Australian (ASX) Stock Market Forum

Trading discipline: Letting your winners run

Re: Trading discipline. Letting your winners run.

Actually, what I did was to add a non hindsight flag which indicated a GO/NO GO to the trade based on the chart at entry. i.e. I TESTED discretionary versus non discretionary as a filter. I was testing my belief that a discretionary component to the system would add an edge - it remarkably did exactly the opposite.

Would be dependant on what that discretionary filter was.

Discretionary MASSIVELY underperformed systematic. This was a consistent observation. Discretion kept me out of many of the outlier winner trades.

We have contrary results as Discretionary trading has placed me in outlier moves particularly short term 50% plus moves more often that Systems trading,infact I cant remember a 50% move in less than a week with any of my systems.---but that may have more to do with the system than Discretionary V non Discretionary.


Not at all.

If you take 10 trades @ 70% win rate with a R:R of 4:1 and I take 30 trades @ 40% win rate and R:R of 3:1 - who comes out ahead?

You'll feel better psychologically about your 70% win rate and how clever your analysis is, but I'll have more money.

True but now we are talking of trade frequency not win rate.
If I take my 10 trades in 2 weeks and you take 30 in 6 weeks I'll be wearing the smile.
 
I thought this was the beginners lounge?

Can some one tell me if i can set up a stop loss on comsec or is it a program i need to purchase?

thanks for starting this Neutral i did get some good stuff from some of the responses.
 
I thought this was the beginners lounge?

Can some one tell me if i can set up a stop loss on comsec or is it a program i need to purchase?

thanks for starting this Neutral i did get some good stuff from some of the responses.

Hi Alfa. Yes you can set up Stop Losses on CommSec. This is also who I use. You have to pass a short multiple choice test first. You can find the link on their website.

Neutral
 
Re: Trading discipline. Letting your winners run.

True but now we are talking of trade frequency not win rate.
If I take my 10 trades in 2 weeks and you take 30 in 6 weeks I'll be wearing the smile.
But that's exactly my point.

You are taking 10 trades in 2 weeks and getting a 70% win rate. You're getting the 70% win rate which makes you feel good because you're discarding many trade entries since they don't meet your setup/entry criteria. I then come along and trade much the same way, except I lower my standards for the entry. All of a sudden I'm taking 30 trades in the same time frame at a 40% win rate and a lower R:R and MAKING MORE MONEY.

i.e. Win rate is inversely proportional to trade frequency all else being equal.

I postulate that if you relax a given set of trade entry parameters you will;
- lower win rate
- increase trade frequency
- lower Risk:Reward ratio
- MAKE MORE MONEY

Think outside of the box. It ISN'T about being right.
 
Hi Alfa. Yes you can set up Stop Losses on CommSec. This is also who I use. You have to pass a short multiple choice test first. You can find the link on their website.

Thanks Neutral i will have a look at commsec tomorrow. I wish i new about these stop losses 6 months ago. I may have saved some of my capital ie MQG (mac bank) am now down 48pc. Thats just one in my RED portfolio started 14 months ago.

Lucky ive got money to burn NOT:banghead: good night

PS i did read some good stuff on one of the threads on stop losses
 
Re: Trading discipline. Letting your winners run.

Actually, what I did was to add a non hindsight flag which indicated a GO/NO GO to the trade based on the chart at entry. i.e. I TESTED discretionary versus non discretionary as a filter. I was testing my belief that a discretionary component to the system would add an edge - it remarkably did exactly the opposite.

I know that i'm a bit late to this discussion and the question is slightly off-topic, but I've seen a lot of people talking about how important testing is to a trading strategy. So far my (limited) day trading has been a combination of educated guesswork about support lines, channels, and momentum, but I'm interested in putting together a more concrete strategy and obviously testing is part of that. There doesn't seem to be a lot of helpful information out there about exactly how you go about doing testing.

I know I'm opening up a can of worms that deserves its own thread (and no doubt has many threads already) but I'll ask now, as well as have a good search through the old threads: do you do it in real time with paper trades, or do you generally back test historical data? And if you back test, how exactly do you get the data? And what software do you use to test it? And by you, I mean what do those in this discussion typically do, not asking what I could or should do - that said, if you do have advice, I'm all ears since I'm literally starting at the beginning here. Presumably I could do either, with backtesting you're running the risk of the 'rules' of the game changing over time, and with real time paper trade testing, you're limited in that you have to wait a certain amount of time before you can determine the results of the test! So as I see it, there is potential for each, but I'm interested in more of the practicalities of how YOU guys do it. Not asking for your trade secrets, just some advice on where to start and why.

I'm assuming if you're going to test a largely mechanical strategy, you need to code the testing with a pretty complicated set of algorithms and rules? How does that work? I suppose Testing 101 is what I'm looking for. ;) Both a brief summary of the theory but mainly how to do it in practice.

I'll take it to a separate thread if I've hijacked this one. :eek:
 
Re: Trading discipline. Letting your winners run.

But that's exactly my point.

You are taking 10 trades in 2 weeks and getting a 70% win rate. You're getting the 70% win rate which makes you feel good because you're discarding many trade entries since they don't meet your setup/entry criteria. I then come along and trade much the same way, except I lower my standards for the entry. All of a sudden I'm taking 30 trades in the same time frame at a 40% win rate and a lower R:R and MAKING MORE MONEY.

No arguement if your talking indentical timeframes and my higher R/R doesnt catch up to your trade frequency and lower R/R.
Id suggest however that a 70% win rate would suggest a lower R/R---in general terms and Low trade frequency would not be common---More commonly a higher trade frequency than those methods with lower win rates. We can both postulate combinations which would out perform each others.

i.e. Win rate is inversely proportional to trade frequency all else being equal.

I postulate that if you relax a given set of trade entry parameters you will;
- lower win rate
- increase trade frequency
- lower Risk:Reward ratio
- MAKE MORE MONEY

Cant agree with this statement.
-Lower win rate--yes--and no--maybe not
-Increase trade frequency --no---the exact opposite.I can have a method which has a high win rate picking up 1.5:1 R/R and trading 30 times a day.
-Lower R/R---yes---and no--depends on a number of things particularly exit,stop and trade management.
With a higher win rate I can have a lower R/R.
-Make more Money--not necessarily.

Think outside of the box. It ISN'T about being right.

Agree but its not about being random either.
For example I could set you a universe which you could never profit from.
I could set you entry parameter which would NEVER profit regardless of trade management.
Sound or even spectacular Money Management wont turn a flawed methodology into a profitable one.

On Systems
Amibroker is the Cheapest and best value for money. If you cant code your ideas there are people you can pay to do this.
I use Metastock and Tradesim combination.
As for Systems trading and Developement 101
Howard Bandy Quantitive Trading Systems is a good place to start although you'll need a mathamatical bent--- lots of time and patience.
 
Re: Trading discipline. Letting your winners run.

On Systems
Amibroker is the Cheapest and best value for money. If you cant code your ideas there are people you can pay to do this.
I use Metastock and Tradesim combination.
As for Systems trading and Developement 101
Howard Bandy Quantitive Trading Systems is a good place to start although you'll need a mathamatical bent--- lots of time and patience.

I'll check it out. I had brief look at his website and there is a PDF file of the contents pages - looks quite practical, even having a chapter for Amibroker, which is nice. Theory is all well and good but not much use if you can't use the information in a practical way.

That said, is testing the be all and end all? I've seen plenty of traders who claim they use nothing more than their own intuition (and obviously experience too!) to identify when to enter and exit trades... Would you still consider this a valid method of trading? I suppose any method that makes money is valid, but do you think can reliably be done or is it asking for trouble?

Reason I'm asking is that I'm still trying to 'find my style' and am open to investigating all options and strategies. Obviously intuitive trading sounds attractive - less dry reading, more edge-of-your-seat decision-making, etc, but also a bit more prone to emotional involvement at the expense of rationality.

I consider myself a fairly stable and rational person but I admit that I have made some bad decisions before, having held onto a stock for far longer than I should have. In this case it was because I hated the idea of breaking even on the sale but then going into the red on brokerage (am currently using Westpac Broking with around 0.03% commissions - more than twice as much as IB I believe). It was a big error or judgement and I watched the stock drop about 20% over the course of a couple of days - about half way through I figured it had gotten far too late to sell so I held on waiting for an inevitable recovery, which did come, but looking back, theres still no reason not to sell a freefalling stock, because has been mentioned in this thread, you can always pick it up at a lower price later on and save yourself significant losses. So I learned my lesson on that one.

I suppose I'm answering my own questions here. Different styles of trading for different people - all have the potential to work and the potential to fail miserably. :)
 
Best answered this way.

How do you know that your "Idea or Plan" is going to be profitable before you trade it?

If you cant answer that then you should find out what makes trading profitable.
Michael and I have been discussing that at length above.

You can trade without systems testing but you cant trade profitably without knowing what it is that will ensure profit!
 
Best answered this way.

How do you know that your "Idea or Plan" is going to be profitable before you trade it?

If you cant answer that then you should find out what makes trading profitable.
Michael and I have been discussing that at length above.

You can trade without systems testing but you cant trade profitably without knowing what it is that will ensure profit!

Well, right. Obviously there are fundimentals, trends, volatility that all make trading profitable. I see your point though, obviously you have to at least have some sort of idea or prediction about what is going to happen and when, and then execute your trades accordingly and pull out if you don't see what you expected to see happen (within reason).

I understand that, but I suppose we've all been saying is that you can't always predict it accurately, and when it doesn't happen as planned, you have to be displined enough to follow through and cut your losses short, right?
 
Well, right. Obviously there are fundimentals, trends, volatility that all make trading profitable. I see your point though, obviously you have to at least have some sort of idea or prediction about what is going to happen and when, and then execute your trades accordingly and pull out if you don't see what you expected to see happen (within reason).

I understand that, but I suppose we've all been saying is that you can't always predict it accurately, and when it doesn't happen as planned, you have to be displined enough to follow through and cut your losses short, right?


No I'm actually not saying that.
Profitability isnt about analysis.

Nothing to do with prediction.
All about skewing the numbers in your favor.
Michael and I are talking about ways in which this can be achieved.

Its a long road and not something that can be answered easily in a few lines.
Its about what you do when your right and what you do when your wrong.
Not how you are right or why you are wrong.
Seems like we are speaking in riddles when infact when you see it it is very clear.

Gotta go.
 
There doesn't seem to be a lot of helpful information out there about exactly how you go about doing testing.

do you do it in real time with paper trades, or do you generally back test historical data? And if you back test, how exactly do you get the data? And what software do you use to test it?
I'm assuming if you're going to test a largely mechanical strategy, you need to code the testing with a pretty complicated set of algorithms and rules? How does that work? I suppose Testing 101 is what I'm looking for. ;) Both a brief summary of the theory but mainly how to do it in practice.

For day trading the best software is EXCEL. I do all my testing with it. I don't use it to test trades but rather to find stats about how the market moves and reacts. With out some idea of what moves the market and the probabilities you are lost.

Well, right. Obviously there are fundamentals, trends, volatility that all make trading profitable. I see your point though, obviously you have to at least have some sort of idea or prediction about what is going to happen and when, and then execute your trades accordingly and pull out if you don't see what you expected to see happen (within reason).

I understand that, but I suppose we've all been saying is that you can't always predict it accurately, and when it doesn't happen as planned, you have to be disciplined enough to follow through and cut your losses short, right?

All you need is reoccurring patterns that you can trade with favourable R:R. in day trading there are many and all it requires is screen time and data collection (hint chart printing works well here) to recognise them.

And another hint most of the tradable patterns are not the text book ones.
 
Re: Trading discipline. Letting your winners run.

Agree but its not about being random either.

Well, now there's an interesting topic in itself. I'm sure you'll never agree with this, but IMO trend following AND analysis based trading (Elliot, VSA, macro pattern recognition, et al) IS basically random and relies more-or-less totally on trade management to succeed. The best trade managers are those that make the most money.

Following on from this belief comes my 40% construct: allow a good trade manager to manage more trades in a given time frame and they will make more money. To allow them to manage more trades, they need to relax their entry criteria if they are at 50%+ win rate. Relaxing their entry criteria will lower their win %.

I believe there is another layer to the onion - where the REAL edges are found, not the illusory ones provided by analysis. I believe this is where the mythical 1000% traders operate, the "5% of the 5%". There are very few of these - only two to my knowledge have ever posted to ASF (I'm not one of them...yet).

diliff said:
I suppose Testing 101 is what I'm looking for.
1. Think up a trading method that you think will work. WRITE IT DOWN IN FULL. Universe, Entry, Exit, Risk Management, Money Management. Everything about how the system will operate.
2. If it is capable of mechanical coding, use a trading simulator to figure out if it worked in the past. Note that complexity does not equate with profitability - usually the opposite. I use TradeSim+MetaStock, but AmiBroker can also be used and TradeStation is another in this field.
3. If you can't use a trading simulator for your system either; a/ simplify it until you can, or b/ hand backtest it on at least 30 trades - very time consuming to do this properly without introducing a bias.
4. If the backtesting shows a profit and the other parameters returned are satisfactory (profit is the LEAST important parameter - much more important are drawdown, win % and number of trades) then proceed to walk-forwards testing. This can be done mechanically if you backtested on a fully historical past (eg backtested 2001-2004, leaving 2005-2007 for walk-forwards testing) or you can do this by paper trading. I personally prefer paper trading as it reveals things about a system that you would never have considered otherwise.
5. If your forwards testing shows performance within specification, proceed to trial trading. Trade the plan with a small amount of capital.
6. If this still performs within specification, then you probably have a robust system which can go into full production trading.
7. You may well need several iterations of this process for any given system.

This is an overview and an oversimplification, but if you at least do this, you'll be ahead of the vast bulk of traders out there.
 
I'm sure you'll never agree with this, but IMO trend following AND analysis based trading (Elliot, VSA, macro pattern recognition, et al) IS basically random and relies more-or-less totally on trade management to succeed.

No not at all.
I once believed as you do.

The best trade managers are those that make the most money.

Plus I'll add---consistantly---and without blowing up.
 
I once believed as you do.
Not so at all.

Many traders have a psychological need for their analysis to make a difference. This inherently limits them to thinking within boxes.

There are very few outside of the boxes. I believe those that are are the ones capable of truly extraordinary returns, returns that are simply unimaginable to those constrained by their beliefs.
 
I would just like to tip this train of thought a bit.

Does anyone consider that the financial system you are dealing with `determines` the amount of success or failure?A trading method and the individual trader has to be accepted by the system for the trader to become financially successful.

First of all your friends have to be supportive of the ideas you have.They can affirm success with their words and help with observation and knowledge.

Secondly the brokers you are dealing with have to want you to be successful.They can push a stock in your favour or give you truths.Your success is the brokers success.

Thirdly the market you are in determines profit or loss from a trade!Not your method!The epitomy of vanity is when you `think` you are in control.We are all players in the game to serve someones purpose and to have our purposes met.
 
I would just like to tip this train of thought a bit.

Does anyone consider that the financial system you are dealing with `determines` the amount of success or failure?A trading method and the individual trader has to be accepted by the system for the trader to become financially successful.

I think you mean say the US financial state of affairs V ours for instance.
If so no.

First of all your friends have to be supportive of the ideas you have.They can affirm success with their words and help with observation and knowledge.

No

Secondly the brokers you are dealing with have to want you to be successful.They can push a stock in your favour or give you truths.Your success is the brokers success.

No--Brokers are not necessary when trading systems (Full service) pretty useless all round in my experiance.

Thirdly the market you are in determines profit or loss from a trade!Not your method!The epitomy of vanity is when you `think` you are in control.We are all players in the game to serve someones purpose and to have our purposes met.

To a degree refer to my comments on Universe,can also be applicable to Futures and indexes where the role of "Market" may have a bearing.
 
Does anyone consider that the financial system you are dealing with `determines` the amount of success or failure?A trading method and the individual trader has to be accepted by the system for the trader to become financially successful.
Accepted??
First of all your friends have to be supportive of the ideas you have.They can affirm success with their words and help with observation and knowledge.
Why??
Secondly the brokers you are dealing with have to want you to be successful.They can push a stock in your favour or give you truths.Your success is the brokers success.
When does a broker push a stock for or against you??
 
Does anyone consider that the financial system you are dealing with `determines` the amount of success or failure?A trading method and the individual trader has to be accepted by the system for the trader to become financially successful.
Nope. In fact, if anything the opposite is true. "Conventional wisdom" is usually wrong.

First of all your friends have to be supportive of the ideas you have.They can affirm success with their words and help with observation and knowledge.
You'll be making money when they're hurting. No support there, just attempts at undermining what you're doing.

You very quickly stop talking about trading with your friends.

Secondly the brokers you are dealing with have to want you to be successful.They can push a stock in your favour or give you truths.Your success is the brokers success.
Totally wrong. A broker succeeds by making you trade more so they earn more commissions. Plus, who will they look after - the big institution who has 1,000,000 of XYZ to sell or the small investor wanting to know where to put their money..."well, XYZ is a BARGAIN at the moment".

Thirdly the market you are in determines profit or loss from a trade!Not your method!The epitomy of vanity is when you `think` you are in control.We are all players in the game to serve someones purpose and to have our purposes met.

This I do agree with. The market will produce whatever profit it produces and you have no control over that. Your only control is over the amount you allow the market to take away from you.
 
Top