Dump it Here

[Beaches]

Out of interest Skate, would you mind showing us the backtest results for the system over the first week? You would expect it to match, it's good to check that it does.

A backtest can never match live trading. The point of a backtest is only to establish a positive expectancy. Think of the back testing as Phase 1 testing. If the system gives a positive expectancy in Phase 1, then you move to phase 2 testing which is live trading. initially, your live trading is still testing.

A backtest can never match real trading for many reasons.

Here is 1 very simple situation that can make backtesting differ from live trading. If we assume a stock has the following Buy bids and Sell offers during the preopen session (7:00am to 10:00am):

Buy Orders
2000 @ 5.21
5000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
2000 @ 5.18
2000 @ 5.17

Sell Orders
1000 @ 5.19
2000 @ 5.20
1000 @ 5.20
6000 @ 5.20
1000 @ 5.21
5000 @ 5.22
2000 @ 5.23
5000 @ 5.23
1000 @ 5.24
3000 @ 5.25

In the above example we will assume that the ASX algorithm sets the opening price at $5.20 at the opening auction. There are 4 Sell bids making a total of 10,000 shares selling at the opening price of $5.20 or less and 8 Buy bids making a total of 14,000 at the opening price of $5.20 or more. In this situation only the first 5 Buy bids (highlighted red) fit into the 4 Sell bids (highlighted red) of 10,000 shares available at the match price. These 5 highlighted Buy bids will get filled at $5.20 during the opening auction. The remaining 3 Buy bids of 3,000 shares at the opening price of $5.20 are known as the Surplus and will not get filled during the opening auction.

If immediately after the opening auction the trading price moves higher than the opening price of $5.20 and stays higher for the rest of the day, the 3,000 buy bids at $5.20 will never get filled and will be purged from the list at the end of the day if they have been entered as ‘good for the day’ bids.

From the above example, we can see that while the opening price was $5.20, some bids got filled but others did not. So not only can the system backtest get a different result from live trading, but of 2 people trading the same system, one may get filed and one mot filled which would give them different results based on where they are in the queue at the opening auction.

This is just one example and there are probably another dozen scenarios that could result in a difference between a backtest and live trading.

 

[qldfrog]

A backtest can never match live trading. The point of a backtest is only to establish a positive expectancy. Think of the back testing as Phase 1 testing. If the system gives a positive expectancy in Phase 1, then you move to phase 2 testing which is live trading. initially, your live trading is still testing.

A backtest can never match real trading for many reasons.

Here is 1 very simple situation that can make backtesting differ from live trading. If we assume a stock has the following Buy bids and Sell offers during the preopen session (7:00am to 10:00am):

Buy Orders
2000 @ 5.21
5000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
1000 @ 5.20
2000 @ 5.18
2000 @ 5.17

Sell Orders
1000 @ 5.19
2000 @ 5.20
1000 @ 5.20
6000 @ 5.20
1000 @ 5.21
5000 @ 5.22
2000 @ 5.23
5000 @ 5.23
1000 @ 5.24
3000 @ 5.25

In the above example we will assume that the ASX algorithm sets the opening price at $5.20 at the opening auction. There are 4 Sell bids making a total of 10,000 shares selling at the opening price of $5.20 or less and 8 Buy bids making a total of 14,000 at the opening price of $5.20 or more. In this situation only the first 5 Buy bids (highlighted red) fit into the 4 Sell bids (highlighted red) of 10,000 shares available at the match price. These 5 highlighted Buy bids will get filled at $5.20 during the opening auction. The remaining 3 Buy bids of 3,000 shares at the opening price of $5.20 are known as the Surplus and will not get filled during the opening auction.

If immediately after the opening auction the trading price moves higher than the opening price of $5.20 and stays higher for the rest of the day, the 3,000 buy bids at $5.20 will never get filled and will be purged from the list at the end of the day if they have been entered as ‘good for the day’ bids.

From the above example, we can see that while the opening price was $5.20, some bids got filled but others did not. So not only can the system backtest get a different result from live trading, but of 2 people trading the same system only one may get filed and would have different results based on where they are in the queue at the opening auction.

This is just one example and there are probably another dozen scenarios that could result in a difference between a backtest and live trading.

My understanding is that if there is more demand than offer during the auction, the price will increase and in the above example, the open price will NOT be 5.2 but 5.21 or whatever to ensure the situation you describe does NOT happen: open price is equal to price P where all sells down to P or below and buys up to P and above can be executed.I could be proven wrong but be disappointed
Nevertheless AB give you an indication if your orders might be unrealistic in real world execution:
see notice 802 below on a backtest execution:

This should not happen much with initial positions but if you day dream and play with 10M portfolio with 12 position, it will obviously happen and yes in that case the Backtest means shxt.
Right now, with 1k positions, I think we are safe

 

[qldfrog]

BTW, yes Backtest will differ from execution, but you should be able to understand and calculate it, otherwise something is wrong in your system code;
As a beginner, I make sure I compare both and track differences, both a fail safe and instructive past time..it is not always easy

 

[Skate]

I thought I might be missing something, I'm often missing something. But I still don't get the statement that Amibroker uses dumb maths. Amibroker will use what we tell it to use. Why can't we do that?

@Lone Wolf, I refer to "Dumb Maths" in relation to calculating static numbers. Amibroker lacks the intelligence to calculate fluid numbers, numbers that move from second to second, jocking for a position in the pre-auction is an experience in itself & Amibroker couldn't keep up with that. Amibroker is a programmable scientific calculator nothing more & nothing less. Self-driving cars have difficulty making decisions by code alone, sometimes a person using "smart maths" can quickly calculate "non-coded" variables that in the long run can save the day.

Equity is a static number
Also, you have raised the coding for Equity (Position size = Equity / 1000) using static numbers it's very easy "but" to code pyramiding of PositionSizing using a fluid "Bank Balance" is a little more challenging. If you are interested have a read how I handle "pyramiding of PositionSize": https://www.aussiestockforums.com/posts/1052081/

Suspended & delisted securities
There are some things that can be coded into a strategy where others are a little more difficult. The analysis report will not include "Suspended or delisted securities" as I've discussed. ASX announcements need to be checked as they hold information Amibroker is not privileged to. https://www.aussiestockforums.com/posts/1071517/

"Just because a buy signal is generated doesn't automatically mean it's a buy"
This is an important point that I was not aware of in this discussion. (I probably missed it) I was under the impression this was a purely code-based trading system in which you trust the system and take the signals without hesitation. But if there's a discretionary element to stock selection that can't be coded then I understand why the backtest won't match.

Discretionary element
The discretionary element is part of the trading plan that overrides the strategy selection & that's an announcement of a takeover or a scheme of arrangement that Amibroker is unaware of.

Skate.

 

[Beaches]

My understanding is that if there is more demand than offer during the auction, the price will increase and in the above example, the open price will NOT be 5.2 but 5.21 or whatever to ensure the situation you describe does NOT happen: open price is equal to price P where all sells down to P or below and buys up to P and above can be executed.I could be proven wrong but be disappointed

The algorithm is more complex than that and you will find on the heavily traded stocks there is nearly always a surplus on either the Buy or Sell side at the opening. There may even be times where part of your order is filled and part not at the opening. If you have a buy order in for 1,000 shares, depending on the bids .. you may only get filled for 300 in the opening auction the the other 700 will sit at the top of the queue for that price until there is a match during the day.

The example was only indicative of what could happen. However, even if the opening price was moved to $5.21, that would only result in 11,000 matching and there would still be 2,000 Surplus

If your broker's trading platform shows the indicated opening price and the surplus (some do not) you can watch a heavily traded stock from say 9:30am to 10:00am and you will see how the bids entered change the predicted opening price and the surplus.

 

[Lone Wolf]

A backtest can never match real trading for many reasons.

Agreed. I'm only advocating that you eliminate any differences in your backtest that can be eliminated. Then check your live results to make sure there are no differences to the backtest that you didn't expect. Different isn't a problem if you expect it and understand the impact on your expectancy. Just a suggestion.

 

[Skate]

I'm playing along as I'm interested to compare the different levels that we sell at and the overall impact it has. I won't necessarily be at the computer at 10:30am on the day we sell, so if I don't get out on the open, it won't be the same as Skate. What difference does this make over time? But my buy positions are identical to Skate as I used the same entry quantities and price limits. End values are different as CMC has $11 per trade ($9.90 if you do more than x trades in a month). So a bit worse than skate. However, I could have up to $10,000 positions and still only pay $11.
View attachment 103530

The CommSec confirmation summary
The 20 position Action Strategy portfolio is now full.





Skate.

 

[Skate]

How to read the Share Trade Tracker Dashboard
For those following along with the Action Strategy, I have made 6 references on the Dashboard capture that needs a closer look. The colour-coded reference boxes denote at a glance the information contained therein. I had planned to explain them in detail but as they say, a "picture paints a thousand words".

Profit Factor
Number (4) is the big one as it's the "Profit Factor" of the Action Strategy & we need to keep the cell coloured "GREEN". If the "Profit Factor" is less than 2 the cell will turn from "GREEN" to "YELLOW" which will be okay but I'm not trading the Action Strategy for "Okay" results.

It's worth remembering
"good is not good when better is expected".

The markets are being kind to us today
The results look pleasing so far today so I'll post a capture because as you know results can turn in a heartbeat.



Skate.

 

[aus_trader]

The markets are being kind to us today
The results look pleasing so far today so I'll post a capture because as you know results can turn in a heartbeat.

Yeah ! Go Goldies !

This is an awesome portfolio if the Gold price rallies like it is doing Skate.

From my analysis almost half of the candidates are Gold stocks. I can confirm at least 9.
i.e. 9/20 = 45% at least.

 

[Skate]

Skate.

 

[Rsthree]

My order for IGL was filled, 1143 @ $0.86

I've found that for me buying at auction over the weekend and being a weekly system helps me stay at an arms length from the process and therefore more emotionally detached, which ofcourse is a good thing.

In the past I was way more involved by monitoring throughout the day and executing throughout the week, so way too absorbed and too easy to get drawn into the drama of the market.

 

[Lone Wolf]

20 buy orders filled so far. All filled at the open price and matching Skate's (brokerage aside).

 

[rnr]

20 buy orders filled so far. All filled at the open price and matching Skate's (brokerage aside).

View attachment 103582

Well done @Lone Wolf as you have kept either below or equal to the Position Size of $1,000.

 

[Skate]

The Action Strategy up one day & down the next
Trend Trading has been proven to be profitable over time & trading the Action Strategy (a trend following system) should be no different but there will always be drawdowns along the way. It doesn’t matter when you start you will always go through a bunch of losing trades while waiting for a winning trend to come along. Many traders are not prepared for drawdowns & will stop trading because they are too impatient to get to the point of profitability & give up because the strategy didn’t perform straight away. A reminder "Trading is a marathon, not a sprint".

All the work has already been done
With the Action Strategy, all the work has already been done eliminating you from spending months & months researching methods & testing new ideas. Trading the Action Strategy allows you to trade a system with confidence because you can’t make money if you're not in the game.

Taking control is a personal responsibility
If you are already trading the Action Strategy but feel stressed along the way having trouble following the rules consistently then there is a good chance this system doesn’t suit you.

Emotions & trading don’t mix
As soon as our emotions go up, our ability to make good decisions drop. Trading mistakes are extremely expensive. Trading without rules, traders will revert to judgment & with judgment comes emotions & with emotions comes mistakes. Trading the Action Strategy is a way of trading without thinking allowing clarity & perspective to be experienced along the way.

Skate.

 

[Skate]

I'm back
After today's trading, I've been made whole again. I've got my profits back (said while punching the air) it's like the past 11 weeks hasn't happened.

Trading was looking great this Financial year
The "virus" has affected so many lives as well as the markets. I was having a great year accumulating a good deal of profits only to see some (70%) evaporate over a two week period. The crash came so swiftly, something I hadn't experienced in my 5 years of trading. But the good news is - "I'm back baby" as Phil Gould (NRL fame) would say.

Giving back profits is not always fun
The recent drop in my portfolio has been horrendous, to say the least, as the value in dollar terms was quite high. I'm lucky I was in large profits & gave back a percentage. Trading the last few weeks hasn't been all peaches & cream. I'm posting the line chart so others can understand what I've had to endure these last few crazy weeks. (but it was worth it)



Skate.

 

[laura]

Share the feeling
But how do you think compushare makes profit: by billing companies for mailing.
There is no real want by these chess manager to reduce
Why would i need to have a specific email, bank account, tfn per share traded, could not we have a default setup atyached to our hin?
Yes we can but where is the money....

So we actually need to log in somewhere each time and update details? Should I give them my TFN or ABN? Does it make any difference?

 

[qldfrog]

So we actually need to log in somewhere each time and update details? Should I give them my TFN or ABN? Does it make any difference?

Yes for every share ticker, with compushare you have to update tfn, credit directions and communication details.their site is atrocious and very slow.was worse 6 month ago.
Links services is better in that they default tfn , coms etc from your profile.
But for each update, compushare sends you a letter and get paid by the company of the share involved, a rort ...no other name
And extremely time consuming until you basically have done it for the whole XAO...

 

[laura]

Yes for every share ticker, with compushare you have to update tfn, credit directions and communication details.their site is atrocious and very slow.was worse 6 month ago.
Links services is better in that they default tfn , coms etc from your profile.
But for each update, compushare sends you a letter and get paid by the company of the share involved, a rort ...no other name
And extremely time consuming until you basically have done it for the whole XAO...

Ok, thank you! I'll update them all then! It asks either for TFN or ABN, which one should I use?

 

[qldfrog]

Ok, thank you! I'll update them all then! It asks either for TFN or ABN, which one should I use?

Not sure if Mr Skate is keen on seeing this type of info here? TFN should do but whether you want to trade as a business or under your name, or even an account with your partner can have tax ramifications.
We (on the forum) can not really offer advices here as individual circumstances vary.Sorry I can not offer much more help.

 

[Adamroweco]

just been looking at the au bank's and it seem's that every time the dow has a big up or a big down the au bank's reflect that , yesterday combank $58.88 jump to 60.o81 then a clime to 60.327 that was with a dow jum of 300p from the ending of a corona virus trial , today dow drop 300 combank 59.782 to 59.068 - and across the line all of the bank's on my list have replicated this .
the only other share's that i have on my watch list replicating this are AGL and Origin .
has any one been using this for profit ?