Are these results good enough to trade?

[Superboot]

Hi all,

Been back testing a trend system and have 3 questions:
1. Would you trade the system with the following back-test results?
2. How far back would you look with respect to historical data? (As you can see my result are not so flash pre-2000)
3. What other factors would you consider before making a decision to trade it ?

Disclaimer - This is my first system, I know it is not perfect, but keen to get started

My in-sample dataset has been from 1/1/2000 to 31/12/2003 - this includes 1 bear year and 3 bull.

Results are below:
2000-2003 (average results)
CAR = 33.9%
MaxDD = 22.8%
% Win = 41.6%
% Loser = 58.4%
Other criteria ???

2004-2010 (now) (average results)
CAR = 29.4%
MaxDD = 31.9%
% Win = 38.3%
% Loser = 61.7%
Other criteria ???

1992-2000 (average results)
CAR = 5.5%
MaxDD = 29.9%
% Win = 35.8%
% Loser = 64.2%
Other criteria ???

Results do not include any commissions.

 

[So_Cynical]

Just interested Superboot in why you are testing a trend system, why have you chosen trend following as your method of trading choice?

 

[Superboot]

Hi So Cynical - good question.

I am ultimately looking to develop other system(s) for non-trending markets however, a trend system is the one I decided on first - possibly because many of the books I have read focused on these.

Also, I am only able to act on eod data and have limited time to manage my system - the trend system I have developed meets these requirements. Some of the other systems I have studied (e.g short term range trading) require more trade activity.

Do you trade a trend/range or other system?

Cheers

 

[Wysiwyg]

Hi all,

3. What other factors would you consider before making a decision to trade it ?

I have near 6 months Ami. experience. Back tests do not replicate real time. To know for sure there is only one way and that is to paper trade it for awhile.

What are the maximum consecutive wins and losses? Could you endure that many or more in reality?

Happy to talk.

 

[So_Cynical]

Hi So Cynical

Do you trade a trend/range or other system?

Cheers

I doubt i qualify as a trader, the average time im in a "trade" is about 6 weeks and have only 2 to 4 "trades" going at a time....i don't have a system , more a set of somewhat discretionary rules, and tend to find myself buying bad news pullbacks and ranging stocks.

Are there range trading mechanical systems? i think sideways is a very profitable and yet somewhat over looked trend.

 

[Space Invader101]

Superboot I trend trade for a living. You need to consider the overall market indices and test your system when the overall market is in a long-term trend. That being the ASX200 and Dow Jones (DJIA) charts.

Test the following bull markets (up trending):
- 14th March 2003 (start of Iraq war II) to Jul 2007 (Beginning of Global financial Crisis)
- March 2009 End of the global financial crisis to Now

Events such as the 2002 downtrend and the GFC in 2008 will skew your results. Also back testing with software won't tell you everything compared to looking at the charts. Some stocks move differently to others. I tend to pick stocks that better suit my trading system.

I would disregard data prior to 2000. Think about how many traders and investors have the internet and software compared to pre-2000. A chart isn't just price, but mass psychology and the masses perception of stocks has changed by having more access to information.

 

[Wysiwyg]

Test the following bull markets (up trending):
- 14th March 2003 (start of Iraq war II) to Jul 2007 (Beginning of Global financial Crisis)
- March 2009 End of the global financial crisis to Now

Events such as the 2002 downtrend and the GFC in 2008 will skew your results. Also back testing with software won't tell you everything compared to looking at the charts. Some stocks move differently to others. I tend to pick stocks that better suit my trading system.

It just so happens that is exactly what I am trying to work around. The bear markets. If superboot throws up an equity chart it will reveal the huge draw down periods.
These four indexes: 1) The ALL ORDS 2) The S&P 500 3) The S&P TSX Composite(Canadian) 4) The FTSE All Shares ... show the flat spots where no trading is done. Hard to make progress going long in a down trend. I am working on a short trading strategy to trade the bear markets in unison with this system.

The system is the same for each and it pretty much mirrors the indices.

 

[Wysiwyg]

I am working on a short trading strategy to trade the bear markets in unison with this system.

First equity chart of Long trades only then next equity chart with "assistance" from the Short trading strategy while the Longs are not trading.
Then table of trades for month and year.

$100k starting capital, 10 % equity in each trade if possible
$30 each way comm.
No stop loss, take profit or trailing profit
Entry and Exit via conditions

821 trades (787 longs, 34 shorts, short contribution = $31k )
Draw down 10.7%
Largest win = $7510
Largest loss = $2921
Max. consecutive losers = 12 at an average of $302
Max. consecutive winners = 7 at an average of $632

Probably do better with a dartboard.

 

[Superboot]

Many thanks for your comments

I am using Amibroker, is there an easy way to view the 'ups' and 'downs' of your equity curve? I have seen monthly bar charts of this, however not sure if this was in AB. I know there is a ~~~~~Equity chart...

I did start trying to add in a short component when the market was in a downtrend but it started getting messy (will nail the long side first).

I do have an XAO filter i.e. only enter positions when the EMA(XAO,5) > EMA(XAO,20). However, I don't know whether this would be enough to dodge the big bears - should I be using a longer term/use it as an exit condition etc?

 

[tech/a]

Take the filter out then see if there is an appreciable difference.
If not --back to the think tank.

How about this one.
180 day ema of the low triangulated

Google triangulated M/A 's if you want to know how they are calculated.

Click to expand

Keeps you out of the big falls and in the big rises---thats trend trading.

 

[yonnie]

google doesn`t know triangulated m/a`s......did you make them up yourself?

 

[captain black]

google doesn`t know triangulated m/a`s......did you make them up yourself?

Here's a link to a TMA written for Amibroker:

http://www.amibroker.com/library/detail.php?id=240

Also a general description of a TMA:

http://daytrading.about.com/od/indicators/a/Triangular.htm

 

[CanOz]

google doesn`t know triangulated m/a`s......did you make them up yourself?

LOL, perhaps you made a spelling error Yonnie?

CanOz

 

[AlterEgo]

Results do not include any commissions.

And what difference does it make when you add commissions in? And if you're entering and exiting with stop orders, don't forget to add slippage in too.

 

[yonnie]

LOL, perhaps you made a spelling error Yonnie? CanOz

hah, if I made a spelling mistake, Tech surely made one too.
am just copying the guy.....

 

[Superboot]

Interesting - I hadn't heard about TMA's before.

Thanks for the links Capt'n, the logic makes sense i.e.
SMA = (P1 + P2 + P3 + P4 + ... + Pn) / n
TMA = (SMA1 + SMA2 + SMA3 + SMA4 + ... SMAn) / n

However, I got very lost in the amibroker code (I might try it myself)

I have tested my current XAO filter (EMA(XAO,5) > EMA(XAO,20) ) and it does have a significant impact, however maybe not enough to keep the maxDD level down to a more appropriate level. I think from memory, when I did test a longer term XAO MA it hurt my CAR.

I think the reason for this is that I am not in the market for those periods. If I re-calculate my CAR over the active up trend periods it should be a better indication of my return (as per Space Invader101 suggestion). I would also expect it to provide a lower maxDD. I can then treat the down trend periods as an additional compenent to my system.

Will see how it goes...

 

[Space Invader101]

What chart intervals are you using, Superboot, for the All Ords and Individual stocks? (Daily, weekly, monthly etc.)

I just focus on the ASX200. Maybe test the top 200 and make a comparison with the All Ords.

 

[tech/a]

Here is another---this more reactive.
Its not An M/A nor is it an ATR.
Its not a canned indicator either.
Its one of my own.Superboot I'll private mail it to you if you want it.
It will be in metastock language but only a line of code.

BLUE is the Triangulated M/A.
Click to enlarge.

 

[Superboot]

And what difference does it make when you add commissions in? And if you're entering and exiting with stop orders, don't forget to add slippage in too.

Hi AlterEgo,

On average I found about a 2% reduction in return once commission had been factored in ($17 each way of the trade) - I understand if I moved to IB it would be even less.

Not sure how I would measure slippage? I use an online broker - I guess you would need to trade/paper trade and compare to your backtest results (?)

 

[Wysiwyg]

An option to trade in relation to index movement is:

Q = Foreign("^AORD", "Close"); // Change to index of choice
Q1 = DEMA(Q, 250); // Change to desired moving average & period
Q2 = Q > Q1; // Open long positions
Q3 = Q < Q1; // Open short positions
Q4 = Cross(Q1, Q); // Close ALL long positions
Q5 = Cross(Q, Q1); // Close ALL short positions

Vary Q and/or Q1. Usable in conjunction with other buy/sell criteria.