Newbie Lessons - All your questions answered

[Sir Osisofliver]

Sir O,

Is it easy to determine what % of a company's shares are held by the board/directos/internally? Is this concept measurable?

Appendix 4b material announcement

At first glance the % of winning trades may be quite low and as such may not suit all traders/investors.

What do you base that on?

I can only see data onthe ASIC website for short %'s held back to June 2010. Is this enough data to perform the required backtesting?

I had the system before I had the data. Looking for other conditions in the market. The data merely adds to confirmation. However you are correct if you just used the data set at ASIC it wouldn't be statistically relevant

Cheers

Sir O

 

[Sir Osisofliver]

When I get sometime.!!

That would be great Tech. I won't be doing any current examples.....but you could. JBH is looking like a target at present. ~24% of the issued capital is in short's. I'm sure that the newbies and myself would appreciate your technical analysis of the stock at this time.

Cheers

Sir O

 

[kid hustlr]

Nice discussion as always Sir O.

I will have a discussion about "other conditions in the market" and report back.

 

[NewbAndy]

Lets get come ranking talk started. here's my initial rankings. Change as you see fit (or add to it).

Pro
1. Measureable short value %
2. Have latent demand (can see this through volume increase when it unwinds)
3. Cuts down on list of companies to research
4. you can see confirmation (% shorts decreasing)


Con
1. The data released by ASIC is up to 5 days after it happened.
2. Company heavily shorted for a reason
3. Data only goes back 1 year. is it enough
4. is the short a function of the market

Hi Everyone,

As has been mentioned- is there a way of getting ASIC data more recent? (I'm guessing not) Perhaps there's a way of estimating this?

In the very brief look I've had, surely the #4 con above could be rated higher? Wouldn’t interference from the XJO play a bigger part in confusing the reason behind rises in stock price?

Cheers
Andy

 

[kid hustlr]

Taking a stab,

I think the overall index will act as an indicator. If the market rallies up then a heavily shorted stock is likely to rally even harder. My reasons for this are:

- Those shorting are most likely more active traders and will look to close there positions, leading to the in built demand you discussed earlier.
- A level of computer buying would occur due to the links to the SPI and this would again squeeze out shorters.
- Market sentiment may begin to change and investors may adjust their opinion/outlook/models and recognise this stock is now a buy.

As an aside, JBH is very close to a strong support level back in Nov 2008. Whether this is coincidence or not I'm not sure.

 

[burglar]

Taking a stab,

I think the overall index will act as an indicator. If the market rallies up then a heavily shorted stock is likely to rally even harder. My reasons for this are:

- Those shorting are most likely more active traders and will look to close there positions, leading to the in built demand you discussed earlier.
- A level of computer buying would occur due to the links to the SPI and this would again squeeze out shorters.
- Market sentiment may begin to change and investors may adjust their opinion/outlook/models and recognise this stock is now a buy.

As an aside, JBH is very close to a strong support level back in Nov 2008. Whether this is coincidence or not I'm not sure.

So you're saying that market shorts will unwind when the index pivots upwards.
And the index is available in real time!!

So how do you find out if JBH is shorted in order to short the market?
I'll take a stab and say we should look for a correlation of JBH shareprice against the index.
Or an earlier breakout!

 

[Sir Osisofliver]

The next step.... Measurement Tools.

So we've stress tested the idea and identified the pro's and con's. We've also ranked them in importance.

(I'm using what's been stated)

Pro
1. Measureable short value %
2. Have latent demand (can see this through volume increase when it unwinds)
3. Cuts down on list of companies to research
4. You can see confirmation (% shorts decreasing)


Con
1. The data released by ASIC is up to 5 days after it happened.
2. Company heavily shorted for a reason
3. Data only goes back 1 year. is it enough
4. is the short a function of the market

What we need to now determine is what we will measure these pro's and con's against. We need to identify the possible indicators that we can use against each of these. Some of these will be easy..and some will require some thought. I ususally spitball or brainstorm several idea's to try and refine the kind of data requirements that the system will use. (I am going to use JBH for this - THIS IS NOT A RECOMMENDATION)

Pro
1. Measureable short value % From here I can download the daily aggregate short positions JBH is currently sitting (with 5 days delay) at an indicated level of 24.01% of the issued capital in short positions. At current price of of $8.60 this is 23,734,597 shares or $204,117,534 worth of shares to purchase to close out the "short overhang". (my terminology)

Question: What indicators can I construct from this data (knowing that the con is that it will have a 5 day delay?)

Question: Is it sufficient to use only this data? Do I want to determine the amount of inactive stock (IE that held by directors) to alter my percentage?


2. Have latent demand (can see this through volume increase when it unwinds)
To determine a volume increase (which does not suffer from latency like number one above), I need access to volume data. With a daily in the above I'm going to want to look at Daily data, with weekly data for longer term confirmation. Our hypothesis is that the shorts (being a shorter term instrument) will unwind rapidly, causing a volume increase.

Question: What volume indicators can I construct from this data?

Possible indicators are:
Moving Average based indicators using volume, (SMA, EMA)
Bollinger Band - Using a 2 standard deviation
Anything else? If so why? And why would I use the above?

3. Cuts down on list of companies to research

We need to construct a watchlist of the potential targets.

Question: At what point do targets enter the system? (IE how much is Heavily shorted?)

5% of issued capital (same as a significant holder)
(We'll use this as a working figure for now and know it may be subject to adjustment).

4. You can see confirmation (% shorts decreasing)

Question: What percentage move in the short position counts as confirmation?

For this we are using weekly data (as mentioned above). A temporary indicator would be 0.5% movement (10%).

Question: Why do you think I used this figure? If you want to use something else...why?

With the con's I'm looking at mitigation. How can I offset these factors? What might I use?

Con
1. The data released by ASIC is up to 5 days after it happened.

Some of the indicators above are real time, (Volume for example, Price movement will also be real time). Does the real time nature of some of the system inputs offset the time delay in this indicator?

2. Company heavily shorted for a reason.

The determination of whether the target has been oversold will be a judgement call against a research process. To me this is where the Fundamental aspects of the System come into play. What fundamental questions do I want to answer? Here's a few....what else would you want?

PE Historic
PE Peer
PEG Peer
Quick
Acid
Director Holdings as percentage
top 10 Holdings as percentage
Recent news

3. Data only goes back 1 year. is it enough?

This con will only be answered once we have some test data.

4. is the short a function of the market?

As pointed out by Burglar and Kid Hustlr...it's proposed that this system will have greater efficiency in different market conditions. That the unwinding of shorts will happen at the fastest pace during an uptrending market, or at the change in trend of the market. We could use a correlation of the target JBH, against an index, to refine our system and give us a greater probability of success.

Question: Which is the best index to use?

XAO
XTO
XJO
XKO
XDJ

A combination?


Sir O pours himself a scotch and dry and waits in the corner again for discussions to happen.


Cheers

Sir O

P.S.

I won't be giving any technical analysis on JBH. I'd appreciate if someone else could give it a shot though. (I've removed a lot of stuff to give you a relatively simply chart with price and volume). ONCE MORE WITH FEELING. AGAIN - NOT A RECOMMENDATION

 

[Trembling Hand]

I won't be giving any technical analysis on JBH. I'd appreciate if someone else could give it a shot though. (I've removed a lot of stuff to give you a relatively simply chart with price and volume). ONCE MORE WITH FEELING. AGAIN - NOT A RECOMMENDATIONView attachment 47403

Yeah looks to be going to sleep. Nothing here, next opportunity for me.

 

[Sir Osisofliver]

Yeah looks to be going to sleep. Nothing here, next opportunity for me.

Cool...why?

Remember my original statement back on the previous page. Why do we sometimes see a stock come down over a long period of time, that then rapidly increases in price? It's a dog isn't it? = It's going to sleep isn't it? I agree that at this moment you haven't got a technical trigger for entry, but we're trying to predict the next move (change in trend) with an increased probability for success due to the unwinding of the short position. So what can you see now?

Cheers

Sir O

 

[Trembling Hand]

Remember my original statement back on the previous page. Why do we sometimes see a stock come down over a long period of time, that then rapidly increases in price? It's a dog isn't it? = It's going to sleep isn't it? I agree that at this moment you haven't got a technical trigger for entry, but we're trying to predict the next move (change in trend) with an increased probability for success due to the unwinding of the short position. So what can you see now?

In fact I would say its actually increasing in its falls. Down trend lines have continually been broken and made steeper falls since 09. Its the kind of index futs chart I would bottom pick on an intraday basis but they, Index futs, on that time frame don't go to sleep, stock ones do.

The only technical thing I would say about it is there is less and less volume as price falls. This could lead into a lack of supply when the shorts do try and cover. They may find no offers above them to cover. That can lead to very nasty moves...... when it happens.

 

[Sir Osisofliver]

In fact I would say its actually increasing in its falls. Down trend lines have continually been broken and made steeper falls since 09.

I would too. In fact I'd go as far to say that...is a compounding pattern, and the increases are part of that compound curve...which is a fractal or chaotic expression of price. Is this relevant to your conclusion?

Its the kind of index futs chart I would bottom pick on an intraday basis but they, Index futs, on that time frame don't go to sleep, stock ones do.

So why is your analysis telling you that this stock will "sleep" - I assume this means a neutral trend sideways for quite a while? What features lead to this conclusion?

The only technical thing I would say about it is there is less and less volume as price falls. This could lead into a lack of supply when the shorts do try and cover. They may find no offers above them to cover. That can lead to very nasty moves...... when it happens.

Ahh the entry trigger...As you can see...we haven't got to that part of the process of system building yet TH. We're building up to it and getting closer though so we could probably begin some discussion on it. Your comment above is interesting. Does this volume indicate a potential bearish divergence? How would you determine if this is the case?

So what kind of technical trigger would you want to see? A break of a previous high pivot? A break of a minor or major trendline? An increase in volume that accompanies one of those features? A fib retracement of 100% +/- 5%, since we've seen a break of the 38.2, 50 and 61.8% retracement levels?

TH - I hope you see that I'm not picking on you...just trying to draw out the benefit of your experience for the newer forum members. In this case we are looking at an example which has not yet displayed the price behaviour I discussed back on page 31. So what price features would you want to see in JBH which would make you go long?


Cheers

Sir O

 

[sinner]

So what price features would you want to see in JBH which would make you go long?

Hope I can input.

At minimum:
* higher swing low than the last swing low

But prefer:
* first pullback (lowest close in N (where N between 3-10) days) after a breakout (highest close in N*10 days)
* trend signal (MA crossover or similar)
* breakout itself

If you are expecting rapid movement then it's probably safer to enter at least a portion of the position on a trend/breakout signal so that you don't get left behind waiting for a pullback.

 

[Trembling Hand]

So why is your analysis telling you that this stock will "sleep" - I assume this means a neutral trend sideways for quite a while? What features lead to this conclusion?

Because thats the pattern that I think it will show at best. it looks like another sp thats heading to zero rather than hero up to this point.

Does this volume indicate a potential bearish divergence? How would you determine if this is the case?

Bearish divergence? huh? It looks to be bullish..... maybe. It is to me showing a lack of supply as price falls. Not much more. Whether that eventually turn into demand without corresponding supply at the moment its just a guess. Thats what makes me think its probably going nowhere. Not a lot of increasing supply nor demand.

So what kind of technical trigger would you want to see? A break of a previous high pivot? A break of a minor or major trendline? An increase in volume that accompanies one of those features? A fib retracement of 100% +/- 5%, since we've seen a break of the 38.2, 50 and 61.8% retracement levels?

For me a decrease in volume AND price rising would signal end of selling. I reckon that would get me in before any sort of HH or retracement etc.

 

[Sir Osisofliver]

Because thats the pattern that I think it will show at best. it looks like another sp thats heading to zero rather than hero up to this point.

And until a trigger occurs I would agree our action is to merely maintain it on a watchlist.

Bearish divergence? huh? It looks to be bullish..... maybe. It is to me showing a lack of supply as price falls. Not much more. Whether that eventually turn into demand without corresponding supply at the moment its just a guess. Thats what makes me think its probably going nowhere. Not a lot of increasing supply nor demand.

Bah Humbug! Could have sworn I wrote bullish divergence. Damn fingers not typing what my brain tells them to. Typo - sorry for any confusion.

For me a decrease in volume AND price rising would signal end of selling. I reckon that would get me in before any sort of HH or retracement etc.

So can you scan for a) a decrease in volume; and crosscheck against b) a defined level of price rise? E.G A 2 standard deviation break to the upside?

Cheers

Sir O

 

[sinner]

For me a decrease in volume AND price rising would signal end of selling. I reckon that would get me in before any sort of HH or retracement etc.

Fair enough, but can you quantify that for us so we actually know what you're referring to? I mean, is "a decrease in volume AND price rising" on a 1 day basis so you'd buy the 'trigger' on the next up day with lower volume than the previous day...or are you describing something more fuzzy like the volume trendline is down while the price trendline is up?

 

[Sir Osisofliver]

Hope I can input.

All Are welcome Sinner.

At minimum:
* higher swing low than the last swing low

A purely visual indicator Sinner. Is it possible to construct a robust trigger from that? I mentioned before one of the things I look for is minimal time for execution of the system. This might suit your system, you might be willing to spend the time to look for this characteristic (or others) in the price action. I also said that the same idea can generate different methods of execution. Would the above (being a more discretionary decision process) suit your trading style?

But prefer:
* first pullback (lowest close in N (where N between 3-10) days) after a breakout (highest close in N*10 days)
* trend signal (MA crossover or similar)
* breakout itself

If you are expecting rapid movement then it's probably safer to enter at least a portion of the position on a trend/breakout signal so that you don't get left behind waiting for a pullback.

1) First pullback - after a break of a major/minor trendline. (Horizontal or descending I assume). I'm going to ask you a probability question here Sinner. How frequently does a stock break from a trendline like you've described and not experience a pullback/retracement until partway through the completed emerging structure? What I mean is you may be looking for a pullback that does not happen early in the new trend, especially if the shorts unwind quickly.

2) This is a good one, easy to scan for, what are your inputs? 10/20 12/26 SMA?

3) breakout of what exactly?

4) Last comment. Once again you've defined something that might be unique to yourself (and also might introduce a risk). You are willing to give up some potential profit for greater certainty. This has introduced the risk that the target moves so quickly that you emotionally assume that the horse has bolted. This is why it's important to ensure that our trigger is a defined one...so that we can test against it and have a degree of expectancy.

Cheers

Sir O

 

[sinner]

A purely visual indicator Sinner. Is it possible to construct a robust trigger from that?

It's not purely visual at all! Definitely possible to construct a trigger. I have a piece of code (IF low > low[1] AND low[1] < low[2] THEN curswinglow=low[1]; swinglows[N]=curswinglow ELSE swinglows[N]=NULL) that measures the high/low price for a swing high/low and keeps it in two arrays, so you can easily check whether the current swing low (what you call a pivot low) is higher than the last one in the array. You can also do things like measure the distance of the current swing low to the N day low to establish the shallowness/depth of the pull-back.

1) First pullback - after a break of a major/minor trendline. (Horizontal or descending I assume). I'm going to ask you a probability question here Sinner. How frequently does a stock break from a trendline like you've described and not experience a pullback/retracement until partway through the completed emerging structure? What I mean is you may be looking for a pullback that does not happen early in the new trend, especially if the shorts unwind quickly.

Trendline? I dunno where you got that from. Unless you count an 'highest close in N days' to be a trendline setup. If I replace the word trendline with breakout in your question, the answer is about 30%.

2) This is a good one, easy to scan for, what are your inputs? 10/20 12/26 SMA?

For newbies sake, I would suggest using 'logical' values for example the SMA of 1 week worth of prices (5 SMA) crossing the SMA of 1 months worth of prices (20 SMA). But this isn't realistic for me, I use adaptive techniques to get an aggregate short term signal independent of any particular parameter.

3) breakout of what exactly?

Highest close in N days. So a breakout of the N-1 day highest close.

4) Last comment. Once again you've defined something that might be unique to yourself (and also might introduce a risk). You are willing to give up some potential profit for greater certainty. This has introduced the risk that the target moves so quickly that you emotionally assume that the horse has bolted. This is why it's important to ensure that our trigger is a defined one...so that we can test against it and have a degree of expectancy.

100% agree with what you say here, although I haven't 'defined' anything, simply suggesting that if your target is trades which "come down over a long period of time, that then rapidly increases in price" then use the mechanical setup which is most captures the most alpha off those trades. No point waiting for a pull-back on rapid price increases IMHO.

 

[Sir Osisofliver]

It's not purely visual at all! Definitely possible to construct a trigger. I have a piece of code (IF low > low[1] AND low[1] < low[2] THEN curswinglow=low[1]; swinglows[N]=curswinglow ELSE swinglows[N]=NULL) that measures the high/low price for a swing high/low and keeps it in two arrays, so you can easily check whether the current swing low (what you call a pivot low) is higher than the last one in the array. You can also do things like measure the distance of the current swing low to the N day low to establish the shallowness/depth of the pull-back.

So the answer to that question is...yes I can construct a robust indicator from that.

Trendline? I dunno where you got that from. Unless you count an 'highest close in N days' to be a trendline setup. If I replace the word trendline with breakout in your question, the answer is about 30%.

Ahh I see what's going on. Without going into massive detail Sinner, there's some specific probability that is associated with fractal structures...certain types of trendlines confirm to fractal structures, (Multi-touch trend-lines) which give greater probability of outcome when broken. I assumed you meant that...my apologies for any confusion, I'm responding when I can get a break from doing other stuff. From what you've said then, you're simply looking at an increasing price movement, (which would appear at the change in trend point from a descending trend pattern. What's the advantage of using your metric as opposed to a Std Dev break? I assume you find it easier to scan for?

For newbies sake, I would suggest using 'logical' values for example the SMA of 1 week worth of prices (5 SMA) crossing the SMA of 1 months worth of prices (20 SMA). But this isn't realistic for me, I use adaptive techniques to get an aggregate short term signal independent of any particular parameter.

Hmm... I don't like to encourage the use of a bad indicator. Using a 5/20 SMA....not as good as using a cross period like a 12/26 SMA, precisely because a 12/26 doesn't correspond to a week/month. But hey, this is your system, use what works for you!

Highest close in N days. So a breakout of the N-1 day highest close.

Right, so you're creating a lagging signal that shows the change from the downward trend. Presumably you can then finetune the N value in your testing phase. I like it.

100% agree with what you say here, although I haven't 'defined' anything, simply suggesting that if your target is trades which "come down over a long period of time, that then rapidly increases in price" then use the mechanical setup which is most captures the most alpha off those trades. No point waiting for a pull-back on rapid price increases IMHO.

"...then it's probably safer"...to me at least, is an emotional "feel", not a statistical statement. It's something we need to be careful of and to make sure we challenge that kind of thinking. To me it often predefines things emotionally, "Stocks that look like JBH are dogs", "It's probably safer to enter after x conditions". We're often correct, but the important thing to understand is why, or more importantly, why we got it wrong. From your comments above...you get it...and to me it's a trade-off between potential alpha and increased expectancy to determine where our system entry exists.

Cheers

Sir O

 

[tech/a]

Hi there.
Had a chance to read through " Stuff ".
I don't wish to add only that we are looking for a measurable set of conditions
So this can be tested.
It is not known whether the conditions chosen are indeed " correct " until placed into a formula which can be tested.

T/H has suggested a low volume rise could signal a change of sentiment.
Tech would suggest that Shorts would be un wound in selling volume not buying volume.
So I would expect any turn to com after some exhaustion of selling.
All of these hypothesis have to be tested against like stocks past data.

Experience has also shown me that prolonged down moves rarely " v " bottom and are likely to rally then fail once again.
Consolidation --- sideways tends to pre cursor sustained up moves.
Mean reversion in my experience works best in corrective up moves--- unless your adding or entering shorts.

In the end it will work or not in the formula tested.

 

[Trembling Hand]

Tech would suggest that Shorts would be un wound in selling volume not buying volume.

Yeah agree but I think we were looking for a short squeeze set up. That will clearly not happen when there is still sellers.