Dump it Here

[MovingAverage]

Having increased bets concentrated on a few positions insures when you have a "win" you have a "decent win".

Skate.

Fiddling with your position sizing to get acceptable results is just masking an underlying system with mediocre performance.

 

[Skate]

Fiddling with your position sizing to get acceptable results is just masking an underlying system with mediocre performance.

@MovingAverage you can't keep misrepresenting what I have said to suit your narrative. This has to stop.

Tennis
Let's not play tennis with each other, if you have an alternative opinion - feel free to express it. You have the right to express an alternative view or even make a few posts on the correct way to position size a "mechanical trading system".

Blurting out one-liners is not educational to anyone
We can both hold a different opinion & neither of us could eventually be right. Expressing why you hold your opinion & why I hold mine helps others to understand the point from different angles.

For the benefit of others
In a recent private message, I remarked how trading has changed for me since COVID & the steps I have taken to bring my strategy in line with current market conditions. I've even explained when trading was good, I took advantage of those conditions having a larger than expected "position size" ranging from (40-53) positions in each portfolio.

Desperate times required desperate measures
As I have amended my trading strategy I've explained why it was necessary. After doing so I posted comparison charts & graphics to lineate the difference it made. I've even taken the time to explain why having a "Take Profit Stop" can be an advantage to a strategy's profitability in these trying times.

Why did I explain all this?
I did it in the hope others would think about what I have done & why. I didn't spew out one-liners but took the time & effort to give a measured explanation & the "reason" for the changes I've made.

How things have changed
The 20th of January 2019 seems like a long time ago & it probably is when it comes to trading. In the 2019 post, I explained my strike rate (win rate) was around 50%. I always post accurate & up to date records so others can experience my trading method. At times I overshare, but that's just me.

Since the 1st of January 2020
My strike rate has hit an all-time low of 39%. I'm just saying, back in the early days "up until the COVID period" trading was so much easier & my equity curve was so much smoother. Those who maintain trading has not changed since COVID well frankly my trading doesn't reflect the status quo. Then again, it just might require more research on my part to establish if my methodology of re-coding made the difference or was the real culprit for the lower strike rate was due to the much tougher trading environment.

My gut feeling is it's the latter
I should also say a 39% strike rate is on par with a trend following strategy but as usual, I'm seeking above par results.

Since COVID my win rate has dropped to 39%.
I had not "realised" my strike-rate had dropped substantially as the profits during this period were still strong. I must confess I only look to solve a problem when one is noticed.

Like the old saying
"If you can't see the problem, there is no answer"



For the advantage of others
(a) Consider if your post adds value to the discussion.
(b) When expressing an alternative view, start off by saying “In my opinion …” and try to focus on the issues rather than the person.
(c) We are all wordsmiths to some degree & it's easy to incite an emotional, knee-jerk response, creating an emotional outburst.
(d) Ask a detailed question & you will get a detailed response.
(e) Ask a one-liner, & all it deserves is a one-liner in response.

Don't post one-liners
Most members ask questions or post a response making a statement as a one-liner, (not all but some) & sometimes their responses are the same, it's unhelpful. If you skim read my post you'll learn nothing & at times you will "completely miss the point" I was trying to get across by jumping to the wrong conclusion.

It takes time
It takes a lot of time & effort to give a measured response so all I'm asking is if you have a question detail it precisely, let me understand why you don't understand as constantly going back & forward with others is tiresome, boring & more importantly, it wastes our time.

Same old same old
Some members tend to follow patterns of behaviour, posting to invoke an emotional response, never answering questions directly to justify their position but they "always demand documentary evidence" from others to support their assertions, while offering none in return.

# Please don't let that be you.

On a lighter note
@MovingAverage thank you for posting in the "Dump it here" as it has certainly helped, without the generosity of senior traders "making a contribution" the "Dump it here" thread would have died long ago.

With respect
If you or others want to make a difference, keep the material on the issue at hand while refraining from telling others their views are flawed. Simply take the effort to explain why from your advantage point.

Skate.

 

[MovingAverage]

@MovingAverage you can't keep misrepresenting what I have said to suit your narrative. This has to stop.


Blurting out one-liners is not educational to anyone
We can both hold a different opinion & neither of us could eventually be right. Expressing why you hold your opinion & why I hold mine helps others to understand the point from different angles.

It's not a one-liner...you like quoting my past posts so why not look back to the post I made sometime ago to you about system testing and position sizing. I said back then that if you really want to understand a system's performance you need to isolate the impact of position sizing--you discounted me with no reason. I clearly stated back then my opinion on position sizing and the artificial influence it can have on a system's performance. Nothing new here. I'm merely suggesting that if your system isn't performing then to fiddle with your position sizing isn't fixing the problem.

BTW, you keep saying the market has changed but you've offered no insight into that. I recently countered your opinion with details about why I believe the current market hasn't changed. So please, do not suggest I offer no insight into my opinions.

Also, not that long ago in one of my prior posts (which I sure you can find) I suggested profit taking exit in breakout systems was a good thing. I even posted up example trade of why it was good. You absolutely lambasted the idea without any supporting evidence and here you are now suggesting they're good

 

[othmana86]

Fiddling with your position sizing to get acceptable results is just masking an underlying system with mediocre performance.

What if position sizing based on volatility forms part of your strategy?

I think radge refers to doing this in his monthly rotational strategy, wether he actually does it or not I’m not sure.

 

[Skate]

For transparency
There was one sell signal this week, a position not held in the Platinum10-position portfolio. The corresponding chart is uploaded to display the entry & exit points.












Skate.

 

[MovingAverage]

What if position sizing based on volatility forms part of your strategy?

I think radge refers to doing this in his monthly rotational strategy, wether he actually does it or not I’m not sure.

Position sizing is very important—and I’m not advocating one position sizing over another. Yes, position sizing is a crucial part of your strategy and if volatility based techniques work for you then great, but there is a subtle point that I was trying to make with my earlier post. In MY OPINION (it’s important to state that in this thread now) a mechanical system (not your overall strategy) should be tested and evaluated with minimal influence from position sizing. This will give you a much clearer insight into how your system really behaves. I’ll put up some sim results later today to better illustrate the issue.

 

[MovingAverage]

So here is the point I am endeavoring to highlight. The below shows backtests for a system (what the system is makes no difference for this purpose). The first results are for the system using the common positioning sizing technique of a maximum of 20 open positions each position being 5% of the portfolio's value. This technique appears a lot in the backtesting results posted in this thread and is somewhat of a defacto standard here.

The returns below look pretty good don't they--and you'd be forgiven for thinking this is great let's go live with this. Pay close attention to the returns from 2011 to 2021.



Well now let's strip out the impact of ever increasing position sizes that is inherent in the ever popular position sizing technique of 20 position each being 5% of the portfolio's value and instead level the playing field by maintaining the same position size over the same 20 or so years. Below are the returns for the same system but using a fixed dollar position size, which in this case is 20 positions each of $5000 ($100,000 starting value) and this remains across the 20 year span. The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance.

 

[ducati916]

So here is the point I am endeavoring to highlight. The below shows backtests for a system (what the system is makes no difference for this purpose). The first results are for the system using the common positioning sizing technique of a maximum of 20 open positions each position being 5% of the portfolio's value. This technique appears a lot in the backtesting results posted in this thread and is somewhat of a defacto standard here.

The returns below look pretty good don't they--and you'd be forgiven for thinking this is great let's go live with this. Pay close attention to the returns from 2011 to 2021.

View attachment 134510

Well now let's strip out the impact of ever increasing position sizes that is inherent in the ever popular position sizing technique of 20 position each being 5% of the portfolio's value and instead level the playing field by maintaining the same position size over the same 20 or so years. Below are the returns for the same system but using a fixed dollar position size, which in this case is 20 positions each of $5000 ($100,000 starting value) and this remains across the 20 year span. The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance.

View attachment 134511

So a couple of questions:

(a) I assume not an issue where position sizing (as alluded above) is a specific variable in the strategy; and
(b) The results above (the fixed position size) rather looks like the markets changed, would you agree?

jog on
duc

 

[MovingAverage]

So a couple of questions:

(a) I assume not an issue where position sizing (as alluded above) is a specific variable in the strategy; and
(b) The results above (the fixed position size) rather looks like the markets changed, would you agree?

jog on
duc

No idea what you mean by point a). Position sizing is what it is—an input to the system. Depending on your definition of variable, my sims show both scenarios—first as a variable and the second where it is not a variable. As for b) you do not need my sim results to come to that conclusion—just look at my earlier chart of XAO and my comments that some aspects and dynamics of the current market resemble our market post the 2008 GFC. In any event, the market is always changing day-to-day, week-to-week, month-to-month, year-to-year and so on so on.

 

[ducati916]

1. No idea what you mean by point a). Position sizing is what it is—an input to the system. Depending on your definition of variable, my sims show both scenarios—first as a variable and the second where it is not a variable. As for b) you do not need my sim results to come to that conclusion—just look at my earlier chart of XAO and my comments that some aspects and dynamics of the current market resemble our market post the 2008 GFC.

2. In any event, the market is always changing day-to-day, week-to-week, month-to-month, year-to-year and so on so on.

1. Well that's not exactly what you said:

"The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance."

If position size as an input is a variable that is actively chosen, specifically to improve performance, then that is different to simply masking poor performance through ignorance and error of what you are doing.


2. So 'change' in markets (obviously) needs a definition. Fluctuations, which you are (seemingly) referring to, are not the definition of 'change' that others are using or possibly only myself.

When we (or I) talk about 'change', we are talking about forces within the markets that drive the secular trends as opposed to simple volatility.

On that basis, would you agree that the market appeared to 'change' in 2011 using your backtest as an example?

jog on
duc

 

[Cam019]

So here is the point I am endeavoring to highlight. The below shows backtests for a system (what the system is makes no difference for this purpose). The first results are for the system using the common positioning sizing technique of a maximum of 20 open positions each position being 5% of the portfolio's value. This technique appears a lot in the backtesting results posted in this thread and is somewhat of a defacto standard here.

The returns below look pretty good don't they--and you'd be forgiven for thinking this is great let's go live with this. Pay close attention to the returns from 2011 to 2021.

View attachment 134510

Well now let's strip out the impact of ever increasing position sizes that is inherent in the ever popular position sizing technique of 20 position each being 5% of the portfolio's value and instead level the playing field by maintaining the same position size over the same 20 or so years. Below are the returns for the same system but using a fixed dollar position size, which in this case is 20 positions each of $5000 ($100,000 starting value) and this remains across the 20 year span. The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance.

View attachment 134511

I'm not sure I agree with you. Haha! That's not surprising, right!

Isn't this just simple maths? Simple vs compounding rate of return.

Let's say your account size started with $100,000. Well $5,000 position size is obviously 5% of your accounts capital. However, the larger your capital grows (which is the goal over time) the lower the percentage of your account balance each $5,000 position size will be, and therefore each realised profit or loss will have a lesser effect on the growing capital base over time - hence the lower returns.

Why bother backtesting with a position sizing model that does not compound realised profits? This does not make sense to me.

 

[MovingAverage]

1. Well that's not exactly what you said:

"The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance."

If position size as an input is a variable that is actively chosen, specifically to improve performance, then that is different to simply masking poor performance through ignorance and error of what you are doing.

I have absolute no idea what point it is you are trying to make here, but that is probably because I have a low IQ. You keep referring to “input variable” and not sure why. I’ve made my point on position sizing having an undue influence on systems and how it can mask an otherwise mediocre system—which I stand by. Given I can’t comprehend your points let’s just leave this here unless you can simply your proposition for my simple mind.

1. Well that's not exactly what you said:

"The above results started with $100,000. Guess what...check out the returns from 2011 to 2021...bloody terrible. And this is the point I was making earlier: fiddling with position sizing to get acceptable results is just masking a system with mediocre performance."

2. So 'change' in markets (obviously) needs a definition. Fluctuations, which you are (seemingly) referring to, are not the definition of 'change' that others are using or possibly only myself.

When we (or I) talk about 'change', we are talking about forces within the markets that drive the secular trends as opposed to simple volatility.

On that basis, would you agree that the market appeared to 'change' in 2011 using your backtest as an example?

jog on
duc

Again, I don’t really understand your point. Can you give me an explicit example or two of what forces you are referring to. The reality is that the charts of the current market suggest we have been here before—yes I know you’re not referring to charts. If these forces you speak of are external then I have no idea or interest to understand those because they manifest themselves in the charts. As I said before, markets are changing all the time, but are we in a period that is radically different to the past—nope I don’t think so. Not prepared to draw any conclusions on market dynamics from my sim results other than to say yes pre ‘11 and post ‘11 are different, but again that is very obvious from my earlier post on XAO

 

[MovingAverage]

I'm not sure I agree with you. Haha! That's not surprising, right!

Isn't this just simple maths? Simple vs compounding rate of return.

Let's say your account size started with $100,000. Well $5,000 position size is obviously 5% of your accounts capital. However, the larger your capital grows (which is the goal over time) the lower the percentage of your account balance each $5,000 position size will be, and therefore each realised profit or loss will have a lesser effect on the growing capital base over time.

Why bother backtesting with a position sizing model that does not compound realised profits? This does not make sense to me.

I agree with you and what you say is right. But my point is all about understanding your system’s behaviour without external influences. Again IN MY OPINION when you’re initially building, evaluating or refining your system I’m just suggesting it be done without the influence of position sizing. By all means factor in position size but that should be done after you’ve settled you’re initial system using fixed position sizing. You can then get a better understand of how much of your strategy’s performance can be put down to the system per se and how much to position sizing. Here’s is a crude example…imagine someone starting live trading that system I posted results for and they started in 2019…do you reckon they would really have any chance of even getting double digit returns for that year?

 

[Skate]

I clearly stated back then my opinion on position sizing and the artificial influence it can have on a system's performance.

In MY OPINION (it’s important to state that in this thread now) a mechanical system (not your overall strategy) should be tested and evaluated with minimal influence from position sizing. This will give you a much clearer insight into how your system really behaves. I’ll put up some sim results later today to better illustrate the issue.

The first results are for the system using the common positioning sizing technique of a maximum of 20 open positions each position being 5% of the portfolio's value.

Well you live & learn
I don't know of anyone who would backtest using anything other than fixed dollar position sizing. Using any other method over time you wouldn't be able to enter a position as the bet size would be enormous. I'll show you that in a few backtest & why it is silly.

Disclaimer
When you see any of my backtests that are uploaded they are all "Fixed Dollar Position Size" meaning the bet size remains constant.

Talk about misleading backtest results
I hope no one is using that methodology when in the development phase of a strategy or backtesting the worth of a strategy. The ASX has a very shallow depth of the market that can't accommodate large bets. Even so, if it could the slippage would be enormous.

Let me post a few captures
"The Platinum Strategy" is my latest creation so I'll use this strategy as an example. The strategy is a 10-position portfolio with $10k bets with no rebalancing. I'll explain rebalancing when it comes to the next bet sizes or (Position Sizing) in the next post. But let's evaluate a backtest using a few different well-established methods when it comes to live trading.

The backtest will be in this order
(a) Fixed Dollar Position Sizing (fixed Dollar bet size)
(b) Fixed Fractional Risk %
(c) Fixed % of Portfolio Equity



Skate.

 

[Skate]

The backtest will be in this order
(a) Fixed Dollar Position Sizing (fixed Dollar bet size)
(b) Fixed Fractional Risk %
(c) Fixed % of Portfolio Equity

The backtested strategy will be "The Platinum Strategy"
Please notice that the position size (bet size) is a fixed $10k with "NO REBALANCING". I'll explain what rebalancing is in the next post



The backtest period will be 4 years in duration
If you follow along you will experience firsthand why backtesting other than a fixed dollar amount can be misleading. This was the point @MovingAverage was making. The backtest period is from 18th December 2017 to the end of trade Friday 17th December 2021. The position size is fixed at $10k with 10-positions.

(a) Fixed Dollar Position Sizing (fixed Dollar bet size)




Fixed (Position Size) bet size of $10k
As you can see the bet size is constant & it never changes.

(a) Fixed Dollar Position Sizing (fixed Dollar bet size)



Skate.

 

[Skate]

The backtest will be in this order
(a) Fixed Dollar Position Sizing (fixed Dollar bet size)
(b) Fixed Fractional Risk %
(c) Fixed % of Portfolio Equity

The backtested strategy will be "The Platinum Strategy"
Please notice that the position size (bet size) is Fixed Fractional Risk %.



The backtest period will be 4 years in duration
Following along you will experience firsthand why backtesting other than a fixed dollar amount can be misleading. This next capture will display how the next bet size will be so large it's untradeable.

The backtest period
The backtest period is from 18th December 2017 to the end of trade Friday 17th December 2021. The position size is a Fixed Fractional Risk % with 10-positions.

(b) Fixed Fractional Risk %




Fixed Fractional Risk % (Position Size) bet size
As you can see the bet size constantly varies & at times untradeable. The bet size starts off around $10k but quick balloons out to around $90k for each bet. It's not achievable to enter the market without slippage "which will" shift the market.

(b) Fixed Fractional Risk %




Skate.

 

[Skate]

The backtest will be in this order
(a) Fixed Dollar Position Sizing (fixed Dollar bet size)
(b) Fixed Fractional Risk %
(c) Fixed % of Portfolio Equity

The backtested strategy will be "The Platinum Strategy"
Please notice that the position size (bet size) is a Fixed % of Portfolio Equity.



The backtest period will be 4 years in duration
Following along you will experience firsthand why backtesting other than a fixed dollar amount can be misleading. This next capture will display how the next bet size will be so large it's untradeable.

The backtest period
The backtest period is from 18th December 2017 to the end of trade Friday 17th December 2021. The position size is a Fixed % of Portfolio Equity with 10-positions.

(c) Fixed % of Portfolio Equity




Fixed % of Portfolio Equity (Position Size) bet size
As you can see the bet size constantly varies & at times untradeable. The bet size starts off around $10k but quick balloons out to around $146k for each bet (what a joke). It's not even achievable to enter the market without slippage "which will" shift the market. Also getting a $146k on would be next to impossible in one transaction.

(c) Fixed % of Portfolio Equity



Skate.

 

[Skate]

I'm now talking about live trading & position sizing
There are different ways to position size other than fixed dollar amounts when you are trading a live account. The benefits are there when you have a limited supply of trading capital. Like most, I have my way of "rebalancing my position size". Let's call it my bet size as it would be less confusing. I've made over 20 posts on how I rebalance my next series of bets (with graphics).

Amibroker "Exploration Analysis"
Allows you to modify your next bet SIZE with a simple line of code no matter the balance of your outstanding funds not in the markets.

Rebalancing is simply calculating the dollar size of the next bet or series of bets
Pyramiding works both ways, the bet will either increase or decrease according to the trading funds available.

Skate.

 

[Skate]

Trading in the pre-auction necessitates dollar re-balancing of positions sizing
Re-balancing will be a direct correlation to the current trading account balance. It's a simple process & it's how I increase the size of the next series of bets. My trading account balanced is averaged over the next series of bets to have all my trading funds deployed in the markets. Using a (+/-) 3% premium results in "unused" fund. The outstanding funds are added to the regular bet size of the portfolio.

"Re-balancing Explanation" (Position Size)
Re-balancing my next bet is referred to as "Position Size". Re-balancing is a simple technique to reinvest profits & make corrections for losses. Re-balancing ensures every trading Dollar is put into the markets "to fight the good fight".

Skate.

 

[Skate]

How do I rebalance?
Re-balancing is adjusting the size of my next bet. Position-sizing (the bet size) uses the trading Bank balance feed to calculate the size of the next bet or series of bets. It's simply a way of putting every dollar to work.

What is the Re-Balancing Formula?
Trading Bank Balance/outstanding positions = new "Position Size" the new bet size. This will now be the new bet for each & every pending trade. The new "Position Size" also calculates the number of shares to buy in the pre-auction. It couldn't be simpler.

Skate.