Australian (ASX) Stock Market Forum

qldfrog weekly Skate inspired system

Jump to new
elbee said:
I have recently switched from Amibroker to Realtest, partly because of the problem you have pointed out.
Realtest allows you to specify the maximum number of setups that can be considered for entry on any one day so if one packet is above the entry limit it will not be replaced by another with a lower position rank.
Click to expand...
thanks interesting point.obviously make sense but imagine me suggesting this on AB forum :)
This realtest gets more and more interesting; how hard was it to switch AB to RT, and how easy was it to do the equivalent of AB Explore? If I may ask @elbee ?
 
qldfrog said:
thanks interesting point.obviously make sense but imagine me suggesting this on AB forum :)
This realtest gets more and more interesting; how hard was it to switch AB to RT, and how easy was it to do the equivalent of AB Explore? If I may ask @elbee ?
Click to expand...
The switch is relatively easy, RT is well documented and there is a helpful user forum. The developer is very receptive to suggested amendments, a contrast to AB where the developer is openly hostile.
There is a scan facility that works the same as AB's explore as well as the facility to produce a file containing the next day's order in a format that can be loaded straight into an Interactive Broker's basket file.
 
qldfrog said:
AB help required:
In real life, some shooting start buys are missed as my:
last close+3% (for example) buy values at open are overshot;
I would like to model that in backtests to ensure backtests are as near real results as possible:
so a Buy if O<ref(C,-1)*1.03
I can do that but in cases like that we want to ensure this buy is missed not just replaced by the next one based on the positionscore
so need to reduce on that day the total number of buys by one
And obviously this should not interfere with explore results;
If someone has already coded this , I would be grateful, otherwise i will do the yard yakka...
And share with you on demand
Click to expand...
Hey Frog,
Not sure if you're chasing the exact AFL code or general approach. Think you have a level of familiarity with AFL so I will describe a general approach you might find helpful. If you want some AFL let me know and will be happy to post up some general code for you to mess with. But here is how I'd do it.

Assuming your system using a fixed number of positions (e.g., 20) the first step is to determine the current number of open positions you have. There's a few ways to do this in AB but personally I do that with the custom backtest. Once you pull the number of open positions I'd define a variable POS_TO_TAKE as initially being your max open position - number of open positions. At this point I would reduce POS_TO_TAKE by 1 when you encounter the condition Open > ref(c,-1)*1.03. I'd use the POS_TO_TAKE in a loop to control the number of buys you take. When POS_TO_TAKE is reduced by 1 it would essentially simulate a missed trade, which what I think you're trying to do. You'll have to run this as an overall loop so POS_TO_TAKE is reset for each day (bar).

Hope this makes sense. Good luck
 
MovingAverage said:
Hey Frog,
Not sure if you're chasing the exact AFL code or general approach. Think you have a level of familiarity with AFL so I will describe a general approach you might find helpful. If you want some AFL let me know and will be happy to post up some general code for you to mess with. But here is how I'd do it.

Assuming your system using a fixed number of positions (e.g., 20) the first step is to determine the current number of open positions you have. There's a few ways to do this in AB but personally I do that with the custom backtest. Once you pull the number of open positions I'd define a variable POS_TO_TAKE as initially being your max open position - number of open positions. At this point I would reduce POS_TO_TAKE by 1 when you encounter the condition Open > ref(c,-1)*1.03. I'd use the POS_TO_TAKE in a loop to control the number of buys you take. When POS_TO_TAKE is reduced by 1 it would essentially simulate a missed trade, which what I think you're trying to do. You'll have to run this as an overall loop so POS_TO_TAKE is reset for each day (bar).

Hope this makes sense. Good luck
Click to expand...
yes exactly what I have in mind, but this is virgin territory as I have never done dynamic check of open positions, etc and not sure of backtest engine behaviour
 
qldfrog said:
yes exactly what I have in mind, but this is virgin territory as I have never done dynamic check of open positions, etc and not sure of backtest engine behaviour
Click to expand...
just need to be able to focus, between sorting some components for arduino, testing AB, checing RealTest and doing the daily explore and system homework/Tax return...
And mind focused on how the hell I am going to be able to get out of Aus and keep my assets while closing trusts and company, ASAP seeing the general public response here vs what is happening
Should probably close all the daily system while I sort my personal emergency exit and stop loss
 
qldfrog said:
yes exactly what I have in mind, but this is virgin territory as I have never done dynamic check of open positions, etc and not sure of backtest engine behaviour
Click to expand...
It’s pretty straightforward—dynamically determining the number of open positions is a very commonly asked question so search the AB forum (plenty of posts about it there) or just google it.
 
@qldfrog I'm expecting to see a good week from your systems. Your fingers should be flying across the abacus tonight.

Forget "the Reset", the conspiracies and allow yourself to revel in the satisfaction that market success brings. We have to take every opportunity to revel because the market doesn't give us too many. Don't disappoint me.
 
peter2 said:
@qldfrog I'm expecting to see a good week from your systems. Your fingers should be flying across the abacus tonight.

Forget "the Reset", the conspiracies and allow yourself to revel in the satisfaction that market success brings. We have to take every opportunity to revel because the market doesn't give us too many. Don't disappoint me.
Click to expand...
was actually brilliant indeed for the systems but today was a step back:lost today, VERY good previous 2 sessions I am doing the tally
 
peter2 said:
@qldfrog I'm expecting to see a good week from your systems. Your fingers should be flying across the abacus tonight.

Forget "the Reset", the conspiracies and allow yourself to revel in the satisfaction that market success brings. We have to take every opportunity to revel because the market doesn't give us too many. Don't disappoint me.
Click to expand...
Jigsaw results:
lost nearly $10k on monday then a few days of good returns and then -$2.7k today!!!
AJX lost 1.3k(-19%) and BCB one of my stars gave back nearly $1k, add a few CXO AGY losing 7 or 8% or the winners did not make up for the losses.Basically SPP galore CXO, POS, SYA, BCB and that is usually neither a good sign for the immediate SP, nor is it for the market trend.SPP often at the top of the curve

At the end of the week ; a week where definitively materials and small caps are back[ which is good for the systems]
the dailies:
Flying Bat:
XAO, fully invested -$2.8k or -4.9% very bad, we will go back to that:

Guppy:
+$1.7k or +3.9%
Only.65% invested as I moved to half packets at the end of july after weeks of capital grinding .talk about a bad idea..
We have had 2 good weeks in a row one at 2.9% weekly, one at 3.9% weekly..as I keep a daily eye on Guppy, the last 10 sessions were quite good so will move back to full packets..maybe suited back to conditions.

volatility ASX:fully invested +$1.3k or +2.6%
volatility US:fully invested +$0.3k or +1.7%

Weekly:
QF sec: +$2.1k or +3% sadly only 50% invested

QFDuc: +$0.6k or +0.6% poor returns here 20% loss on PEK, closing gains and big loss on PDI
overall so disappointing after such a great perf last year


overall:+$3.2k or around 1% with 25% cash as we start ramping up again so basically just the XNT after cash is taken into account
following post on Flying Bat
 
Checked Flying Fox vs backtests..kind of match.But I found one issue this week: some of my buys disappear on the next day explore..usually a sign of look forward code issue
I will go into half packet there until I find the reason; must be something obvious but a bit busy now
So FB down to half size,
Restoring Guppy full size for the daily we should have made a bonza....
QFSec starting to gain traction at last and QFDuc disappointing this week but that can happen
 
peter2 said:
@qldfrog I'm expecting to see a good week from your systems. Your fingers should be flying across the abacus tonight.

Forget "the Reset", the conspiracies and allow yourself to revel in the satisfaction that market success brings. We have to take every opportunity to revel because the market doesn't give us too many. Don't disappoint me.
Click to expand...
I hope I did not disappoint you @peter2 , not stellar results
but on the way back
The Flying Bat is a lame duck in disguise I am afraid...

Sadly , the Reset and its trail will have bigger implications for me/all of us and my finances than making x% on a system today so that does not help to motivate system work:
how can I physically get the F out of this country and keep some of what i built in the last 30y here.
A lot of unwinding and few certitudes.
If I am wrong cost is minimal and I will travel abroad; if I am right: priceless
I suspect the German Jews who focused too much on the recovery from the 1929 crash paid it dearly in the following decade.
OK back to the trend systems
 
Argh. Yes, I'm disappointed to read the latest update. I fully support your trading efforts and because their yours, you. I respect all investors and traders who work diligently in this business. It tough emotionally. We learn so much about ourselves as we hit our personal boundaries and fall victim to our personal biases.

I suspect that there's a few details that have escaped your notice or that you don't understand when back-testing and then implementing your systems. I can't give you an example because I don't know your systems. It's just an impression I get when reading your updates. It's like hearing a 8 cylinder engine running on 6. Something needs a tune up.

A theoretical example, it could be that your entries are late into the current price swing. This will be disastrous when trading micro cap price rallies in daily charts. I've thought that many of @Skate's entries were a little late into the trend but the weekly charts seem to offer more protection. Daily charts don't provide the same protection. eg AJX an entry higher than 0.073 is too late. With the close at 0.079 your trade should still be profitable. Sure, it's lost some open profit but your system back test should give you the confidence that this is OK and across a larger batch of trades your trade mgt earns more with the current settings.

Your last post indicates that you've got much larger concerns than your trading business. They may be having a negative impact on the discretionary aspect of your systems. eg turning them on/off, increasing/decreasing position sizes.
 
peter2 said:
Argh. Yes, I'm disappointed to read the latest update. I fully support your trading efforts and because their yours, you. I respect all investors and traders who work diligently in this business. It tough emotionally. We learn so much about ourselves as we hit our personal boundaries and fall victim to our personal biases.

I suspect that there's a few details that have escaped your notice or that you don't understand when back-testing and then implementing your systems. I can't give you an example because I don't know your systems. It's just an impression I get when reading your updates. It's like hearing a 8 cylinder engine running on 6. Something needs a tune up.

A theoretical example, it could be that your entries are late into the current price swing. This will be disastrous when trading micro cap price rallies in daily charts. I've thought that many of @Skate's entries were a little late into the trend but the weekly charts seem to offer more protection. Daily charts don't provide the same protection. eg AJX an entry higher than 0.073 is too late. With the close at 0.079 your trade should still be profitable. Sure, it's lost some open profit but your system back test should give you the confidence that this is OK and across a larger batch of trades your trade mgt earns more with the current settings.

Your last post indicates that you've got much larger concerns than your trading business. They may be having a negative impact on the discretionary aspect of your systems. eg turning them on/off, increasing/decreasing position sizes.
Click to expand...
Thanks Peter
I hope this LONG post will not bore people and be instructive..and help me as i write it down..
There are definitively areas in the dailies FB and Guppy which are really failing:
1) price of entries;
I used to assume that the buy at next open is a done deal..not at all;
with these micro caps, even if it opens at 0.086c my buy up to 0.087 entered the night before might be missed and I then increase a bit or miss the entry
a lot of the buy at open also fails as the price shots well above my 3% margin from last close, and I can not (broker rejected) place higher value, these are often the ones making a lot of the profit;

the dailies, even in backtests are a game number, sell buy sell buy with a fine margin, even a few missed entries are enough to move the balance into negative.The whole purpose of my dailies is to get out quickly at the next crash but so far has cost money to no advantages

2) Another area I am really NOT happy is the ranking of my selected buys;
Never been happy or found a method i like
So far best I found is a mix of TSIndicator and SP (softened)
dailies are to be quickly invested and quickly out, so my buy criteria should be loose and they are; I get dozens of buy each days, then rank them..but ranking is a mystery; if you use smaller is better[ which i used to], great in backtest, but you quickly realise you circle between a set of half dozen of penny stocks bleeding to death by multiple cut so not right

About the late entry potential, not sure: to take the AJX example I entered in a weekly on the 2/08 at 0.067 so still a great 18% profit at 7.9c tonight but yesterday it was at 9.7c....
It is more that I exit too soon for these type of stocks..then buy again 2d later higher, and exit again after a small loss; rince and repeat
yet statistically, this is the way to go


Overall the AB backtests match the tradings ..which is good and convenient
Have I been curve fitting? maybe..who knows

There are things for sure I can not understand :

as a general rule, if the ASX is gaining more than 0.3 or 0.4 % , I lose money;
anything stable or down is usually good for my systems; ???

my 11AM portfolio value is usually much higher than my end of day usually nearly 1% portfolio difference , so this is big
to the point I was trying to check if a buy at close, sell at open system would work
Today : sum of systems -$2.7k at close; this morning was +0.5k at 11AM-> here we have again the 1% portfolio variation, always the same way, even on my good days with flat asx, I go from +$4k to $1k
-> I am wondering is this is not a crude pump and dump daily activity, which would maybe mean big money offloading their shares lately ???
I do not remember experiencing that behaviour 1 y ago

For the better news:
FB current version is smoother..except for this week, so hopefully just a glitch or effect of the old code still affecting the portfolio
volatility systems going ok, and giving me some uncorrelated returns
The Bat started in May,
I give below the systems backtest charts for the current year
older version:
1628865390889.png
newer since mid July
1628865439819.png
So should I persist, probably and resist the gloom you mentioned.
This Reset taking place is definitively influencing my discretionary side, negatively
so let's step back and yes, I carry on with the BAT unchanged maybe i need to give my newer code system a couple months teething and refining...

Guppy: started last December, current version since mid June mehhh just missed the best 2 weeks of the year?
1628865978053.png

weekly:
Sector influenced: hum....
1628866112249.png

QFDuc did well last year but I probably toyed too much with it:
back to a May version since this month so basically a new start

was:(before last week)
1628866562560.png

now: and only since this week so in a way a new system
1628866395970.png

PS: all BT graph from 01/01/21, various initial investment
As a conclusion, I have to let the changed systems take ownership and probably give them more trust
so not playing around with packet sizes
It will take a few ticks to erase the old code effects

PS2, graph since 1/1/21 but yes I test my systems on more than this
 
Great post. I'm sure that post helped you sort through a few concerns that you may have had. You may have made some decisions by thinking through your written words.

(1) I'm not sure the 3% is appropriate for smaller priced stocks where one tick may be 3%.

I can't help you with the problem of your broker not accepting orders that they consider too far away from the last close. I use conditional order to catch breakouts and re-entries and the trigger can be much higher than the close. These orders don't go into the market though and create limit orders when triggered.

Have you considered scanning the market and entering (buying) just before the close instead of the open next day? I scan on the 3pm data that I get at 3:20pm. I've got 40 min to scan and make my decisions before the close.

(2) Can't help you with ranking as I'm not a system trader. I use discretion to select the trades from those available.

You can apply a time stop to minimise the erosion of the penny stock bleeders.

AJX = good entry. The only way get the price spike exits is to target them with your trade management. This may or may not earn more profit overall. Your back testing should indicate the best mgt approach for your entries.

If the AB back tests match the trades then your doing your job. Tick.

I've noticed the same thing on many days. The portfolio value is usually higher in the early morning than at the close. Anecdotally I say it seems like 90%. I don't know what it is actually. It's so common that I don't want to know what it is in the morning because i'll be disappointed in the afternoon. I'll check the share prices but don't want to know the summary. LOL

The new Bat looks much better than the old one. Although your comparison periods are short. Compare over a longer period, although you know this.

Guppy. I'm curious as to the underlying tactics you're using in this system. I understand Guppy's work quite well and know his tactics should be a good robust producer. If it's not, that's a concern. There are two aspects to Guppy's trade entry, a setup and an entry trigger (confirmation). Both steps need to be defined for a trade to be valid.

The Duc's volatility entry is similar to my 1stBB entry. I know this produces too many entries. It may work better with a smaller universe based on price or volume traded or market cap or basic FA info. Ideas only.
 
qldfrog said:
I found one issue this week: some of my buys disappear on the next day explore..usually a sign of look forward code issue
Click to expand...
I guess you would know but this is handy to see those sneaky signals Tools>Bar Replay.

About the general stuff above, I excluded those below 10c (too volatile for me) plus it always made sense to me to use turnover as the ranking. Like you, I used 3% on entry but 5% on exit to place the entry on open.
 

Similar threads

sharks
Trading systems/methods guidance
Replies
14
Views
6K
rnr
R
MovingAverage
Emergency Exit for Weekly System
2
Replies
27
Views
5K
MovingAverage
MovingAverage
MovingAverage
Swing System
Replies
19
Views
5K
MovingAverage
MovingAverage
quanttraderx
Weekly Breakout Strategy For QQQ ETF
Replies
17
Views
5K
KevinBB
K
Linus van Pelt
System optimisation: IS training vs. OOS testing, and Paper Trading
Replies
10
Views
2K
Garpal Gumnut
Garpal Gumnut
Top