Dump it Here

[Captain_Chaza]

Thanks @Skate for your posts on the AI analyses.

Is it just me or do those "analyses" read like a description of the system code. There's no real analysis for me there.

If I asked the AI to analyze a cat, I'd expect the result to conclude that the cat's four legs make it more stable than if it had less.

Is it possible to ask the AI for ideas to improve the systems? eg in high volatile conditions, in unsuitable market conditions (ie down trend)?

If I can do it ( SOMETIMES)
Any robot/A.I. may be able do it (All Times )

Keep the good work up Captain Skates
Seems like I am your Biggest Fan
PS I cat may even come up looking like "a Bitch"

Who knows? lol!

Salute and Gods' Speed

 

[Skate]

Thanks @Skate for your posts on the AI analyses.

1. Is it just me or do those "analyses" read like a description of the system code. There's no real analysis for me there.

2. If I asked the AI to analyze a cat, I'd expect the result to conclude that the cat's four legs make it more stable than if it had less.

Is it possible to ask the AI for ideas to improve the systems? eg in high volatile conditions, in unsuitable market conditions (ie down trend)?

@peter2, you're spot on. Regardless of the AI model used, it can only interpret information into readable English as far as coding goes. At times, Amibroker AFL coding can be confusing, and those who aren't coders may struggle to make sense of it.

Regarding your first point
I understand your concern. The AI analyses can sometimes read like a description of the system code, lacking sufficient insights or analysis for some users. However, it's important to note that these analyses are generated based on the data provided to the AI. While they may not offer the same level of depth or insight as a human analyst, they can still be valuable in providing a different perspective on the data.

On your second point
It's absolutely possible to ask the AI to explain a cat, and it does the same for a coder. Ask AI to improve a cat and the results will be disappointing and it's the same when it comes to coding it responds with. However, where AI excels is in generating ideas to improve trading systems. It can suggest ways to enhance the stability of a system in highly volatile conditions or propose strategies for trading in unsuitable market conditions, such as a downtrend. While the AI's suggestions may not always be practical or feasible, they can serve as a starting point for further exploration and analysis.

Please keep in mind
While AI can be a useful tool for generating ideas and exploring different scenarios, it's essential to carefully evaluate and consider any suggestions it makes. The AI's suggestions are based on the data it has been trained on, and may not always be applicable in real-world scenarios. Ultimately, your own judgment and expertise should guide your decision-making process.

Skate.

 

[Skate]

If I can do it ( SOMETIMES)
Any robot/A.I. may be able do it (All Times )

@Captain_Chaza, it might not be as easy as you suggest. The current state of AI technology, specifically in the realm of Amibroker coding, is still in its early stages and has a long way to go before it can produce reliable, accurate coding results.

Mistakes are plentiful
One of the main issues I've encountered with AI-generated code is that it often contains mistakes and inaccuracies. For example, it might use non-existent function names or make up new ones on the fly, which can lead to errors and confusion. Additionally, AI models don't always grasp array processing well, which can result in suboptimal inaccurate code.

AI has no concept of fact or truth
Another problem with AI language models is that they don't have a concept of "truth" or "fact" and apologise when an error is brought to their notice. This means that they can produce code that looks okay on the surface but lacks depth and accuracy. While AI can be a fun tool to play with, it's important to remember that it's not yet a replacement for human expertise and judgment.

In summary
While I acknowledge the potential of AI in trading, I believe it's crucial to be aware of its current limitations and to approach its use with a critical mindset. My recent series of posts aimed to shed light on a novel aspect of trading that others may have overlooked. Remember, AI can provide answers to any question, but the accuracy of those answers may vary. It's essential to evaluate AI-generated content carefully and not rely solely on it for making trading decisions.

Skate.

 

[Telamelo]

Here’s a fun fact
Most traders don't trade with their "serious money". Serious money is always allocated to "dividend-earning investment”. Meaning, most who have serious money will select an investment strategy over a "trading strategy" every day of the week. Admittedly investing can be seen as a long-term trade but that’s where the similarity ends.

Investing allows you to ride the emotional rollercoaster
Doing so allowing "Father-time" to do all the heavy lifting whereas trading relies on timing the market with precision. Timing the markets admittedly is much harder to do. When investing, "timing of the purchase" becomes less of an issue. If you can pull it off (trading rather than investing) you will have the ability to upscale the "profit factor" by multiples.

Skate.

Tracey Edwards take on "Does stock price matter for Dividend Investing"

 

[ducati916]

@ducati916, the results of both strategies are disappointing, but not unexpected. The timing when starting a new trading strategy can impact its performance, especially over such a short time frame. The truth is, there's never a good time to start trading, as luck and timing play a significant role in the performance early on.





Skate.

Surely the whole point of a technical approach is exactly that: 'timing'.

Investors are often criticised for incurring significant drawdowns because they prefer 'time in the market' as opposed to 'timing the market'.

Mechanical trading is a subset of technical trading, it sure ain't investing.

We now have a dichotomy of: (i) a technical methodology that is (ii) subjugated to issues of timing.

Hence I return to my original question: if you are trading a long only system do you know from your testing that:

(a) it can be profitable in a bear market; and
(b) if not, then the first entry condition must be confirmation that a bull trend is in place in the wider market.

The wider market accounts for 70% of a stock's movement.
It's sector accounts for 20% of it's movement.
An individual stock's earnings/news accounts for only 10%

If you are trying to go long into a bear market, the system will struggle.

Now the explanation for 'Pressure' makes more sense as an entry criteria: an upward trending SMA (higher than it was previously).

Your market:



For a long only system....sucks balls.

At best you are looking for oversold bounces.

jog on
duc

 

[divs4ever]

' time in the market ' works when there is a realistic inflation factor

timing the market ( roughly ) works in uncertain markets and economies

now 'time in the market ' ( normally ) ignores survivor bias , but does it include 'bonuses via demrgers ??

now when i started investing ( 2011 ) it was quickly apparent that opportunistic adding to holdings ( averaging down ) was a slightly better mouse-trap than 'set and forget ' ( in most cases )

 

[Skate]

Surely the whole point of a technical approach is exactly that: 'timing'.

@ducati916, timing can play a significant role in the performance of a trading strategy, especially in the early stages. However, I believe that a well-designed technical approach can still be effective in mitigating the impact of this.

Both trading systems under the exercise of "paper trading" are pure technical systems designed to provide a statistical edge in the long run, and I believe that they can operate effectively in all market conditions. While there are no guarantees in trading, I have confidence in my system's ability to identify profitable trading opportunities and manage risk effectively, even though their methodology varies.

My trading approach utilises a combination of technical indicators to determine entry and exit points that also consider market sentiment, such as the "PercentageUp" filter, to assess the overall market trend. While the "PercentageUp" filter has recently experienced a streak of negative weeks, I believe that incorporating this factor into my strategy still gives me an edge, as it helps me make more informed trading decisions and increases my chances of success in the long run.

One thing that's worth considering is the potential impact of market volatility on your trading strategy. As you mentioned, market sentiment can be a significant factor in determining trading performance, and it's possible that increased volatility could make it more difficult to accurately assess market sentiment.

Overall, I aim to maximise my trading performance and achieve consistent profits over time, "well that's the end goal".

Skate.

 

[Skate]

' time in the market ' works when there is a realistic inflation factor
timing the market ( roughly ) works in uncertain markets and economies
now 'time in the market ' ( normally ) ignores survivor bias , but does it include 'bonuses via demrgers ??
now when i started investing ( 2011 ) it was quickly apparent that opportunistic adding to holdings ( averaging down ) was a slightly better mouse-trap than 'set and forget ' ( in most cases )

@divs4ever, "Time in the markets" is another way of saying "Investing for capital gains and dividends". This can be a good alternative to "timing the market", as it allows you to benefit from the consistent income stream provided by dividend-paying stocks.

@Telamelo has shared a great video on this topic, which you can check out that has been "posted above" for more information. If you're interested in learning more about investing for dividends, I recommend checking out Tracey's YouTube channel for some helpful tips and advice.

https://www.youtube.com/c/TraceyEdwards

Skate.

 

[ducati916]

' time in the market ' works when there is a realistic inflation factor

timing the market ( roughly ) works in uncertain markets and economies

now 'time in the market ' ( normally ) ignores survivor bias , but does it include 'bonuses via demrgers ??

now when i started investing ( 2011 ) it was quickly apparent that opportunistic adding to holdings ( averaging down ) was a slightly better mouse-trap than 'set and forget ' ( in most cases )

YES.

And since 1987 in the US that has been driven by a secular decline in rates all the way to ZIRP. Not only rates, but inflation through a variety of methodologies and secular forces has been kept artificially low:



All changed.

Time in the market is getting its teeth kicked in. Timing the market is now absolutely crucial if you are playing almost anything other than gold, oil and a few other commodities that will kick in soon (copper). You can still buy-the-dip here.

Banks are kryptonite.

We are fast approaching the final denouement. Approximate date Q2 2024 when the recession turns ugly.

jog on
duc

 

[peter2]

At times, this thread and in particular @Skate's system performances motivates me to think that I can do better than I am. I should be able to distill my trading experience to a set of rules that can be coded into a trading system.

If I can do it ( SOMETIMES)
Any robot/A.I. may be able do it (All Times )

I agree with the Capt. I have periods where I'm in sync with the market and the trading is profitable. Then there are periods where I'm out of sync with the market and frustrated. Previously out of favor stocks become in demand and I find it hard to buy them. Emotional sabotage, that what it is. I'm a captive to my psychological biases.

System trading should minimise the negative effects of our biases. Knowing this, motivates me to consider systems but I still resist.

The main reason for me not going down this path and disappearing up my own ar$e is that it's a LOT of work. There's different market conditions (MC) to describe, entries, exits based on the current MC and entwined in all of this is a robust risk management system that responds to the different MCs.

IMO skate's systems (as good as they are) are only about half way there. Restricting the universe to an index is limiting the opportunities IMO. Liquidity is more important than index participation. Market regime filters do their job reasonably but they're far from perfect. Currently the %Up filter is off and is missing the opportunities that have triggered. For a portfolio only holding 10 positions there are plenty of charts showing strong bullish trends that should still be held by the system.

@ducati916 mentions that equity price movement are 70% due to the general market, 20% sector and 10% corporate. I'd agree but it's only a general observation. Consider the recent demand for the uranium companies, 90% sector related and opposite to the current market. There are short periods of sector out-performance all the time. Systems with general market regime filters will miss these opportunities when they move up in a down market.

I know what I'm suggesting here. Lots more work for a system builder. Many stocks need their own market filter. Uranium stocks need their own uranium market filter. Gold stocks need their own market filter. This applies to most commodity companies.

Long only systems are only suitable when market are moving up. Skate's systems miss out on the benefits that the inverse ETFs provide because they're not in the systems' trading universe.

I can see the potential of using systems but the mountain of work required stops me.

 

[Skate]

Long only systems are only suitable when market are moving up. Skate's systems miss out on the benefits that the inverse ETFs provide because they're not in the systems' trading universe.

I can see the potential of using systems but the mountain of work required stops me.

@peter2, thanks for sharing your thoughts. I appreciate your perspective and the detailed insights you've shared. It's refreshing to see such honesty and transparency.

I agree that limiting the universe to an index can restrict opportunities, but liquidity is crucial in system design and trading. While buying filters like "PercentageUp" can help, they're not foolproof, and long-only systems may not be suitable for all markets, as we're seeing now. Inverse ETFs offer benefits that are missed in such systems.

I understand that building a trading system can feel overwhelming, especially when considering the amount of work required to adapt to different market conditions and minimise emotional biases. However, breaking down the process into smaller, manageable tasks can make it more manageable. By focusing on a specific area or strategy that interests you, you can gradually build out your system and develop a deeper understanding of the markets and can be a valuable learning experience that helps you become a better trader.

As you mentioned previously, even though the markets may not be conducive for trend traders at the moment, a well-designed system can still perform well when the markets change for the better. If you feel that a system would help you achieve your trading objectives and reduce emotional biases, then it may be worth investing the time and effort to create one. However, if you prefer a more discretionary approach to trading or a mix of both, then that's perfectly fine too.

Skate.

 

[ducati916]

1. At times, this thread and in particular @Skate's system performances motivates me to think that I can do better than I am. I should be able to distill my trading experience to a set of rules that can be coded into a trading system.





2. IMO skate's systems (as good as they are) are only about half way there. Restricting the universe to an index is limiting the opportunities IMO. Liquidity is more important than index participation. Market regime filters do their job reasonably but they're far from perfect. Currently the %Up filter is off and is missing the opportunities that have triggered. For a portfolio only holding 10 positions there are plenty of charts showing strong bullish trends that should still be held by the system.

3. @ducati916 mentions that equity price movement are 70% due to the general market, 20% sector and 10% corporate. I'd agree but it's only a general observation. Consider the recent demand for the uranium companies, 90% sector related and opposite to the current market. There are short periods of sector out-performance all the time. Systems with general market regime filters will miss these opportunities when they move up in a down market.

1. Converting trading experience into a set of rules, whether mechanical or discretionary is entirely possible. They are both rooted in the same error. The error is that both are subjective.

1 + 1 = 2. Unfortunately that does not apply to markets. In markets 1 + 1 can = 7 or (-4) or any other number on a whim. The rules set up for 1 + 1 = 4.5 do not work when it = (-3).

Mechanical systems lack flexibility. This is of course by design. Flexibility is presupposed as a negative trait. Discretionary systems have that flexibility. Of course they can have far too much flexibility which gets us in trouble as all manner of rules are ignored or thrown out.

2. Systems are backtested against history. Now this is a can of worms. Put aside for a moment the amount of historical data used.

History is forever being revised. Ultimately, unless you were there AND accurately knew what was going on (an impossible task) then saying that because X happened in 1929 then X will repeat in market data is just a huge error.

But mechanical and discretionary methodologies have little else to actually use.

So time periods are important. We have had this discussion on this thread before, however some historical data:



Data from 1750. Almost 300 years. Across 3 different nations. Do you seriously think that a backtest of 720 days, across a backwater like the ASX will actually disclose the macroeconomic forces driving all economies currently?

Is a bear market bounce, the same as a bull market? At the very least this should have been tested. The massive bounce in 2020: qualitatively the same as the 2023? The answer is no. They are polar opposites.



Markets do not operate in isolation. In 2020 that 20yr TLT would be very positive and positive into 2021. If you backtested stocks in their decline from 2019 into 2020, that decline was arrested due to Central Banks flooding liquidity. Stocks became a single decision outcome with real yields negative... the 'TINA' phenomena.

The short point is: testing for a mechanical system must test other markets in addition. If not, the test is flawed.


3. Agreed. The truism: there is always a bull market somewhere.

They may just not be in stocks currently. Commodities generally. Drill down. Oil is in a massive bull market. Anything energy, non-green related is bubbling away (uranium). Gold/Silver are both in a bull market. Art is also in a bull market.

This is a bear market. Can you trade long? Absolutely but it is a timing strategy.



You sit out the decline and trade the rip higher, selling at the top and sitting out again. LOL.


So what is the answer?

It has to be comfortable for you to trade. If you are fighting yourself, then no matter how great the method, it will be miserable. Your strategy must be aligned with your personality. That may in fact mean a mechanical system.

It may seem that I am down on mechanical. Incorrect. Both my strategies are 100% mechanical. My 'rules' are about 3 lines for the more complicated strategy, 1 line for the simple strategy. I avoid complexity. Complexity hides the fact that you actually gaming the process. The term is 'Optimising'.

It must manage risk effectively. If it or you fail here, it's over.

It must have a shot at being profitable. If it doesn't lose money (see above) then it has a good chance to make money.

You must align the correct strategy to the correct market. For example: commodity markets are rangebound (always). Yes the range is huge, but true nonetheless. I employ a very different strategy for commodities than for stocks. One of my two strategies is for commodities, t'other for stocks.



You must choose an effective time period to compete in. Technology has speeded everything up. Trading is conducted in micro-seconds. Unless you can compete technologically, then you need to move out. The shorter the timeframe chosen, the higher the level of competition and difficulty that will be encountered.

Daytrading is using massive leverage to snatch pennies. Get on the wrong side of a dollar move and you will feel the pain.

I have increased my timeframe to 1 year or longer depending on the strategy (one strategy trades a 1 year position, the other is open ended, no timeframe other than my lifetime). It's a snooze fest. Boring. But that is exactly what I want. A decision, buy/sell can be mulled for a day or two, sometimes a week. A bit extreme for the more hyper-active out there and will cause issues because it is not aligned with their personality.

Technicals: are subjective. They promise much but so often fail. Longer term technicals are more reliable than short term technicals. If you are going to be 100% technical...extend out as far as your personality can tolerate.

Do NOT base your strategy solely on the technicals. You must have an independent metric by which you measure buy/sell decisions.

Fundamentals work but have 2 major issues: (i) it takes a long time to actually play out and (b) are very reliant on your intellect.

Combine fundamentals with very long term charts to observe the secular forces playing out. Don't waste your time daytrading fundamentals.



Snoozefest.

I may have gone off piste somewhat.

jog on
duc

 

[Skate]

I may have gone off piste somewhat.

@ducati916 I believe you are somewhat on the right path and I commend you for being brutally honest. Your comment provided is a thoughtful and insightful analysis of the challenges and limitations of developing a trading strategy. The valid points you raise about the difficulty of relying solely on technical analysis hit too close to home as far as I'm concerned.

Mechanical systems do lack flexibility and are often inadequate for dealing with the complexities of trading. It's also refreshing when you acknowledge that discretionary systems can be prone to errors and biases and that a balance between mechanical and discretionary approaches may be necessary.

Managing risk and risk tolerance is an important point and it's the same with the time frame you elect to trade must also align with your personality is the foundation of mechanical system trading.

Overall, your comments provide a nuanced and thoughtful analysis of the challenges of developing a trading strategy while suggesting that a balanced approach may be the most effective way to navigate the complexities of the market. Even though I agree with the rationale of your comments, in the next post I'll give you my opinion from my perspective.

Skate.

 

[Skate]

Snoozefest.

I appreciate both @peter2 and @ducati916 for sharing their thoughts on trading. I appreciate their perspective, and I understand where they are coming from. However, as one who likes to have his say - I'd like to offer an alternative view from a long-only technical mechanical trend trader perspective.

Firstly, I believe that mechanical trading systems can be flexible and adaptable to changing market conditions. While it's true that some mechanical systems may struggle with unexpected events or sudden changes in market sentiment, this can be mitigated by incorporating risk management strategies and using multiple layers of confirmation in the trading process.

In my experience, a well-designed mechanical trading system can be incredibly effective in identifying and capitalising on trends in the market. By using a combination of technical indicators, I can identify high-probability trading opportunities with a high degree of accuracy even if they fail to follow through. Of course, no trading system is perfect, and there will always be some degree of risk involved. However, I believe that a mechanical approach can help to minimise emotions and biases in trading decisions, which can be a significant advantage in the long run.

While it's true that machines can't replicate human intuition and expertise, I believe that a mechanical trading system can still be incredibly effective in identifying trading opportunities. Additionally, I can use risk management strategies and a series of exit strategies to limit my exposure to potential losses, which can help to mitigate the risks associated with trading.

While it's true that some mechanical trading systems may rely on high-frequency trading strategies, this is not necessarily a requirement for all mechanical systems. In my case, I focus on weekly trends and use a lower trading frequency, which can help to reduce the impact of short-term market noise and volatility.

My trading journey since July 2015 has had its ups and downs, but my faith in my methodology has been consistently rewarded. In fact, my latest exercise of "paper trading" two new trading ideas has only strengthened my conviction. When viewing the short-term results of both strategies I'm sure others wouldn't have the same confidence as I have.

In conclusion, while I acknowledge the apprehensions regarding mechanical trading systems, I firmly believe that a well-crafted long-only technical mechanical trend trading approach can be remarkably effective in identifying and capitalising on market trends. By integrating risk management strategies, utilising multiple layers of confirmation, and adopting a weekly perspective, I firmly believe I can minimise the risks associated with trading and optimise my potential returns over time.

Skate.

 

[Skate]

It may seem that I am down on mechanical. Incorrect. Both my strategies are 100% mechanical. My 'rules' are about 3 lines for the more complicated strategy, 1 line for the simple strategy. I avoid complexity. Complexity hides the fact that you actually gaming the process. The term is 'Optimising'.

@ducati916, the "Paper Trading" exercise is all about simplicity. I wanted to show that a simple entry into a breakout with confirmation with precautions could be an effective way to trade. However, I recently discovered that the data I was using from Norgate was flawed due to old remnants left in place after my subscription downgrade. Despite this, both strategies were developed and optimised on this false premise. It remains to be seen if they will crash and burn.

The "paper trading" exercise still has value, as it demonstrates the process of getting a tradable strategy to live trading. This is a crucial step that many traders overlook. My own research shows that the "Dual Breakout Strategy" performs better, but the "Pressure Strategy" is simpler and relies on the difference between two moving averages.

In summary, the "Pressure Strategy" is a trend-following and breakout strategy that uses a moving average comparison and a momentum filter to generate trading signals. It also uses some additional filters to improve the signal quality. However, it may also have some limitations and drawbacks as I've just mentioned.

Skate.

 

[Skate]

Skate.

 

[Skate]

Skate.

 

[Skate]

The Power of a Well-Placed "Take Profit Stop"
In today's trading session, the stock price of (DYL) plummeted by 7.25% from yesterday's close, leaving many traders in a state of panic. However, by using a "Take Profit Stop" in the "Dual Breakout Strategy", I was able to successfully lock in my profits and avoid the temptation of holding onto the position for too long.





The "Take Profit Stop" is a predefined profit target that closes a trade once it reaches a certain price level and is designed to lock in profits to avoid the risk of losing those gains due to market volatility.

As the price chart below shows, the "Aqua" dashed line indicates the level at which the "Take Profit Stop" was triggered. Despite the significant drop in stock price, I was able to turn a profit.




Skate.

 

[Skate]

It's a bit weird
The phrase "If you don't poo when you first go to the toilet, you always get a second go" may not be a well-known proverb because I just made it up, but it's a weird way to express the idea of persistence and second chances. In a broader sense, it implies that if we don't succeed at something on our first attempt, there's always an opportunity to try again.

So, the next time you encounter a setback or failure, remember that it's a normal part of life and the learning process. Don't give up – keep pushing forward, and know that success may be just around the corner.

Skate.

 

[investtrader]

Couldn't really find anywhere to put this. XSO making a new 3 year low today. Nearly 30% down since Nov. 21. In case anyone was wondering why it is tough going!!!