Dump it Here

qldfrog · 12 April 2020

If trading a purely weekly system, you can have up to 5 down sessions
A DD of 20/25% would still be reasonably possible even if you trigger the gtfo on that friday night
And an average fall of 4 or 5 pc per session is not exactly unrealistic coming from our highs
The only way to reduce the risk is to have a daily SL ..i toyed with that idea in January i initiated then canned it and lost 20k minimum the following month.

Bazzi · 12 April 2020

qldfrog said:
If trading a purely weekly system, you can have up to 5 down sessions
A DD of 20/25% would still be reasonably possible even if you trigger the gtfo on that friday night
And an average fall of 4 or 5 pc per session is not exactly unrealistic coming from our highs
The only way to reduce the risk is to have a daily SL ..i toyed with that idea in January i initiated then canned it and lost 20k minimum the following month.

Hi @qldfrog Do you use SL in all trades now? Any maximum SL or a guideline you follow in general?

aus_trader · 13 April 2020

qldfrog said:
If trading a purely weekly system, you can have up to 5 down sessions
A DD of 20/25% would still be reasonably possible even if you trigger the gtfo on that friday night
And an average fall of 4 or 5 pc per session is not exactly unrealistic coming from our highs
The only way to reduce the risk is to have a daily SL ..i toyed with that idea in January i initiated then canned it and lost 20k minimum the following month.

I remember you mentioning about this earlier actually. The reasons were valid at the time and I agreed with the reasoning that positions can be exited too early if there was an intra-week low that reversed and closed higher by the end of the week. This time it happened so quickly that an over-riding GTFO type of exit was probably the only way to reduce the draw down as Peter2 mentioned earlier.

ducati916 · 13 April 2020

peter2 said:
1. This thread has seen another interesting twist. One that both surprises and disappoints me. The major reason for advocating a fully systematic trading approach is to minimise the emotional interference that plagues discretionary traders.

2. Currently the emotions are in control of the systematic trader(s). They're doubting the effectiveness of their systems. They're looking for better exits strategies. The recent market behaviour was so unexpected that they're reacting purely on emotions. May I shout, "STOP IT, GET A GRIP".

3. IMO the PANDA system is one of the best I've seen. It works brilliantly in mildly bullish markets and kills it in fully bullish markets. It's a seriously good system. So is your Hybrid system. How can you now be in any doubt?

"it fell short of my expectations" : I'd suggest that your expectations were wrong then. Didn't you back test this system to your satisfaction? Didn't you try to break it? Didn't you test it to see what would happen if the market suddenly tanked?

Looking at the 2020 test results shows me exactly what I'd expect to see when the underlying market falls 35% suddenly. A 17% DD is within the normal range for a trend following DD. I remember remarking on some back test results posted her earlier that the DD shown was too low. It was suggested that it was a timing issue. Portfolios started earlier wouldn't have such a large DD. Now we're seeing the larger DDs.

4. All long only equity systems have been thumped by this sudden selloff. Trend following systems require wide trailing stops (large downside exposure) for maximum reward. The market price action triggered the GTFO exit and we followed our plan.

1. I would argue that such introspection and reflection on the various systems and underlying philosophy of system based trading are extremely valuable. Without it, progress grinds to a halt. From what I have seen, discipline was maintained, the systems followed: discipline is an emotional state.

2. The emotion of disappointment is also valuable in that it drives the discipline of looking for improvement. The recent market fall highlighted a weakness in the system(s). The important point here is that these 'types' of markets are rare. To date 1987 and 2020. The search for improvement that is currently underway I agree is (IMO) misguided in that it is addressed intrinsically to the system, when in reality it should be addressed extrinsically to the market [my previous (a) and (b)].

3. I agree. In the right conditions, it printed money. Thus, the fix is not in the system, it is in the 'conditions', ie: correct market conditions, which is extrinsic to the system. Now given the rarity of the wrong conditions, this should be eminently fixable and well worth the time in considering (searching for) answers that result in improvement.

4. I would agree with this through observation of the various systems disclosed here on ASF.

I already have the fix. The problem is whether it can be coded. I am going to converse with Mr Skate and see if he can code it. If he can, it may well be problem solved.

jog on
duc

qldfrog · 13 April 2020

Bazzi said:
Hi @qldfrog Do you use SL in all trades now? Any maximum SL or a guideline you follow in general?

No as i do not believe the next down will be as brutal nor do i have enough invested at this stage.
As we have more time at home and a new interesting period to backtest against, i am fine tuning my 2 systems.live they are but hardly invested so this is a good time for fine tuning
To get the best average return, SL have to allow 20 -25pc fall from peak on a daily view..not much which is not caught by existing exits

Rsthree · 13 April 2020

SL = ?? Probably obvious, but not forgetting that we are in a beginners thread.

rederob · 13 April 2020

Rsthree said:
SL = ?? Probably obvious, but not forgetting that we are in a beginners thread.

Stop Loss

Rsthree · 13 April 2020

DOH...totally obvious I was looking for a more complex answer.

frugal.rock · 13 April 2020

Another meaning could be
SL = screamingly late

Am wondering if/ how dark orders and centre point trading methods could be taken advantage of by the coders?
From my understanding, the systems/ strategies used for backtesting could be forward implemented into special dark orders with the right order coding.

The trick as I see it is getting the order coding correct.
From my basic understanding, orders can be coded with probably pretty much any conditions nests you can think of... and am thinking that some of you on this thread would be capable of doing it.

Personally, have been working on identifying algo/bot trading and the desired outcome of the algo, intended to take advantage of the money flow.
As most here know, the DOM (depth of market) that is visible is just the tip of the iceberg.
The rest of the burg is what interests me, the darkness...

@Skate, the buy sell pressure indicator you mentioned (Karthik?) and posted charts of, look interesting, but seem to be stock specific?
(Was thinking of using indices as a combination. A gross tide change indicator.)
In any case I don't have any software as yet to be able to use it.
My problem.

F.Rock

Skate · 13 April 2020

frugal.rock said:
Am wondering if/ how dark orders and centre point trading methods could be taken advantage of by the coders?
@Skate, the buy sell pressure indicator you mentioned (Karthik?) and posted charts of, look interesting, but seem to be stock specific?

@frugal.rock I've done some research how to identify Dark Pools Patterns using Technical Analysis but it's above my pay grade. What I have found - Pool Chart Patterns are mostly bottoming formations that move erratically sideways, violently up and down causing most traders to be whipsawed out with losses, so this style of trading is better left to the big boys.

Here is an interesting fact
Did you know that the Flash Crash of May 2010 was caused by a simple error of the wrong order type being used. Instead of using a Time Weighted Average Price (TWAP) order, the trader used a Volume Weighted Average Price (VWAP) order. The order resulted in a spike in the "Volume" which then in turn triggered more and more orders thus causing the crash. It was as easy as confusing (TWAP) orders with (VWAP) orders - who would have thunk it !

Karthik Marar Buy and Sell pressure indicator
The Buy and Sell pressure indicator & smoothed histogram is not stock specific as it handles all markets no matter the periodosity used. I have attached a daily chart for the Dow Jones as well as the All Ordinaries as examples.

Summary
I've not been able to capitalise using the "Buy and Sell pressure indicator" in any of my trading strategies - its not been for the lack of trying though.

The Dow Jones

The All Ordinaries

Skate.

Newt · 13 April 2020

Skate said:
@frugal.rock
Karthik Marar Buy and Sell pressure indicator
The Buy and Sell pressure indicator & smoothed histogram is not stock specific as it handles all markets no matter the periodosity used. I have attached a daily chart for the Dow Jones as well as the All Ordinaries as examples.

Summary
I've not been able to capitalise using the "Buy and Sell pressure indicator" in any of my trading strategies - its not been for the lack of trying though.

Skate.

You've shared this a few times Skate - wondered if/how you employed it.

Newt · 13 April 2020

qldfrog said:
If trading a purely weekly system, you can have up to 5 down sessions
A DD of 20/25% would still be reasonably possible even if you trigger the gtfo on that friday night
And an average fall of 4 or 5 pc per session is not exactly unrealistic coming from our highs
The only way to reduce the risk is to have a daily SL ..i toyed with that idea in January i initiated then canned it and lost 20k minimum the following month.

In the spirit of what Peter2 said don't think you should beat yourself up QF. As alluring as "intrabar" stops are (and Amibroker makes them very possible), the stark reality is that trend following over the medium to long term benefits from wide trailing stops. The short term benefit is usually far exceeded by the extra profit potential of wide stops in the long run.

We're experiencing some extremely unusual circumstances lately, and modern financial markets are so automated and quick to respond the positive feedback loops have been crazy. I think I read somewhere recently US employment weekly data last month spiked to 30 SDs. 6 SDs is getting close to 1/million so goodness knows what 30 SDs is.

Pretty tough times to have every angle covered.

Newt · 13 April 2020

ducati916 said:
1. I would argue that such introspection and reflection on the various systems and underlying philosophy of system based trading are extremely valuable. Without it, progress grinds to a halt. From what I have seen, discipline was maintained, the systems followed: discipline is an emotional state.

2. The emotion of disappointment is also valuable in that it drives the discipline of looking for improvement. The recent market fall highlighted a weakness in the system(s). The important point here is that these 'types' of markets are rare. To date 1987 and 2020. The search for improvement that is currently underway I agree is (IMO) misguided in that it is addressed intrinsically to the system, when in reality it should be addressed extrinsically to the market [my previous (a) and (b)].

3. I agree. In the right conditions, it printed money. Thus, the fix is not in the system, it is in the 'conditions', ie: correct market conditions, which is extrinsic to the system. Now given the rarity of the wrong conditions, this should be eminently fixable and well worth the time in considering (searching for) answers that result in improvement.

4. I would agree with this through observation of the various systems disclosed here on ASF.

I already have the fix. The problem is whether it can be coded. I am going to converse with Mr Skate and see if he can code it. If he can, it may well be problem solved.

jog on
duc

Duc, would love to know if this ever comes to something. Knitting such a diverse set of macro/commodity/stock data into a cohesive filter would certainly be something. Suspect it would envitably involve observations, experiences, opinions that were heavilty dependent on the experience of the observer - not something many people could do no matter how much raw data you started with. You could go mad trying to do multi-variate regressions or goodness knows what.

qldfrog · 13 April 2020

rederob said:
Stop Loss

Indeed apologies for using the acronym without first explanation
Phone typing is not an easy sport

frugal.rock · 13 April 2020

Can the Karthik indicator average all ASX indices on the 1 chart?
Really just looking for a overall indicator.

The dark pool thing still interests me. Have seen stocks with selling pressure that largely absorb the pressure with little to no price movements, which to me indicates hidden support.
Eg, A smaller priced stock may absorb relatively large sell downs with no apparent support. Will trade the daily average volume at one price all day, with apparent support vacant.
Large volume trades go through, the buying volume- usually only a pathetic visible amount, remains similar after a trade size that has the capability of knocking the SP down a tick or 3 if allowed. Just some observations.

F.Rock

qldfrog · 13 April 2020

One more point to add regarding stop loss:SL
Used in that context during drastic movement, it is easy to actually miss the falling knives
If you trigger your SLat $n, what is your sell order at?
2pc below?, more? From previous old experience on daily trading attempts you have to go a wide berth from trigger point to be executed and are often getting the low of the bar
You would need to add consequent slippage in your backtest
Disclaimer: i always assume a weekly system with present orders and potentially SL, with no human interaction during the week /day

Skate · 13 April 2020

Newt said:
You've shared this a few times Skate - wondered if/how you employed it.

The logic
The best solution so far was applying the Buy and Sell pressure indicator as a directional price indicator that displayed better than average volume - sometimes it worked perfectly, at other times it produced many false signals. Frankly it's an unreliable indicator & I haven't found any trading value using this indicator to date.

How do I use the Buy and Sell pressure indicator these days
I use the "Buy and Sell pressure indicator" as a "snooping tool" to satisfy my curiosity judging market sediment & at times understanding the sediment of individual stocks.

Newt said:
In the spirit of what Peter2 said don't think you should beat yourself up QF

# Easier said than done.

Skate.

ducati916 · 13 April 2020

Newt said:
Duc, would love to know if this ever comes to something. Knitting such a diverse set of macro/commodity/stock data into a cohesive filter would certainly be something. Suspect it would envitably involve observations, experiences, opinions that were heavilty dependent on the experience of the observer - not something many people could do no matter how much raw data you started with. You could go mad trying to do multi-variate regressions or goodness knows what.

Well I have just PM'd Mr Skate to see if he can code it up. Actually, all but 1 should be relatively straightforward. I guess we wait and see if he can.

As to the issue of 'objectivity' (coded) and subjectivity (observation), there may be issues, hence my initial challenge to Mr Skate (to see if it is possible) to code it, taking it from discretionary to systematic.

jog on
duc

qldfrog · 13 April 2020

Skate said:
# Easier said than done.

Indeed, so far rational thinking still winning so was tempted but did not go that way, more generic weekly exits checks in GTFO style

Trendnomics · 13 April 2020

peter2 said:
This thread has seen another interesting twist. One that both surprises and disappoints me. The major reason for advocating a fully systematic trading approach is to minimise the emotional interference that plagues discretionary traders.

Currently the emotions are in control of the systematic trader(s). They're doubting the effectiveness of their systems. They're looking for better exits strategies. The recent market behaviour was so unexpected that they're reacting purely on emotions. May I shout, "STOP IT, GET A GRIP".

Respectfully @Skate I was disappointed to read of your doubts regarding this system.

IMO the PANDA system is one of the best I've seen. It works brilliantly in mildly bullish markets and kills it in fully bullish markets. It's a seriously good system. So is your Hybrid system. How can you now be in any doubt?

"it fell short of my expectations" : I'd suggest that your expectations were wrong then. Didn't you back test this system to your satisfaction? Didn't you try to break it? Didn't you test it to see what would happen if the market suddenly tanked?

Looking at the 2020 test results shows me exactly what I'd expect to see when the underlying market falls 35% suddenly. A 17% DD is within the normal range for a trend following DD. I remember remarking on some back test results posted her earlier that the DD shown was too low. It was suggested that it was a timing issue. Portfolios started earlier wouldn't have such a large DD. Now we're seeing the larger DDs.

All long only equity systems have been thumped by this sudden selloff. Trend following systems require wide trailing stops (large downside exposure) for maximum reward. The market price action triggered the GTFO exit and we followed our plan.

Boom! Best post on this whole thread!

I've been a long time reader of this thread, and two notable things have become apparent:

Most of the traders on here, still haven't FULLY accepted that systematic trend-following, will always involve losses and large draw-downs - the whole design basis of all systematic trend-following systems are to take losses consistently (<50% win rate) - this is what makes these systems so robust (you have to become comfortable with something that can be unbearable at times, to avoid emotional tinkering - in a few months it'll be my system's 7 year live-traded-non-stop-un-tinkered anniversary).
Most of the traders on here, still believes that running multiple trend-following systems will provide some form of diversification or overall draw-down protection (IMO: only more temptation for emotional tinkering). Concurrently, the profitability foundation of all TF systems, is grounded in a single statistical anomaly that occurs in price action (the real reason why all TF systems are profitable) - thus, running multiple TF systems would provide a negligible long term advantage (particularly over similar time-frames). Over my years of discovery and application of advanced mathematics (Engineer), I am certain that I have identified this statistical anomaly (simple proofs possible in Excel) - this has provided tremendous confidence in my system and has reduced emotional tinkering temptations. There is little mention of the statistical reasons, why TF systems are profitable, on this thread - it would be great to see, so I can compare some notes (peer comparison/validation).

Dump it Here

Similar threads

Privacy & Transparency

Privacy & Transparency