My Trading System

[Zentrader]

Hi,
As I said in an earlier post, I am forward testing my mechanical system since early January with an account size 1/10th of that I intend. The system cuts out when there is anything greater than a 10% loss in account. unfortunately, that happened already, have been stopped out of first 7 trades, which is too much for my system....but given that the XJO has also lost 10% in the same period and it looks like things have gone a bit panicky and bearish....I may have entered a period when my system would show a draw down given it is a turtle style trend following system. Not that surprising when looking at the Big picture. This has certainly been informative for me. One good thing i am happy with is that with my CFD provider, all of my stops were honoured, and I kept tight money management. So it showed I can faithfully follow my system despite the mounting losses. Now I intend to continue paper trading the system until I get a trigger to enter with minor capital again. that is recover above 10% underwater equity curve. My system only risks 1.5% so happy with that kind of risk. I only wished I'd developed a shorting strategy too he he. But there you have it, we have to trade our system...and ride through its negative points.

 

[nizar]

My system only risks 1.5% so happy with that kind of risk.

Only?
I use 1.5% fixed percent risk and I thought it was aggressive!

 

[kingie_d]

Zentrader,
That turtles system you are testing, is that where you get into stocks when they make a new high of so many days? I think I read somewhere about it where they were using 100 days, which was all good at their time, but some other people more recently did it with 50 days or something like that, which in more modern times turned out to be better. If this is the system I am thinking of (basically a channel breakout I think) then can't you (and didn't they) also use new lows for shorting?
I think we are in the same boat when it comes to using a small amount of capital for testing for a certain amount of time before putting more capital into trading that system. I think that paper trading might be good for some people but they won't get the psychological experience that using real money gives. For me, I could paper trade for years and come out full of beans thinking that my system was great, but until I had learnt to contol my emotions while watching my account diminish thru a losing streak, then I wouldn't have learnt what I'm learning now.

 

[kingie_d]

I've just done some calculations of my expectancy so far. I've only completed 23 trades with this system so it is probably a bit early to do this but I did it out of curiosity. This includes a couple 1.5R and a nearly 2R losses.
Remember that my starting capital was only $5000 so transaction costs are significant (and this proves it)(23 trades = 23 x $40 = $920 OMG nearly 20% of my starting account!)

Expectancy: 1.057

Expectancy without transaction fees: 1.56

I think I have just proven the point you guys were making about the size of a trading account! Wow, I didn't think it was that significant!

 

[Wysiwyg]

Good stuff bro, I agree wholly.
In fact, it could have easily been me who said that what is quoted above.

Late in my discretionary days I just stopped trading. I just couldn't do it anymore because I realised I had no clue what i was doing (even though I was making money, Hey its a bullmarket). How and why should I stick to some set of rules without even knowing if the system I used was even a profitable one?

Okay, i was thinking about how a comparison would be as you develop a `system` to trade stocks.I plan to remain discretionary and adopt a keep it simple approach.I`ll continue with basic fundamentals/charts/patterns decision making.The outcome in 5 years would be an interesting comparison.Who knows, either may get lucky. Time will tell.

 

[kingie_d]

I just realised i did the wrong calculations for expectancy. What i did was:

(#wins * Avg Win)/(#losses * Avg Loss)

Just went and checked it and the correct calculation is:

Expectancy = (Probability of Win * Average Win) - (Probability of Loss * Average Loss) according to: http://tradermike.net/2004/05/trading_101_expectancy/

In Van Tharps books it is:

(Total wins - total losses - transaction costs) / # of trades

Both come to the same figures:

Expectancy: 5.96
Expectancy without transaction fees: 46.621

These numbers seem a bit crazy. They will probably look more "normal" after another 70 or 80 trades....

 

[AndrewM123]

Good Thread and some great inputs.
When I first started with a small a/c ($5,000) a few years ago, I also found the commissions a drain, along with poor decisions !! I switched to forex and only traded with currencies with low spreads e.g AUDUSD 2 pips.and no commissions. I also follow the 2% "rule" and now ALWAYS set a stop and will trail it to lock in a profit. Initially I Hated the thought of 'losing' and would stay in a trade far too long as if it were some form of personal attack. However after getting some good Money Management and systems in place things improved dramatically. You seem to be well on the way..

Good fortunes in your trading

 

[MS+Tradesim]

These numbers seem a bit crazy. They will probably look more "normal" after another 70 or 80 trades....

It's like flipping a coin. Winning or losing streaks early in the piece skew the picture. The more trades available, it will smooth out to long-term probability. Over time the expectancy will smooth out with the winning/losing trades %.

Example:

 

[hems13]

Thanks for sharing. Most of the traders and even educators and reluctant to give out any technique.
Thanks

 

[kingie_d]

Hello to everyone,
Just an update and a few new questions.
Reading back on my original posts in this thread I laugh at some of my early perceptions about trading. One of my big problems ever the last year of learning was inconsistency and impatience. I have since fine tuned my method but it still has a long way to go before I will use a larger account. Firstly, to obey the 2% position sizing rule, I was getting slaughtered in commisions by using such a small account - a point which I believe some of you guys stated when I first started. I am slightly down now in my trading account (compared to when I started with $5000) but I am fairly happy with the result (and knowledge/experience I've gained so far) considering that since I started I have paid over $4000 in commisions!
Someone mentioned Forex and I started reading all I could find about the Forex market. I got onto Oanda, on which I could use a demo account and paper trade new ideas. I can also trade in units which means I can position size for my small account without the fees (in the form of pip value of spread) being so high, relative to my account size.
I have now been trading using 1% postion sizing instead of 2% as I consider I am still learning and want to trade defensivley while I gain more experience.
I originally toyed with the idea of trading longer term in order to offset the (relatively) high commisions I was paying before, but as most of the gurus say, you have to find a style that suits you, not what suits others. And my preference/style is shorter term.
I have read about and studied nearly every indicator around thinking that I could find the "Holy Grail" but now I have overcome that part of my evolution as a trader. I have dropped all indicators and come to understand that they are just derivatives of price action anyway and are all lagging.
I now only have the long term/investor part of the Guppy MMAs (35 to 60) on my charts and only use them as a rough picture of what the market is doing. I am still looking for retracements in the longer term trends (the GMMA), and use them as entries for short term swings, but now I trade using price action "now" rather than indicator entries/exits.

One part of my trading that I would like to improve on possibly is the money management part of the trade as in scaling in/out of trades.
What are your general opinions on scaling in or out of trades. Is scaling into a winning trade better than scaling out and taking profits.

For example: (not real figures, just a generalisation)
If your trading system produced:
25% 1R losers,
25% losers less than 1R,
25% winners less than 1R,
25% winners greater than 1R,

and used 2% for total money (1R) risked.

Would it be better to either:
* Use the whole 2% as one trade, enter and exit with the whole lot,
* Use a smaller amount on initial entry (say 1%) then scale into a winning position with the other 1% if the trade goes your way, exiting with both trades at your systems exit point, (thereby reducing your overall R lost in a trade that goes against you),
* Use the whole 2% on entry and scale out (take profit) on 1% at a specified amount (1R, 2R, 3R or whatever) and let the remaining 1% run for the big winners.

This is just a quick generalisation, but is another part of money management which I am looking into as a way to improve Expectancy (Wins and winning amount compared to losses and loss amounts).

What are you opinions?...

 

[kingie_d]

The idea with scaling in or out of trades would be to improve a systems expectancy.
For instance, the scaling out (as in the above post) would mean capturing half (1% or the 0.5R component) of your trade at say 2R (or whatever your target is) before it pulls back, exiting the other half at "less than 1R". I know this is just a generalisation, but this would theoretically improve your expectancy right??? Because your winning amounts should go up. (Although, if prices keep going up you would miss out on half of your potential profits by exiting early which could have the opposite result)
With the scaling out, you would only lose 0.5R in a "bad" trade which, again should make your expectancy improve as the amount lost in your losers would reduce.
I hope you guys/gals understand the question I'm asking, I'm not talking specifics, just in general. But then maybe with this sort of question you need specifics?!? I might have to get the algebra out and think back to high school maths...

 

[traderc]

I think a lot of people place too much importance on position sizing realistically.

Fair enough it IS important and I'm sure I'll get lots of head shakes to say I'm giving the wrong message from this but most new traders don't invest enough capital to effectively position size or use systems such as scaling into and out of positions even though they SHOULD be aware at least of the concept and how it works.

With brokerage fees often working out to about the same as a 2% position size unless of course you're trading penny stocks, until you build up a few winning trades, position sizing is irrelevant and I think it's far more important to teach about HOW TO TRADE rather than HOW MUCH MONEY TO TRADE.

Realistically I think more impetus should be placed on learning about what indicators do, how to apply them and showing someone that is new to trading the characteristics of the markets and stocks, explaining your reasoning behind why you believe which stocks are the most likely candidates to trade successfully (i.e. what characteristics they display to potentially earn the trader money) and why particular indicators are good to use.

Baffling a newcomer with ideas that can't be used immediately is a waste of time.

I know most of you will probably have heard of Aussie Rob and despite the fact that he wants to simplify everything to one indicator, a lot of information can be gleaned from this mentality. K.I.S.S. principal I think is the key and by looking at some of the formulas people create, the simpler the better, especially for new traders.

I think trading at the moment is very similar to how computer systems used to be (I'm a computer system administrator by trade) in that "the geeks" used to hide behind the veil of being "all-knowing" and charge ridiculous prices for their services, even if they didn't know what the hell they were really talking about.

This is the way I see a lot of traders as well; they try to impart knowledge of systems they really know nothing about which can really cause the "newbie trader" to get into strife pretty quickly.

I think the most important factors in trading are:

• Not listening to hot tips from anyone
• If you get a tip, analyse it via your trading system
• Trade shares that are highly liquid
• Trade cheap shares (if you have a small amount of capital to invest)
• Ensure the share or instrument you're trading is trading consistently (we don't want too many flat-spots to affect our indicators)
• Learn how/where to place a stop loss
• Find shares that are volatile and follow a nice up/down consistent pattern
• Diversify your portfolio - e.g. $10,000 = 10 x $1000 or 20 x $500 parcels

Trading is a gamble; yes it's an educated one and you can "guess" what is going to happen from previous historical data by saying, I've seen this happen 110 times before so I can say next time "I think it will happen" but the reality is, it might not.

Whilst many people these days in the stock trading markets seem to want to hide behind their knowledge or baffle newcomers to the stock markets with BS or to rephrase that, too much information that a new stock trader is not yet ready to understand, it's really just overwhelming for the individual and doesn't help any.

I think the list I've created will help any new trader in sifting the wheat from the chaffe and help them understand the underlying fundamentals to technical trading or at least provide an idea of the DOs and DO NOTs to at least create a foundation of a trading mentality.

Maybe when someone comes up with an idea and says "I'm a beginning stock trader" it should be said that instead of treating them like a child, a simple explanation of methodology or an idea of the concepts is far more preferable than chewing down to the bone on one specific area that really, for a beginning trader, until they get some money in their kitty, is irrelevant.

I'm sure we've all been there and felt like an idiot at some point when learning something new and especially with stock trading, there is so much information to take in on different aspects of the markets that we know can be a steep, difficult and costly learning curve.


Regards,


Christian

 

[tech/a]

Indicators are purely a derivative of price.

A combination of either
Open
Close
High
Low
Range
Volume
Open interest.

With all of the above available why on earth would you trade indicator based?

Trading is a gamble; yes it's an educated one and you can "guess" what is going to happen from previous historical data by saying, I've seen this happen 110 times before so I can say next time "I think it will happen" but the reality is, it might not.

Its only a gamble if you dont have an answer to how do you manage the trade if it goes your way and how do you manage the trade when it doesnt.

Realistically I think more impetus should be placed on learning about what indicators do, how to apply them and showing someone that is new to trading the characteristics of the markets and stocks, explaining your reasoning behind why you believe which stocks are the most likely candidates to trade successfully (i.e. what characteristics they display to potentially earn the trader money) and why particular indicators are good to use.

If you really believe that this is where the edge is found then you will choose this path. Once you find its not here then you'll choose the path of expectancy.

Have a long look at those traders you KNOW are profitable and tell me what they have in common. I'll bet its NOT a secret indicator or combination of indicators.

I think the list I've created will help any new trader in sifting the wheat from the chaffe and help them understand the underlying fundamentals to technical trading or at least provide an idea of the DOs and DO NOTs to at least create a foundation of a trading mentality.

All good.
But unless you know that the method of trading your applying to the above is profitable you'll risk ruin just as easily as someone who has never seen your list. You wont know when its failing,your likely to cease trading it when it is still well within its parameters (if proven profitable). Many a profitable trading method has been discarded whilst in drawdown---without knowledge of what an acceptable drawdown for it actually is.

Here are the test results for a system which doesnt use an indicator other than an M/A the rest is price only. These are the Montecarlo results for 20,000 portfolio simulations. The edge is clear but its NOT in the Buy or sell INDICATOR/S

The EDGE is in the very LAST table.

 

[johenmo]

Indicators are purely a derivative of price.

A combination of either
Open
Close
High
Low
Range
Volume
Open interest.

Never thought of it this way before but they are, aren't they?

With all of the above available why on earth would you trade indicator based?

Tech - is this because it's "conventional thinking"? You responded to my first attempt to build a system 6 months ago and told me it was too complicated. It really was. What I'm playing with at the moment uses indicators and is OK - I think my filters (stock selection) needs more work before I can work on entry/exit stuff.

But I'm failing to grasp what one would use if one didn't indicators. I'm guessing the reply is something like "price action". How do you define/explain price action?

 

[Chorlton]

But I'm failing to grasp what one would use if one didn't indicators. I'm guessing the reply is something like "price action". How do you define/explain price action?

I would suggest "Price Action" is simply how price moves in relation to past price. In addition the use of Volume is used to confirm the likelihood of a particular price move. The analysis of combining Price & Volume is known to many as VSA. I believe this is what Tech/A is referring to.

Although all indicators are derived from price (O,H,L,C) and/or Volume, I do believe there is some merit in using them IF (and only IF) you understand how the indicator is derived and what the indicator is actually showing you on the chart.

If you want to achieve a particular output and an indicator can do that easily then I see no harm in using it. For example, although a trend can be defined by price action (higher lows, higher highs, etc, etc) it is sometimes a lot easier to use a Moving Average. Nothing wrong with that.

However, in contrast if you have simply added a few random indicators to your charts (RSI, Stochs, etc, etc) and then decided on a way to find reliable trading signals from using them, then this is probably the wrong approach....


Just my 2cents,

Chorlton

 

[wabbit]

I just realised i did the wrong calculations for expectancy. What i did was:

(#wins * Avg Win)/(#losses * Avg Loss)

Just went and checked it and the correct calculation is:

Expectancy = (Probability of Win * Average Win) - (Probability of Loss * Average Loss) according to: http://tradermike.net/2004/05/trading_101_expectancy/

In Van Tharps books it is:

(Total wins - total losses - transaction costs) / # of trades

Both come to the same figures:

Expectancy: 5.96
Expectancy without transaction fees: 46.621

These numbers seem a bit crazy. They will probably look more "normal" after another 70 or 80 trades....

I have been writing some applications which model expectancy and have been researching some of the different ways people have computed expectancy in the past; none of them is perfect IMHO but it does highlight the need the for the expectancy result to have a common unit of measurement, i.e. expected return for each unit of currency invested.

e.g.

Lets pretend you invest $25,000 in each trade
P(Win) = 0.417
P(Loss) = 0.583
Avg(Win) = $745
Avg(Loss) = $235
(I got these numbers from one of Nick's seminar notes)

If you just use the formula as it is written in many books with little or no further explanation:
E = P(Win)*Avg(Win) - P(Loss)*Avg(Loss)

E = (0.417*745) - (0.583*235)
E = 310.66 - 137.00
E = 173.66

What does this value represent? You will often see numbers like this bandied around by system-sellers when they "demonstrate" their system expectancy.

If we use the values at Trader Mike's site:
As an example let’s say that a trader has a system that produces winning trades 30% of the time. That trader’s average winning trade nets 10% while losing trades lose 3%. So if he were trading $10,000 positions his expectancy would be:
P(Win) = 0.3
P(Loss) = 0.7
Avg(Win) = $1000
Avg(Loss) = $300

E = (0.3 * $1,000) - (0.7 * $300)
E = 90

So is this system about only half as good as Nick's system?

These "extreme" values of E are showing not the expectancy for each $1 invested, but for each $25,000 and $10,000 invested! To return a more meaningful, important and usable value for E, divide the result by the investment:

For Nick's system
E = 173.66 / 25000
E = 0.00695

i.e. The expected return by this system on $1 invested is 6.95cents.

For Mike's system:
E = 90 / 10,000
E = 0.009

i.e. The expected return by this system on $1 invested is 9.00cents.

The alternative method to arrive at the same value for E is to divide the Avg(win) and Avg(Loss) values by the investment value before entering them into the equation. I believe a better notation for the expectancy formula is:

E($) = P(Win)*Avg(Win per $ invested) - P(Loss)*Avg(Loss per $ invested)

Mike would have been better to leave his formula using the percentage returns.

E = (0.3 * 0.10) - (0.7 * 0.03)
E = 0.03 - 0.021
E = 0.009

If I had a system with a natural E of $5.96, I'd be ecstatic, regardless of how much commission I was paying!



Hope this helps.


wabbit


P.S. Another discussion could be had about changing the expectancy formula to:
E($) = P(Win)*Avg(Win per $ RISKED) - P(Loss)*Avg(Loss per $ RISKED)
where the amount RISKED is brought into play, but this requires a loooong lunch with many beers and wines!

 

[tech/a]

Wabbit
Some interesting and valid points.
I'll have to look into how Tradesim calculates Expectancy and ask David his thoughts. Perhaps its a matter of having various expectancy formulas at our disposal.

[/quote]Tech - is this because it's "conventional thinking"? You responded to my first attempt to build a system 6 months ago and told me it was too complicated. It really was. What I'm playing with at the moment uses indicators and is OK - I think my filters (stock selection) needs more work before I can work on entry/exit stuff.[/quote]

Firstly you can have a profitable indicator based system,Ive seen them.
But I dont use fully indicator based trading methods.
If you test your system against Random entry often you'll find that the results for random entry give a higher return than your own entry criteria.

So clearly the success of the system DOESNT lie in the entry conditions.
If you disect most of the conventional indicators you'll find that they are simply a graphical representation of a combination of the components mentioned above over a period of time. The problem is is that they wont conform to the same behavior often enough at POINTS of time after their period of calculated time.

As an example.
A close greater than a 14 day ema.
There are closes above 14 day ema's all he time but not enough of them continue above that 14 day ema to be reliable as a sure fire indicator of what price will do.

The other important question in system design is HOW LONG do you want an instrument to be in a positive position to make a profit.
A day
A week
A month
A year.

How important are your entry conditions going to be in a year. Id say of no value. So all your really doing is hopping on MOMENTUM ---positive or negative. In todays market expecting momentum to last for a year in a positive direction is un realistic.

But I'm failing to grasp what one would use if one didn't indicators. I'm guessing the reply is something like "price action". How do you define/explain price action?

VSA type setups can play a part. But there are countless others.
Pivot point reversals,Highest value for X bars,Bars since X occured,Gaps.
Fillig of gaps,Retracement % coupled with price and volume action,All the Candlestick setups,Hooks,Inside days. The list really does go on.

I found myself in a situation just reciently.
Since the boom finished I have been trading in a discretionary manner.
Results have been good and not so good. Hardly worth the time devoted to trading. After a chat to Nick and without Nick saying as much,it is obvious that a return to basics.(system developement and a proven trading methodology is what I personally need for these times) My time and inclination isnt suited to discretionary trading. Good money can be made trading in a discretionary manner but my view is that you need screen time---something I dont have and something I dont wish to make time for.

 

[Trembling Hand]

I found myself in a situation just reciently.
Since the boom finished I have been trading in a discretionary manner.
Results have been good and not so good. Hardly worth the time devoted to trading. After a chat to Nick and without Nick saying as much,it is obvious that a return to basics.(system developement and a proven trading methodology is what I personally need for these times) My time and inclination isnt suited to discretionary trading. Good money can be made trading in a discretionary manner but my view is that you need screen time---something I dont have and something I dont wish to make time for.

So can I as what you are trading? or how?

 

[tech/a]

Trading in frequently.
SPI
and the Odd stock.
Problem is for instance.
I'll have a position on, expecting only a day trade.
I'll be called out to site having to leave the screen for 2-3 hrs.
I close the trade.
Or I'll be late in due to appointments and I will have missed a trade or 2.
Or be in a meeting and both situations above occur.

If my situation were different then I would have the screen time required to trade this way it is profitable but not "Profitable enough".
As my trading isn't and is unlikely to ever make the returns my Company is my focus is primarily business in time allocation.

 

[motorway]

Never thought of it this way before but they are, aren't they?



Tech - is this because it's "conventional thinking"? You responded to my first attempt to build a system 6 months ago and told me it was too complicated. It really was. What I'm playing with at the moment uses indicators and is OK - I think my filters (stock selection) needs more work before I can work on entry/exit stuff.

But I'm failing to grasp what one would use if one didn't indicators. I'm guessing the reply is something like "price action". How do you define/explain price action?

Never thought of it this way before but they are, aren't they?



Tech - is this because it's "conventional thinking"? You responded to my first attempt to build a system 6 months ago and told me it was too complicated. It really was. What I'm playing with at the moment uses indicators and is OK - I think my filters (stock selection) needs more work before I can work on entry/exit stuff.

But I'm failing to grasp what one would use if one didn't indicators. I'm guessing the reply is something like "price action". How do you define/explain price action?

A pure expression of realprice action
is a Point and Figure chart

A chart that by it's nature is a trading system
As mechanical or discretionary as you want it to be

You can not but be systematic with a Point and Figure chart

motorway

eg