Win percentage

[tech/a]

Quote: Originally Posted by nizar tech/a and others, just a question, what happens if for example u have $x, and each position size is $x/10, so u can have upto 10 positions open at any one time. If you are fully invested, is it true that u will not exit your position until your (trailing) stop is hit? coz this means that u could be holding onto slow movers and missing out on rockets potentially... if this is the case, what happens if while no stops are hit, your stocks that u have positions in are moving very slowly (albeit upwards) and ur system picks up a stock which is really firing, but your cash reserves have dried, so u cant take a position, do u decide to close one?? do u know what i mean, and what do u do in this situation? obviously the one u miss out on could be a real winner and what determines the (degree of) profitability of your portfolio yet u have missed out because u are "fully invested" also - wat im trying to get at, is that is there any other reason u would exit a stock apart from it hitting your stop? Thanks

[Bobby]

Quote: Originally Posted by ducati916 et al Well the numbers disagree with all of you; When linked to probability studies, the following was calculated; Example; An investor who can earn a return of 15% in excess of the Treasury rate, with 10% volatility. Calculated via standard deviations; Probability of making Money Scale...................................Probabilit y 1yr.........................................93% 1Quarter.................................77% 1month...................................67% 1day......................................54% 1hour.....................................51.3% 1min......................................50.7% In the short-time frames, it is a 50/50 proposition, but stretched out to the longer time frames, the probabilities rise very high. jog on d998

[Bobby]

Quote: Originally Posted by tech/a I have found an answer.

[It's Snake Pliskin]

Nioka,

Quote: If the company is researched before the investment is made the chance of failure is reduced considerably.

[happytrader]

Just a comment:

Quote: Originally Posted by stevo Hands up those who have sold something for a 5% or 10% gain only to watch the stock double or triple over the next 6 months - time frame is important.

[Devil_Star]

Hi, you guys have many good points on the topic. I believe everyone learns different things from their own experience. I particularly like tech's point ' consistency'. I have been searching for a system capable of minimising uncertainties and advancing consistently for some time.

Just wonder what you guys think how long of an 'overal winning' timeframe or how many numbers of 'overal winning' trades can prove a system as a robustly or consistently winning one.

After a year of trading ASX shares and observing the dow jones and nasdaq indices, I recently developed a trading system to day-trade the US CFD indices. I started testing it with $10,000 and completed 15 trades evenly in three weeks, 13 wins and 2 losses, made 40% return. Profit/loss for each trade is
$354, -$263, -$395, $758, $157, $532, $312, $559, $363, $222, $1,138, $159, $201, $56, and $278. (The biggest win is $1,138 and biggest loss is $395. ).

For every single trade I risked maximum 5% of my capital ($500 out of $10,000). I am open to advices from you guys of how long later it is considered to be suitable for me to increase the trade size to maximise my profits while still keeping the risk under control. Anyone knows a way to work out the reliability of a trading system via historical performance?

[tech/a]

D/S you still have a discretionary trading methodology.
The bigger the sample size for testing the better.
I like to see 1000s of trades.
If its trading indexes I would like to see how it performs on many bourses.
Why not code it up and test it.

[Private Investor]

Quote: Originally Posted by Snake Pliskin Nioka, Quote: If the company is researched before the investment is made the chance of failure is reduced considerably. This is a fallacy. It is quite clear that analysis is only a small part of trading and investing. Snake

[professor_frink]

Quote: Originally Posted by tech/a D/S you still have a discretionary trading methodology. The bigger the sample size for testing the better. I like to see 1000s of trades. If its trading indexes I would like to see how it performs on many bourses. Why not code it up and test it.

[tech/a]

No not a lot.

I do have intraday SPI tick data but havent had the time to work through any ideas.
Personally the Longerterm trading approach suits me best due to time restrictions.

I'm sure something could be worked out with the benifit of trading both ways.
Dont know that the SPI is ideal.

Radge has an interesting daily ASX thread going over on Reef.

[professor_frink]

Quote: Originally Posted by tech/a No not a lot. I do have intraday SPI tick data but havent had the time to work through any ideas. Personally the Longerterm trading approach suits me best due to time restrictions. I'm sure something could be worked out with the benifit of trading both ways. Dont know that the SPI is ideal. Radge has an interesting daily ASX thread going over on Reef.

[nizar]

Quote: Originally Posted by tech/a This is a good question one which I and many others have asked. I believe I have found an answer. Firstly the adage of you'll never make big profit taking little profits is fundamentally true.There will be times when a stock outperforms even the wildest expectations and THESE are the times when a TRAILING STOP should be incorporated into your trading---to capture that singual event which may never be seen again in the near future and give you a return on that trade which could have normally taken years. But I'm jumping ahead of myself. To answer the question in an encompassing way,you need to look at the whole way we trade. As most know I prefer to trade Systems where I have "Blueprints" from which to "Benchmark" results. To me the whole idea is to have CONSISTANCY.If I have this then I am a lot more comfortable trading considerably larger capital bases on Systems trades than I am in a Discretionary manner. I can make the Big $$s using Leverage and pyramiding (Or reinvesting ) profits from closed trades.As I have shown even mediocre system performance can bring spectacular profit. Specifically to the question. When portfolio trading we have 8-12 positions when fully capitalised.Most systems will continue to show potential buys throughout the weeks and months---even years we trade it. What you find or should find when testing your method is that the results will not vary greatly REGARDLESS which trade you take or when you take it.This is seen clearly by those with Montecarlo analysis capability. [In Simple terms this form of analysis allows you to allocate as many alternative portfolios over x time all with the same capital base and all traded over the period.All will be different--some by a stock or so and some by all stocks.At the end it reports the performance of all of them] What you find is a deviation from the average of X%.As an example the average maybe a 100% return with the worst being 60% and the best 150%. We all want the 150% or the higher end of the scale. The point is at the time of purchasing stocks we have no idea if that which we are purchasing will be a huge winner or be stopped out.We simply cant know. The fact remains though that at WORST testing has given us the confidence that in this case a 60% return will be had. What can we do to better increase our chances of trading in the top areas. (1) We can use Trailing Stops to capture the moves described about---but be careful that a trailing stop isnt placed on ALL stocks and it becomes an exit. (2) We can place a time and or Performance exit. (3) We can test various position sizing and numbers of stocks in our portfolio. Even so these inclusions will effect the very "Blueprint" our system has generated,so at best we are looking for overall better performance (And this can be measured in many ways other than profit) and a decrease in the deviation from the average. What does become clear though is that IT DOESNT MATTER which trade you take or how long you sit in it---as long as your within the Blueprint of the method designed you will have confidence in your positive expectancy.This releases you emotionally and financially to take advantage of those factors which will have a far greater impact on your profit than attempting to get the very best of EVERY move in EVERY trade. Can you associate with that Nizar? Takes a bit to get your head around.

[tech/a]

Nizar.

A trailing stop is normally introduced to capture the body of a move that would be seen as extraordinary.
Say a stock increased 50% in a few days or weeks and it would normally have taken months or years or you were on a move that had NEVER happened before.

This would be the time to introduce a trailing stop.
Normally a lot tighter than an exit,and if trading a portfolio you would only use it where a price shock occured.
What your trying to do is preserve to quick profits from a just as quick loss.

If trading in a discretionary manner and in smalls a great idea.Still keep in a normal exit but introduce a trailing stop.

One good one I like is last support,moving it up as support finds higher bases.

Here is an example with APG

[nizar]

Thanks tech,

do u place your trailing stop loss some percentage below the current previous close and re-adjust daily?

what determinants dictate where exactly the trailing stop loss is placed?\

great example as well..

[Devil_Star]

Quote: Originally Posted by tech/a D/S you still have a discretionary trading methodology. The bigger the sample size for testing the better. I like to see 1000s of trades. If its trading indexes I would like to see how it performs on many bourses. Why not code it up and test it.

[tech/a]

Nizar.
In the example above the support which is the low.
They would be reset when a new higher low becomes aparent.
Another I have used is Parabolic SAR,and another is a 4 day weighted average.

D/S
I guess your left then with discretionary trading.
All you can do is develope your "Blueprint" from the trades you have made and will make.If your trading begins to deviate strongly from the Norm you'll know somthing is up.
The main thing though is to record and have that info.

It wouldnt give me the confidence to trade $100,000s which of course isnt that hard with CFD's.

[Private Investor]

Devil Star,

I think 100 paper trades is plenty to give you at least a good idea on whether your trading strategies are likely to be profitable in the long run.

Obviously, the more you paper trade the more reliable your test results will be.

Based on what I have seen friends do, 200 - 300 paper trades is more than adequate to determine whether you should start committing your hard earned to actual trading.

For most people, more than 200 - 300 paper trades starts becoming tedious and time consuming for only slight less reliable results than computer modelling exactly the same parameters over 000's of trades.

[It's Snake Pliskin]

Quote: Originally Posted by Private Investor I doubt very much that the likes of Warren Buffet would agree that researching a company to reduce the chance of failure is a fallacy For punters who day trade then obviously research is not of high importance but for investors, especially for those with low risk tolerances, then research and fundamental analysis is an essential part to minimising the investment risk imo.

[WaySolid]

100 paper trades? I'm impressed. Certainly anybody who has this level of discipline has some positive traits going for them.

Personally paper trading has never worked for me, I like backtesting and then running with small size to start. Some of the systems I trade just could only be paper traded with a live simul account, not from the data alone anyhow.

Could just be my personality type potentially.

[It's Snake Pliskin]

Quote: Obviously, the more you paper trade the more reliable your test results will be.