect a miraculous always right answer
Just another good tool
You are totaly right!
Don't expect it gives always the right answer. Only expect it gives a better result than the human logic does.

And you are right with the trouble of neural networks deep learning considering the future as an extension of the past. But this is the base of the physic (except in certain manner in quantum mechanics where probably the time /space is different from the macro world). The limit of the learning is it possibility to integrate all the paramaters (ie the big one earthquake in SF). But in reallity all of those disaster, or good news (iie vaccin) come gradually and the market adapt itsef gradually also.
 
Joined
Jun 8, 2008
Posts
5,675
Reactions
4,871
You are totaly right!
Don't expect it gives always the right answer. Only expect it gives a better result than the human logic does.

And you are right with the trouble of neural networks deep learning considering the future as an extension of the past. But this is the base of the physic (except in certain manner in quantum mechanics where probably the time /space is different from the macro world). The limit of the learning is it possibility to integrate all the paramaters (ie the big one earthquake in SF). But in reallity all of those disaster, or good news (iie vaccin) come gradually and the market adapt itsef gradually also.
not sure I agree with the gradual side..9/11, etc: but you are right, the virus scare came out of nowhere but it took absolutely ages before seeing some reactions
some are still saying today: it is just a flu....imagine if the economic effect was the same every year because of the flu....
Good for some who covered their asses well before it became headlines be it having masks and supplies to getting out of markets into bonds and gold
;-)
Sadly I was wrong as last night in the US was another bloodshed..I was forecasting a rebound...
Maybe all the wise money is now exited (I man edge funds and the master of the universe types) and we are left with private investors, super funds and LICs to share the losses
 
Joined
Jun 8, 2008
Posts
5,675
Reactions
4,871
Please keep sharing your outcome;
I assume your system is self learning and so "automatically" weighting the most recent period higher?
Have you thought about the fact many big money companiesy may use similar toosl and so that their neural network predictions become self fulfilling?
big money predicts tomorrow is a bloodbath, put a sell all at the open
et voila: the market falls...big money prediction was right
rinse and repeat
 
Joined
Mar 3, 2020
Posts
33
Reactions
35
Please keep sharing your outcome;
I assume your system is self learning and so "automatically" weighting the most recent period higher?
Have you thought about the fact many big money companiesy may use similar toosl and so that their neural network predictions become self fulfilling?
big money predicts tomorrow is a bloodbath, put a sell all at the open
et voila: the market falls...big money prediction was right
rinse and repeat
Studies had shown that the use of automatic decision trading (which is not the same as deep learning) let the market with fewer risk of brutal down or high. I agree with that. So if everyone used a similar tool the market probably will be less chaotic. Fund managers have access to this kind of tools but dont know how to tune them. Finnaly they prefere using ultra fast reaction algos to place or quite orders.

I think the actual financial trading market is much more a Ponzi sheme which gives great rewards because a lot of persons want great rewards and are ok to buy more expensive to get great rewards... and so on until the big crackdown.

To make money, it is better to invest (buy) for a long period (years) with the hope a brutal disaster does not happen...
 
Joined
Jun 8, 2008
Posts
5,675
Reactions
4,871
Studies had shown that the use of automatic decision trading (which is not the same as deep learning) let the market with fewer risk of brutal down or high. I agree with that. So if everyone used a similar tool the market probably will be less chaotic. Fund managers have access to this kind of tools but dont know how to tune them. Finnaly they prefere using ultra fast reaction algos to place or quite orders.

I think the actual financial trading market is much more a Ponzi sheme which gives great rewards because a lot of persons want great rewards and are ok to buy more expensive to get great rewards... and so on until the big crackdown.

To make money, it is better to invest (buy) for a long period (years) with the hope a brutal disaster does not happen...
Always educate yourself to the Japanese market before this long term view.
My opinion is that market mix ponzi scheme and growth addiction due to population increases.
When this stops in the west trouble begins
But you can make money on systematic approaches
 

bluekelah

KangaBanga
Joined
Mar 25, 2013
Posts
65
Reactions
83
not sure I agree with the gradual side..9/11, etc: but you are right, the virus scare came out of nowhere but it took absolutely ages before seeing some reactions
some are still saying today: it is just a flu....imagine if the economic effect was the same every year because of the flu....
Good for some who covered their asses well before it became headlines be it having masks and supplies to getting out of markets into bonds and gold
;-)
Sadly I was wrong as last night in the US was another bloodshed..I was forecasting a rebound...
Maybe all the wise money is now exited (I man edge funds and the master of the universe types) and we are left with private investors, super funds and LICs to share the losses
Black swan event is happening mate. Started with poor PMI numbers from Japan. And I knew the virus
would spiral out of control into a pandemic right from day one due to the transmission characteristics and lowish but significant fatality rate. That's why i was asking around on how to get physical gold despite being largely a stocks person lol...

There are some drugs used for other similar viruses being trialled at the moment, but those will take a few more months to show results if any. And vaccine the quickest be a years time or so.

Very high chance china's debt bubble has started deflating and nothing much we can do about it.

Markets should bottom out in about a months time though when virus is running rampant and killing thousands in USA.

China will become Japan 2.0 after this is over. We are coming to the end of another credit supercycle. Some central banks were already buying gold heavily last year so many are ready to let the market crash and reset whilst they sit back and watch the carnage.

I believe more carnage next week for stocks as more earnings are missed and more flights cancelled etc.. Deaths outside china are stacking up as well due to largely top heavy elderly demographics of many OECD nations.

All the credit markets have been too high for too long, I forsee another GFC coming!! Warren buffet meanwhile rubbing his hands in glee whilst he waits for the bottom with his billions in cash... :D
 
Joined
Dec 28, 2013
Posts
2,319
Reactions
10,834
March 9th Dow Jones forecast: should down
View attachment 101119
@jmbonni thanks for the update..

TMT Capture combined.jpg


Request
1. Would you please make a comment about your latest chart - as the more you tell us the more we will learn
2. What is the chart indicating for you, meaning how does this chart benefit readers?
3. You interpretation of your latest chart will give readers an insight into the long term predictive value of your thread.
4. Why has the middle green dot moved from its previous position of 5th March?

Skate.
 
Last edited:
Joined
Mar 3, 2020
Posts
33
Reactions
35
@jmbonni thanks for the update..

View attachment 101121

Request
1. Would you please make a comment about your latest chart - as the more you tell us the more we will learn
2. What is the chart indicating for you, meaning how does this chart benefit readers?
3. You interpretation of your latest chart will give readers an insight into the long term predictive value of your thread.
4. Why has the middle green dot moved from its previous position of 5th March?

Skate.
1->4/ Each computed forecast is unique, even with the same data. This is due to the specific genetic like neural network evolution. If you make 2 calculations you will obtain two distinct NNs, and two forecasts, but generally showing a same trend.
More, each day has a different history.
More, each day may have a specific set of "rules".
This is why you need to train a new NN each new day. What is important is obtaining a right trend for the last known day (ie 5th to 6th March)
Dont forget, it is not a Model, just an atempt to find weak signals with a plastic logic.
 
Joined
Dec 28, 2013
Posts
2,319
Reactions
10,834
1->4/ Each computed forecast is unique, even with the same data. This is due to the specific genetic like neural network evolution. If you make 2 calculations you will obtain two distinct NNs, and two forecasts, but generally showing a same trend.
More, each day has a different history.
More, each day may have a specific set of "rules".
This is why you need to train a new NN each new day. What is important is obtaining a right trend for the last known day (ie 5th to 6th March)
Dont forget, it is not a Model, just an atempt to find weak signals with a plastic logic.
@jmbonni thanks for your further clarification that leads me to asking another question & making a few statement so you understand the position where I'm coming from. @qldfrog has knowledge & experience in this area where I have none. Your chart are intriguing & the topic of Neutral networks, Machine Learning & the results of your research may have applications of being a frontrunner, predicting market movements.

Question
Your website is relatively new & I was wondering if you have extensive experience/education in this field "or" is a new side hobby?

Trading applications
Pattern, chart patterns in particular & how they apply when it comes to trading plays a large part in how I trade. As a mechanical system trader I'm constantly looking for repeatable chart patterns that can be coded & backtested confirming the validity, profitability & if a pattern can be turned into a trading strategy.

Chart Patterns
Patterns are everywhere. Some Patterns are deadly leading us into making a poor decision. Chart patterns are easy to spot when they are in the middle of the chart but as traders we can only trade at the hard right edge. Chart patterns tell you a story of what has happened in the past & not what will happen tomorrow, that's where your research maybe of some assistance moving forward.

Stocks are just like people
Understanding traders past behaviour, may allow you to predict their future behaviour. Stocks are just like people, because that is who makes them look the way they do. The buying and selling creates an emotional ebb and flow pattern & just like people, stocks can be unpredictable, but certain emotions and behaviours do occur regularly. Humans have a unique ability to spot patterns that has been tuned by evolution & we superbly adapted to understand the visual world. Pattern recognition is for our survival as a species & its the very reason a cloud formation looks like a bunny or elvis, sometimes our mind deceives us presenting pattern when in fact there isn't any. We unconsciously take for granted how good we are we at making sense of what our eyes show us & sometimes we don't usually appreciate how tough sometimes a new patterns without explanation is difficult for our visual system to solve.

New patterns without explanation is difficult for our visual system to solve
When you post your next predictive chart may I request that an explanations accompanies the chart as I believe it will carry a higher educational value.

Skate.
 
Joined
Mar 3, 2020
Posts
33
Reactions
35
CAUTION! CAUION!
The markets are going down! More than 20% of draw down from the higher past month may lead to a financial and economic cascading crash! In this case you should protect you personal money from banks and brokers lack of liquidity...
 
Joined
Mar 3, 2020
Posts
33
Reactions
35
so this chart shows a 2000 pt range
Doubt you’ll get it wrong with that.

can’t see the point.
The computed range is important (read my previous posts). It means just the trend forecast, up or down. No more.
 
Joined
Mar 3, 2020
Posts
33
Reactions
35
@jmbonni thanks for your further clarification that leads me to asking another question & making a few statement so you understand the position where I'm coming from. @qldfrog has knowledge & experience in this area where I have none. Your chart are intriguing & the topic of Neutral networks, Machine Learning & the results of your research may have applications of being a frontrunner, predicting market movements.

Question
Your website is relatively new & I was wondering if you have extensive experience/education in this field "or" is a new side hobby?

Trading applications
Pattern, chart patterns in particular & how they apply when it comes to trading plays a large part in how I trade. As a mechanical system trader I'm constantly looking for repeatable chart patterns that can be coded & backtested confirming the validity, profitability & if a pattern can be turned into a trading strategy.

Chart Patterns
Patterns are everywhere. Some Patterns are deadly leading us into making a poor decision. Chart patterns are easy to spot when they are in the middle of the chart but as traders we can only trade at the hard right edge. Chart patterns tell you a story of what has happened in the past & not what will happen tomorrow, that's where your research maybe of some assistance moving forward.

Stocks are just like people
Understanding traders past behaviour, may allow you to predict their future behaviour. Stocks are just like people, because that is who makes them look the way they do. The buying and selling creates an emotional ebb and flow pattern & just like people, stocks can be unpredictable, but certain emotions and behaviours do occur regularly. Humans have a unique ability to spot patterns that has been tuned by evolution & we superbly adapted to understand the visual world. Pattern recognition is for our survival as a species & its the very reason a cloud formation looks like a bunny or elvis, sometimes our mind deceives us presenting pattern when in fact there isn't any. We unconsciously take for granted how good we are we at making sense of what our eyes show us & sometimes we don't usually appreciate how tough sometimes a new patterns without explanation is difficult for our visual system to solve.

New patterns without explanation is difficult for our visual system to solve
When you post your next predictive chart may I request that an explanations accompanies the chart as I believe it will carry a higher educational value.

Skate.
Forget patters and mechanics. Better macro economy for swing trading.
Economic Nobel prizes and other economists are helpless, impotent, for short/long term forecasts because they focus on mechanical interaction.
At this time the only way to have a healthy economic system is to avoid financial speculations: no short selling, no derivative products, avoiding selling stocks before one year hold...
What can do a central banks can do? printing more money, change the rate of the loan, helicoptery money, doing studies. It is not a lot, everyone can apply to manage them.
 
Joined
Mar 3, 2020
Posts
33
Reactions
35
so this chart shows a 2000 pt range
Doubt you’ll get it wrong with that.

can’t see the point.
@tech/a
Your last comment has something good. It shows us how much human beings are bad when it comes to deal with fuzzy logic domain. Subsediary it shows shows the pretentiousness of human beings, particulary when they don't admit their limits.
Don't criticize without proposing better forecasts or without justifying your criticism.
Sadly the Down Jones fallen in the same range as it was forecasted, more than 2000 points!
 

tech/a

No Ordinary Duck
Joined
Oct 14, 2004
Posts
19,442
Reactions
3,691
I think there is a very good chance that the DJIA will have an up day
Which the neural network will learn from.

10/3/2020 Should be UP!
 
Top