Becoming better at X

[DeepState]

From skc in thread "Shares Outstanding vs price (darkhorse70)" within the Beginner's Lounge

1. Observe and study the market
2. Develop hypothesis to explain behaviour observed
3. Acquire new knowledge and data to check and test hypothesis
4. If hypothesis is valid, try to develop trading strategies around it
5. Repeat step 1

Everytime you go through this loop you learn something new

 

[DeepState]

From skc in thread "Shares Outstanding vs price (darkhorse70)" within the Beginner's Lounge

1. Observe and study the market
2. Develop hypothesis to explain behaviour observed
3. Acquire new knowledge and data to check and test hypothesis
4. If hypothesis is valid, try to develop trading strategies around it
5. Repeat step 1

Everytime you go through this loop you learn something new

However you have come upon this loop, it is exactly that related to expertise development and is very seriously undertaken by the most elite in any endeavor where skill and effort are related to outcome. Guys like you.

I read this and was totally rapt. It is rare. If you respond, thanks.

My first question related to high and low uncertainty environments and the application of expertise loops. Something like chess has no uncertainty. All the rules are known and never change. Then you get into financial markets where many things don't make sense for long periods of time. Feedback loops are long and mashed up with other factors whose interaction is weird and often never seen before, or only slightly similar. Worse, as soon as someone cracks the market, they get cracked as the markets figure out how they are doing it and replicate it. That's high uncertainty. You have a hard time figuring out probabilities at all and outcomes are often a...crap shoot.

You work in trading with pretty short horizons. My understanding is that this is pretty noisy stuff but your report a month or so ago of your second year in Propex (?) suggests that you are achieving high hit rate and risk managing very well. Even your trade frequency (which isn't massive for a trader) basically backsolves for very nice average return if you are betting symmetrically without elevating exposure into losses as a pairs guy. If you are martingale, then I'd come to a slightly different conclusion, but still positive of course given the record you have posted and the limited market directionality.

In my world, that's an amazing outcome. We don't trade with anything like that kind of frequency but we didn't enjoy the liquidity advantage (at the time...hehe).

Question(s): Efficient learning is really tough in this type of environment. Logically prepared strategies with strong backtest fail with alarming regularity. Stuff you think is worn out suddenly springs back to life without warning.

How do you cope with such high uncertainty when seeking to learn?

Do you find that your worldview changes a lot or only incrementally or all over the place depending...

How different is what you are doing today from, say, 6 months ago, 1 yr or 2 yrs ago?

What is your standard of proof required to change?

How do you monitor efficacy of your approach to learning?

 

[Faramir]

I read this and was totally rapt.

I am rapt with nearly everything written by nearly everyone. Thank you DeepState. I realise that you are looking for an experienced contributor but is it okay if I can add some input??

Efficient learning is really tough in this type of environment.

How about any learning for a beginner? There are some great beginner threads here. Should I spend more time everyone's posts and less time with pondering what the market is doing?

How do you cope with such high uncertainty when seeking to learn?

I made mistakes with timing and mis-reading charts. I have to forgive myself and start again.

How do you monitor efficacy of your approach to learning?

It is very early days for me. I can't answer this.

 

[DeepState]

1. Is it okay if I can add some input??

2. How about any learning for a beginner? There are some great beginner threads here. Should I spend more time everyone's posts and less time with pondering what the market is doing?

I made mistakes with timing and mis-reading charts. I have to forgive myself and start again.

3. It is very early days for me. I can't answer this.

1. Hi Faramir, thanks very much for dropping in to this conversation.

2. There is a belief in this approach that it really doesn't matter where we start from. The belief is that purposeful and targeted examination of what you/I are doing, when combined with informed ways of altering what we are doing, will lead to high expertise over time. Some of the proponents espouse the 10 thousand hours concept. It takes 10 k hours to become a really solid expert in any field (chess, violin, law..) and pretty much anyone can do it. I don't believe that. I think we all have different gifts which make us unique and our gifts may not be suitable for every endeavor. I haven't yet seen any 5 ft NBA players in the starting 5.

In relation to getting beyond your self-styled beginner approach, I have the following view. I am sure there will be a huge variety of other opinions...which makes things harder, but should make you realize there are many paths to Rome. You need to find your own.

How? Experience really is the best teacher. Put yourself at some risk. Enough to be interested, not enough to hurt you. Pick the most reasonably sounding way of managing money that you have come across and start there. See how it goes. One event is not data, it's anecdote. So you need to examine each outcome and figure out what went right or wrong (both are important) and why. The best investors read an unbelievable amount and are, yet, very targeted in what they choose to read. Check out what Munger says to read. You'll get the idea. This will take many years, but the idea is to help you understand what is actually going on and how you might react to that. In there will be your own temperament. How does that knowledge impact you as a person? Daunting? Everybody started sometime with zero experience too.

Figure out where the highest useful content per minute spent (as part of a wider picture) can be found. Do that. Maybe, for you, it involves a lot of time on the site, or maybe not. Do what works for you. Everybody is different, so almost every piece of opinion needs to be seen in that lens. Facts are facts.

You probably have noticed how incredibly supportive most people are on this site when someone genuinely seeks advice. It's a fantastic resource to help highlight issues and suggest ways forward. However, you take responsibility for what you do with it. Your trades are yours. What you get is 'just' input.

Write it down. It's a great practice to see your screw-ups in writing. It is also a great practice to see where things worked out and become a reasonably permanent part of your approach.

Forgive yourself a bit. This is bloody hard despite seeming easy as might be portrayed from time to time. If you are seriously good, you'll still make bad calls at least 40% of the time.

I'll tell you point blank that the best investors are far from the smartest. However, they are the ones who can figure things out from experience and prior knowledge/contacts faster than the rest. Some people call it 'street smart'. Get dirty. Get prepared to take some massive hits. Get up. Keep moving forward. This is what makes investment/trading so fantastic.


3. Neither can I.

 

[burglar]

... 3. Neither can I.

1.) Is it good to have a "Mission Statement"?
i.e. I want to beat the index!

2.) Does it help to have a "Narrative"?
i.e. The China Commodity Boom.

 

[Smurf1976]

Some of the proponents espouse the 10 thousand hours concept. It takes 10 k hours to become a really solid expert in any field (chess, violin, law..) and pretty much anyone can do it. I don't believe that. I think we all have different gifts which make us unique and our gifts may not be suitable for every endeavor.

I am very sure that I could become an expert on the subject of football. Not necessarily at playing it myself, that requires physical ability not just intellectual, but an expert on the subject itself. I'm sure it's within my ability to learn all the rules in detail, memorise the names of players on every team, who won each of the past 20 grand finals and so on.

I have never had even the slightest interest in the game however. It bores me beyond belief.

I am therefore almost certain to fail any attempt at becoming an expert on the subject, for the simple reason that I have absolutely no interest in it.

Same concept applies to everything I think. I've always had an interest in all things electrical, mechanical or financial and was that way from a young age. A strange combination perhaps, but that's me. So I've never found it even slightly onerous to put the time into learning about such things.

 

[nulla nulla]

I am very sure that I could become an expert on the subject of football...I'm sure it's within my ability to learn all the rules in detail, memorise the names of players on every team, who won each of the past 20 grand finals and so on.

I have never had even the slightest interest in the game however. It bores me beyond belief.

I am therefore almost certain to fail any attempt at becoming an expert on the subject, for the simple reason that I have absolutely no interest in it....

Alternatively your lack of interest (and emotional involvement) combined with your learned knowledge could make you more objective and more able to see outcomes where the emotionally fanatic supporters fail.

 

[skyQuake]

Question(s): Efficient learning is really tough in this type of environment. Logically prepared strategies with strong backtest fail with alarming regularity. Stuff you think is worn out suddenly springs back to life without warning.

Not speaking for SKC's pair trading bit I think I can have a go at this one:

If something is sufficiently logical and sound, has performed well in backtesting and forwardtesting then it shouldn't fail too hard or too fast. Sure it gets crowded but if you trade it day in day out you can see it happening and eventually there is no alpha left (as with all good strategies).
Specifically, I try to look for strategies that are unconventional and/or difficult to backtest, which should lead to a longer lifespan (still forward testing that idea!)

How do you cope with such high uncertainty when seeking to learn?

IMO, if significant time and effort has been spent working out why x did y, but its still unclear (step 3), then move on.
Look for easier prey, but keep an eye on the anomaly for future reference.

What is your standard of proof required to change?

Diminished returns from the strategy if you're highish frequency. If mid/long term hold... Well I got nothing

How do you monitor efficacy of your approach to learning?

Haven't seriously considered the efficacy of my approach to learning. But generally, I know that if im managing a trade, im not exactly learning or doing research.

Here is an interesting grahpic on not forgetting what has already been learned!

 

[peter2]

Great questions DeepState and they made me review my circumstances. Thinking about these topics help me realise what I've learned and the experience I've gained.

1. Coping with uncertainty:
IMO all who fail to earn consistent profit fail to cope with the uncertainty inherent in the market. Overcoming the beginners cycle is the biggest obstacle for all new traders.

In my case I chose one trading approach (one break-out pattern) and traded that exclusively. I took advice from Mark Douglas and thought of each trade as only one of a batch of twenty. Knowing my worse case loss and being comfortable with it lessened my emotional response after each trade and helped me focus on the trading process. Once a batch was complete I reviewed each trade and saw what I did well and my mistakes. I learned over time and many trades what I had to do to create an edge. Of course has soon as I become competent with the break-out methodology the markets changed (GFC) and opportunities were few. It was time to learn another approach and the obvious one was a reversal strategy as the bear market wasn't going to last forever. The confidence and experienced earned in mastering the break-out strategy made mastering the new reversal strategy much easier. The markets changed again and the ASX went sideways for what seemed like ages. My usual profits became a growing accumulation of small losses and I had to do something different. I learned how to trade pullbacks and I did that through the forex markets.

2. World view: My current world view is like a 200d MA. It captures the main trend and will change when the markets turn. Experience has helped me ignore most issues raised by the media and I accept that the media drives sentiment that drives prices both up and down in the short term.

3. Changes: I seem to be trading much shorter time frames, days instead of weeks and much more leveraged instruments (less margin in accounts). I'm trading more intraday than ever before.

4. Proof required to change: Reviewing my application of my trading plans helps me monitor my performance and shows me the aspects to improve, change or eliminate. The performance of the basic strategies shows me what's working and what's not in the current market conditions. I attempt to match the trading strategy to the market type.

5. Efficacy of your approach to learning: Based on progress towards goals.

 

[DeepState]

IMO, if significant time and effort has been spent working out why x did y, but its still unclear (step 3), then move on.
Look for easier prey, but keep an eye on the anomaly for future reference.

IMO you are right on the money. Suggestion: Check out Arrowstreet Capital. They have a Sydney rep office. Could be worth the investment of a coffee to say "Hi" from one sharp investor to another. Probably get a nugget or two.


For clarity, I have no interests whatsoever in Arrowstreet Capital or its products.

 

[Trembling Hand]

1. Observe and study the market
2. Develop hypothesis to explain behaviour observed
3. Acquire new knowledge and data to check and test hypothesis
4. If hypothesis is valid, try to develop trading strategies around it
5. Repeat step 1

Everytime you go through this loop you learn something new

From my view the reason people have so much trouble is they fail to fulling embrace the correct steps to lead to understanding then haven't got a hope in hell of doing the hard repetitive work of practise and review that is required to actually develop expertise. If your hypothesis are incorrect but have formed your belief system you will have an extremely hard time practising something that is always going to fail. Yet at the same time if this is part of your fundamental belief system you will probably never be able to give up on the idea. End result is stuck in a process that doesn't work yet too hard to give up on.

However you have come upon this loop, it is exactly that related to expertise development and is very seriously undertaken by the most elite in any endeavor where skill and effort are related to outcome. Guys like you.

I read this and was totally rapt. It is rare. If you respond, thanks.

My first question related to high and low uncertainty environments and the application of expertise loops. Something like chess has no uncertainty. All the rules are known and never change. Then you get into financial markets where many things don't make sense for long periods of time. Feedback loops are long and mashed up with other factors whose interaction is weird and often never seen before, or only slightly similar. Worse, as soon as someone cracks the market, they get cracked as the markets figure out how they are doing it and replicate it. That's high uncertainty. You have a hard time figuring out probabilities at all and outcomes are often a...crap shoot.

You work in trading with pretty short horizons. My understanding is that this is pretty noisy stuff

How do you cope with such high uncertainty when seeking to learn?

For example the two bits I have underlined of yours are in direct opposition to my own believes. As a result I have been able to find 'something' in what you call "pretty noisy stuff" and repeatedly practise and review that practise to learn and grow skills. As the world lines up with my view it makes a self-fulfilling feedback loop to catch interest which then feedbacks into enthusiasm which feeds more practise and review and on and on.

 

[DeepState]

1. From my view the reason people have so much trouble is they fail to fulling embrace the correct steps to lead to understanding then haven't got a hope in hell of doing the hard repetitive work of practise and review that is required to actually develop expertise. If your hypothesis are incorrect but have formed your belief system you will have an extremely hard time practising something that is always going to fail. Yet at the same time if this is part of your fundamental belief system you will probably never be able to give up on the idea. End result is stuck in a process that doesn't work yet too hard to give up on.



2. For example the two bits I have underlined of yours are in direct opposition to my own believes. As a result I have been able to find 'something' in what you call "pretty noisy stuff" and repeatedly practise and review that practise to learn and grow skills. As the world lines up with my view it makes a self-fulfilling feedback loop to catch interest which then feedbacks into enthusiasm which feeds more practise and review and on and on.

Thanks TH.

1. Might another phrase be "hold your beliefs lightly"? Also, be "data-driven"?

2. My query there is not meant to imply this is impossible. Finding something from near static is what I do as well. We seek to find something useful in the midst of a very large amount of noise. I guess the major question for you given the thread content now is:

a. how do you prefer to go about finding the signal within the noise in the first place?
b. is it more important to you to find a few good ideas and develop them, or find a large number of working hypotheses and continue to trim them (as others are added)?

 

[tech/a]

I may be a little different.

I believe there will always e someone smarter than I am----if I need those smarts I'll hire them.
I know I cant know everything
I know I don't need to.
We all rise to our level of incompetence

As for trading.
I'm visual hence technical analysis fits for me.
Just as you have an inbuilt sense with people so you
Can and should have with anything your directly involved with.
If your in a relationship you know if it's strong and healthy or not.
You know if you or your kids and your friends are ok.

For e same with trading
T/H put it best with it's like recognizing your mother in a crowd.
It really is

As for noise.
Nothing moves without noise. You need noise to sort out direction.
The trick is to e able to understand what it is that the noise is
Sorting out.-----observe it long enough and you'll see it.

One persons noise is another's road map.

 

[KnowThePast]

I have an opinion too!

In addition to sorting out all the noise, how do you judge success, or the validity of the hypothesis? Just because one made money 3 years in a row does not mean he/she was right, perhaps just lucky. Statistically, it takes a very large number of overperforming years before you can confidently attribute result to skill, rather than luck.

And so, you start with hard data and economic theories that provide a plausible explanation of how something works. You verify that it is still plausible it may work today. And you take the plunge, knowing that the past is not an indication of the future.

How do you improve on something when you are now even sure that your base is correct? I don't know. But I think you still go through the learning loop and keep exploring new things. A holy grain will not be found, everything you discover will have only some degree of confidence to it. A combination of various concepts, however, often provides a framework for understanding the whole picture better. And so it comes back to experience. Experience and curiosity, as experience of doing the same thing over and over again is not terribly helpful (unless it works).

There's no gradual slope to learning anything. You may sometimes have no improvement for months, and then have a lightbulb moment and everything falls into place in one afternoon. But you still need to spend those months.

 

[Trembling Hand]

In addition to sorting out all the noise, how do you judge success, or the validity of the hypothesis? Just because one made money 3 years in a row does not mean he/she was right, perhaps just lucky. Statistically, it takes a very large number of overperforming years before you can confidently attribute result to skill, rather than luck.

Gee I would disagree. How do you come up with the idea that 3 years is not proof but could be luck. I doubt for a start if you are discretionary you would actually be trading the same after 3 years. Rarely does a market stay that static. But if we are talking short term trading that is literally 1000s of trades. If you equity curve is smoothly moving in the right direction that is proof.

 

[KnowThePast]

Gee I would disagree. How do you come up with the idea that 3 years is not proof but could be luck. I doubt for a start if you are discretionary you would actually be trading the same after 3 years. Rarely does a market stay that static. But if we are talking short term trading that is literally 1000s of trades. If you equity curve is smoothly moving in the right direction that is proof.

Hi TH,

I am referring to this:



And therein lies the difficulty - how/when do you change approaches, knowing that it takes years to be certain that any one approach is right.

With short term trades, you are right, it's more data and should be easier to draw conclusions from. But to counter that - thousands of trades in a market that is moving up over 3 years may produce a very different result in a falling market. So you still need years of data.

 

[Trembling Hand]

Hi TH,

I am referring to this:

View attachment 59712

And therein lies the difficulty - how/when do you change approaches, knowing that it takes years to be certain that any one approach is right.

With short term trades, you are right, it's more data and should be easier to draw conclusions from. But to counter that - thousands of trades in a market that is moving up over 3 years may produce a very different result in a falling market. So you still need years of data.

I reckon you could tell statistically in quite a short time with most short term trading. Remember this is not about beating a trending market or hanging around for a long term trend. You have to basically be flat at very regular intervals ie end of each day or every few days. There is no way to be getting a ride from a trending market because trends happen in overnight gaps not in buying the open and selling at the close. A long term trader/investor is judged on their decisions only a few times a year, luck can easily play a role. Short termers are judged many times a day every day.

 

[KnowThePast]

I reckon you could tell statistically in quite a short time with most short term trading. Remember this is not about beating a trending market or hanging around for a long term trend. You have to basically be flat at very regular intervals ie end of each day or every few days. There is no way to be getting a ride from a trending market because trends happen in overnight gaps not in buying the open and selling at the close. A long term trader/investor is judged on their decisions only a few times a year, luck can easily play a role. Short termers are judged many times a day every day.

Good points, TH. You are probably right with shorter term trading.

 

[skc]

However you have come upon this loop, it is exactly that related to expertise development and is very seriously undertaken by the most elite in any endeavor where skill and effort are related to outcome. Guys like you.

I read your post and have been thinking about a respond for the past few days. It is a difficult topic and I don't pretend to be an expert in the field of becoming an expert. The "loop" that I've come up with is simply how I saw my own development, although day-to-day I don't sit there and analyse whether I am doing Step 1 or Step 2 consciously.

I also don't know if I'd classify myself as elite or expert trader. I think I am beating the general crowd, but I am far from knowing everything about every situation. I am like your local university's Physics Professor, rather than Steve Hawkins. And in turn, Steve Hawkins probably barely knows anything compared to the vast body of knowledge within the entire universe.

I wil attempt to respond to your questions, but I apologise in advance that it will float randomly between general 50,000 feet views and nitty-gritty detailed specifics.

My first question related to high and low uncertainty environments and the application of expertise loops. Something like chess has no uncertainty. All the rules are known and never change. Then you get into financial markets where many things don't make sense for long periods of time. Feedback loops are long and mashed up with other factors whose interaction is weird and often never seen before, or only slightly similar. Worse, as soon as someone cracks the market, they get cracked as the markets figure out how they are doing it and replicate it. That's high uncertainty. You have a hard time figuring out probabilities at all and outcomes are often a...crap shoot.

Yes market is highly uncertain.. but the ultimate objective of trading (and investing) is not truth-seeking. It's just making profits. So you just need to be right some of the time, make as much as possible while you are right and manage the risk while you are wrong. It's starting to sound like some generic technical trading strategy, but any looped learning should focus on improving these three things.

You work in trading with pretty short horizons. My understanding is that this is pretty noisy stuff but your report a month or so ago of your second year in Propex (?) suggests that you are achieving high hit rate and risk managing very well. Even your trade frequency (which isn't massive for a trader) basically backsolves for very nice average return if you are betting symmetrically without elevating exposure into losses as a pairs guy. If you are martingale, then I'd come to a slightly different conclusion, but still positive of course given the record you have posted and the limited market directionality.

In my world, that's an amazing outcome. We don't trade with anything like that kind of frequency but we didn't enjoy the liquidity advantage (at the time...hehe).

No I am not martingale and I doubt any prop shop would be in business for long letting someone trade martingale with their capital. The result I posted was a combination of a pairs strategy and outright intraday trades. The outright trades were ~20-25% of the total P&L and does help to smooth the equity curve even further. And my size, while large relative to a private trader, is still tiny tiny in the whole scheme of things.

Question(s): Efficient learning is really tough in this type of environment. Logically prepared strategies with strong backtest fail with alarming regularity. Stuff you think is worn out suddenly springs back to life without warning.

How do you cope with such high uncertainty when seeking to learn?

Do you find that your worldview changes a lot or only incrementally or all over the place depending...

How different is what you are doing today from, say, 6 months ago, 1 yr or 2 yrs ago?

I'd describe my trading as evlolutionary... continuous learning and adjustments on the fly. I am always doing something different from what I was doing before.

For example, I used to pair up the iron ore miners, but I've pretty much stopped pairing them since early 2014. So how did I "learn" about this changing market and "improve" my trading by avoiding the sector?

1). I made observations that company share price correlations have broken down along with the iron ore spot price falling
2). I developed hypothesis that, with lower spot price, the margins are getting smaller and fluctuates more in percentage terms. Different producers are affected differently so share price correlations in the past should no longer be relied upon
3). I investigated the facts... it was handy that plenty of reserach analysts cover iron ore producers and their costs. It was clear some companies are still enjoying profits while some will struggle to survive.
4). I made the adjustments to stop trading them.

But as I said, all these were simply happening as I watch the market everyday. It's only when I have to put it in writing that I spend time analysing what I actually did. I didn't wait for my system to generate a string of losing signals before I stopped trading these pairs, as a slightly lazier and more stubborn pairs trader might have done.

And this is just one exmaple on one factor influcing one sector. I would go through the exact same loop when other things change... like gold, or rising $USD, or pending regulation changes, or overall market volatility changes etc. I am definitely not attempting to optimise my trading by finding the absolute truth or valuation or anything like that, I am just trying to continuously improve my win rate and reduce my losses by observing widely, applying some logic and verifying with facts.

What is your standard of proof required to change?

Pretty low. These days I will change when it "feels" right to do so. I don't think you can have a hard set of rules (like X failed trades in a row) because chances are they'd be too slow for both switching off and switching things back on.

However, the iron ore example is a simple one and it happens cyclically and there are plenty of parallels to learn from. If a new unknown situation arises and affects things more subtly, then I'd probably get hit few times before I react to what hit me (if I ever do).

How do you monitor efficacy of your approach to learning?

Not really in a structural manner. I guess I do ask myself... do I feel mostly in control, or do I feel something is broken. If I am in control, then I am learning/adjusting just adequately to keep on surviving. If I feel out of control and I don't know why, then I may be missing something. And the practical way is simply reduce the size and wait for it to pass. It's like driving to a brand new place... you slow down until you become familiar with the place, or until you get out of the foreign place and re-enter a familiar place.

 

[Trembling Hand]

Skc what you have listed above is what I was, probably unsuccessful, in getting at in bonkers "TA which pattern to trade" thread.

If you want to trade short term, imo, there is no "pattern" to learn. There is something far more dynamic going on because you are dealing with an uncertain market that has constant change in not just direction/trend but volatility, time, speed, volume and even where the action is.

Do you feel like what you are doing once you take a look at the progress over the years is actually quite different every say 6-12 months? But at the same time the process that leads to the change is actually the same?