Dump it Here

[ducati916]

Since we are talking all things $VIX



You'll have to excuse a couple of my lines (could have been more precise). I have also omitted the horizontal lines (resistance/support) entirely because this particular chart is just a short history. You would want a longer time frame to plot the horizontals on.

I use the VIX in combination with other indicators (probably about 4 in all). Mine are obviously not coded (big disadvantage, far more time consuming having to look at stuff) but you can still get there (even if you can't code) and have a pretty good idea what is going on.

Obviously (or maybe not so obvious) you can use the VIX in both directions (as Mr Skate is doing). There is a significant amount of information contained in this (seemingly) simple indicator. Clearly from the posted backtests, the advantage is clear.

So a practical application currently: look to far right of the chart. You will see that the steepness of the blue line is unlikely to be maintained. Any fluctuation now will create a 'break' in the trend line. This will in the current market trigger the 'pullback' that a number of posters have been talking about for the last week. Note also, had horizontal lines been added, the falling VIX would be at +/- the support point and would likely rally, which means falling prices in stocks.

Today, we have a bit of a pullback. Is it a pullback or failing (bounce) trend? It is simply a pull back because the other indicators do not confirm a trend (significant) failure.

But that identifies an important point (that Mr Skate has solved via code): how do you differentiate a 'signal' from noise? This takes some effort and is tricky in isolation. I use other confirmations, but you will have to find something that gells with your style of trading.

jog on
duc

 

[Roller_1]

@Roller_1, I'm not saying I'm at a level of competency but when you put a few thousand trades under your belt, fiddling around the edges becomes second nature & less stressful. To me, trading is a straight forward endeavour, yet many make a mess of. The most common problem that traders have is that they always tend to over complicated their trading when trading doesn't need to be complicated at all. We obsessed looking for confirmation for every decision or code alteration we make. It's a trait most of us suffer from, fiddling & optimising till eventually, we end up with a meaningless system that makes a fortune on paper but performs miserably in real trading "if you don't know what you are doing".

Let's not forget the basics
As traders we buy a position in the hope sometime in the future we will be able to offload our position to someone at a higher price than we brought it. Traders make money in the markets by exploiting changes in the prices & that's exactly what I try to do.

Skate.

No problem Skate, i agree that simple works best a lot of the time.

Personally though i need that conformation, or what feel is conformation.

Cheers

 

[Saqeeb]

@Saqeeb I was looking at your portfolio and wondered if you have thought about using (backtesting) Rotational Trading to weed out some of your lesser performing stocks that might tie-up critical funds that could be used in other positions that you would not normally be able to take.

@Trav. Thank you for suggesting and the example code. I will read up and look at implementing this and see if this will improve the backrest results.

You are correct about the typo. I will fix the error when I can access my computer after work.

 

[Warr87]

@Trav. Thank you for suggesting and the example code. I will read up and look at implementing this and see if this will improve the backrest results.

You are correct about the typo. I will fix the error when I can access my computer after work.

That also seems similar to the 'stale' exit that Skate uses, and the 1 I implement in my MAP strategy as well.

 

[Skate]

This brings me to the end
It's been exhausting, explaining how we can take advantage of the (MACD & the VIX Index) as a more efficient way to gauge the "sentiment of the markets" as trading is all about "predicting changes"

ASX200 XVI index
Well, I thought my research of using the (VIX) was done & dusted till I received a "PM" asking if I had also considered using Aussie (VIX) Norgate code ($XVI).

I was just reading through your thread regarding the use of the S&P500 VIX index. Have you considered also using the ASX200 VIX index?

Looking at both charts
Looking at both charts below they appear similar (to a blind man) but the responsiveness between the two is quite different. I made mention in a previous post that the VIX needed to be manipulated requiring a hefty amount of mathematical gymnastics to have it work efficiently. "Well blow me down" there was not one piece of code that was compatible between the two indexes. First off, I had to heavily modify the smoothing of the adaptive moving average before cutting the look-back period. The "parameters & settings" were so foreign to each other, they were like "chalk & cheese". The issue with squaring the average of the HHV/LLV over the lookback period was a real eye-opener.

Long story short
I got there in the end. For those heavily invested in my latest series of posts searching for a better "sentiment filter" it's only fitting I should share my findings.


Chart of the S&P 500 (VIX)





Chart of the ASX 200 (XIV)



Next post - I'll be comparing the Aussie (XVI) to the S&P 500 (VIX)
In the next post, I'll display a side-by-side comparison between using our own homegrown ASX 200 ($XVI) compared to using the S&P 500 ($VIX) as the "sentiment filter". (displayed in two different time frames)

Skate.

 

[Skate]

Two backtest periods (displayed)
Below is a side-by-side comparison graph between using our own homegrown ASX 200 ($XVI) compared to using the S&P 500 ($VIX) as the "sentiment filter". There is no need for a story, the backtest results speak for themselves.

Timing the markets
Some trading systems may benefit from an attempt to time the market more efficiently indicating when we should be entering & exiting the markets. The probability of success increases when we trade with the prevailing market trend & we need to "GTFO" when the sentiment changes.

Finally
There are of course many more indicators other than the (MACD) or the (VIX) you could use before deciding which is the best "Sentiment Filter" or "Index Filter" for your own Strategy as trading is all about "predicting changes". My ultimate desire is for you to be stimulated enough to think about developing your own unique filter to gauge "market sentiment". Finding a trading method that allows you to trade consistently & confidently, well then, you might have just found your "Holy Grail".

Comparison (1) Financial Year (YTD)





Comparison (2) Calendar Year (YTD)



All I ask
Please don't lose sight of why I've made this series of posts of searching for a more efficient way to gauge the "sentiment of the markets". These posts were made for you. "I'm just repeating stuff I already know"

Scorecard
Let's chalk this win up to @ducati916

(END)

Skate.

 

[peter2]

Thanks @Skate for that research series on the VIX. BIG thumbs up from me.
I was also wondering about the ASX VIX. Thanks again for posting your work.

The reason the US VIX is more effective as a sentiment indicator is that the SP500 is a better representation of the overall US market than the XJO in the ASX. Our market is dominated by two sectors, financial and materials. Other industries aren't properly represented in the main indices.

I liked this paragraph of @Skate 's.

Finally
There are of course many more indicators other than the (MACD) or the (VIX) you could use before deciding which is the best "Sentiment Filter" or "Index Filter" for your own Strategy as trading is all about "predicting changes". My ultimate desire is for you to be stimulated enough to think about developing your own unique filter to gauge "market sentiment". Finding a trading method that allows you to trade consistently & confidently, well then, you might have just found your "Holy Grail".

We can borrow ideas and indicators from anywhere and anyone but we must first accept them as our own in order to have the confidence to use them properly.

 

[qldfrog]

thanks for mentioning the XVI, I was not aware and it is available in the Premium data package under XVI (no $)
Will check if it improves my systems vs VIX

 

[qldfrog]

Thanks Skate,
if it helps
using a similar code as what I do with the VIX(and I assume different from yours), a trial with XVI was less than stellar, even after optimisation
since 1/01/2015 to today, daily:
longer period as I wanted significant entries to be statistically meaningful
using the same/similar code focusing on leveraging XAO trends either way:

the XVI is beaten flat so I will stick to VIX for my systems
@peter2 might be on something as to why and I would add XAO is quite different from ASX200.Hope it helps and reinforcse @Skate posts;
I did the tests and checks, better sharing than keeping for myself
Another great day trading, even with the VIX falling: have a great day

 

[Trav.]

That also seems similar to the 'stale' exit that Skate uses, and the 1 I implement in my MAP strategy as well.

Fair enough, I don't have the MAP code but with a system with limited funds or fully invested how long do you hold onto a losing position (s) ??

Million dollar question I know, but can be tested with a stale exit if stock is held for x number of weeks (bars).

In the example @Saqeeb has only held PDN for 4 weeks but with a loss of ~ 18% I would be looking at options, but that is me and hence why I like daily systems, so comes down to the individuals "sleep at night factor"

 

[Willzy]

Using $VIX as a sentiment filter on my version of the Weekend Trend Trader was an excellent idea Skate!

Cheers!

More investigation required but simply swapping it out from my usual index/sentiment filter even without optimization yielded an additional 2% of CAR and a more visually pleasing equity curve.

I'll also have a play with using $VIX as an additional feature in a logistic regression machine learning model for some US stocks to see what it can do there. I'll report back!

 

[Warr87]

Fair enough, I don't have the MAP code but with a system with limited funds or fully invested how long do you hold onto a losing position (s) ??

Million dollar question I know, but can be tested with a stale exit if stock is held for x number of weeks (bars).

In the example @Saqeeb has only held PDN for 4 weeks but with a loss of ~ 18% I would be looking at options, but that is me and hence why I like daily systems, so comes down to the individuals "sleep at night factor"

I give mine a bit of time, as in the nature of trend following where some stocks will pullback before continuing up. But if it goes nowhere I'd rather move my money somewhere else. Your idea isn't the same, though, as the 5 worst stocks may not necessarily be losing money.

 

[Skate]

Using $VIX as a sentiment filter on my version of the Weekend Trend Trader was an excellent idea Skate!
More investigation required but simply swapping it out from my usual index/sentiment filter even without optimization yielded an additional 2% of CAR and a more visually pleasing equity curve. I'll also have a play with using $VIX as an additional feature in a logistic regression machine learning model for some US stocks to see what it can do there. I'll report back!

@Willzy what a great post & thank you for your kind words.

More investigation required
I'll be looking out for your report with a frank & honest appraisal if the (VIX) is beneficial to any of your systems. I'm so happy you were motivated enough to make your first post in the "Dump it here" thread. Your words were "music to my ears", it's exactly what I wanted to hear, well done.

Why post ?
If my posts affect anyone's behaviour, I'm a winner - that's my end game..

My ultimate desire is for you to be stimulated enough to think about developing your own unique filter to gauge "market sentiment". Finding a trading method that allows you to trade consistently & confidently, well then, you might have just found your "Holy Grail".

Skate.

 

[Trav.]

Your idea isn't the same, though, as the 5 worst stocks may not necessarily be losing money.

I think that the idea of the rotational trading is that each stock in your watchlist (eg, S&P ASX 300 or user defined) is ranked via the PositionScore function as defined by you and if that stock falls out of the top 5 then it would be replaced.

So with this example Saqeeb has a portfolio of 20 stocks (max positions) so you would set up the ranking for that portfolio size as per below

PositionScore = 50 - RSI(); // PositionScore as defined by you
SetOption("WorstRankHeld",30); // Remove stock if outside of top 30 ranked stocks​

Then once a month / quarter etc you would run the Rotational portion of the code and the non performing stocks (if any) positions would be exited, allowing you to take on a new position and address any portfolio re balancing sections of your plan.

Non performing stocks in this case would be the ones that are ranked outside the top 30 stocks, so with a portfolio of 20 stocks you are bound to have a few going poorly and their ranking slipping down the list.

This may not be for everyone but if your system is running at 40% winners then you may want to turnover stocks periodically.

Just food for thought.

 

[qldfrog]

I think that the idea of the rotational trading is that each stock in your watchlist (eg, S&P ASX 300 or user defined) is ranked via the PositionScore function as defined by you and if that stock falls out of the top 5 then it would be replaced.

So with this example Saqeeb has a portfolio of 20 stocks (max positions) so you would set up the ranking for that portfolio size as per below

PositionScore = 50 - RSI(); // PositionScore as defined by you
SetOption("WorstRankHeld",30); // Remove stock if outside of top 30 ranked stocks​

Then once a month / quarter etc you would run the Rotational portion of the code and the non performing stocks (if any) positions would be exited, allowing you to take on a new position and address any portfolio re balancing sections of your plan.

Non performing stocks in this case would be the ones that are ranked outside the top 30 stocks, so with a portfolio of 20 stocks you are bound to have a few going poorly and their ranking slipping down the list.

This may not be for everyone but if your system is running at 40% winners then you may want to turnover stocks periodically.

Just food for thought.

That would favour stock at end of trend and remove the one you get just as they start...
Really see the benefits in some systems but also the huge costs on other.
Stale checks would get them out on my systems, the longer i own a position the higher the profit or it get the chop
Just to say consider wisely if this is for you, look at average longest holding periods for winner loosers...or just backtest inbuilt rotational systems

 

[Trav.]

That would favour stock at end of trend and remove the one you get just as they start...

@qldfrog I'm not sure that is correct. The entry would still be based on your normal system entry requirements which wouldn't be at the end of a trend, it is how you program it so really up to your imagaination

Really see the benefits in some systems but also the huge costs on other.

Again not sure if this a factor as rotation would be done monthly or quarterly so impprovemnts in portfolio should outweigh a few extra transactions

Stale checks would get them out on my systems, the longer i own a position the higher the profit or it get the chop

Yes that is a good way to do it, what is the criteria....n-bar stop ROC ??? many ways to skin a cat.

Just to say consider wisely if this is for you, look at average longest holding periods for winner loosers...or just backtest inbuilt rotational systems

I think that this is what we are talking about. Backtest / optimise to see if it suits your system or trading mindset.

 

[qldfrog]

Yes that is a good way to do it, what is the criteria....n-bar stop ROC ??? many ways to skin a cat.

indeed I played with: after n bar ownership, ROC in the most recent bars has to carry on above specific % otherwise get out as you may statistically find better opportunities getting in other position.
I noted as well you were considering rotation on very long term: quarterly etc, indeed if you keep positions that long, makes a long of sense;
i rarely have positions reaching that duration on my weeklys

 

[frugal.rock]

The discussion about rotation is great.
Bias confirmation of my thoughts as mentioned re Ausnet in Mr Skates (brilliant overall) action strategy.

F.Rock

 

[Trav.]

Just to put up a quick back test result with the worst rank feature being optimised you can see the difference with the different re balancing options.

Discussion only - not using this in any trading system

universe = ASX200

$100k Equity, Fixed position $10k, Max Positions 10

Rotated out end of each month

Ranking system = ROC(Close, 60);

Extra column "worst" is the WorstRankHeld option for determining if to rotate stock out of portfolio, so I have 200 stocks in my watchlist the system has selected;

Optimal worst ranking system for best CAR/MDD



or sorted for maximum profit.



​

 

[Skate]

What have I been thinking about? "Thinking"
I've made 168 posts referencing the word "thinking" that may have been missed by many.

Thinking
Thinking is so much a part of us that we usually don’t think much about thinking. If we pause for a moment & observe our thoughts, we become aware of the activity of our mind. Our thinking is tirelessly telling us what to believe & we tend to accept whatever it's telling us "as an accurate reflection of reality". We have become accustomed to accepting our thoughts "as fact" even if it's a "distorted fact". No matter what our thinking conveys, we rely on it heavily every day.

Specialised skill
Thinking controls you without you even realising it. It's very rare that you can control your thoughts & we tend to accept the way we think as the norm, not realising that thinking can be controlled & developed into a specialised skill.

Why post about thinking?
Because trading is a "thinking game".

Skate.