Disbelief - It's luck not skill - page 10

 
bigFish:
How can you say that Forex is luck?

I've been trading the same method for around 8 years. The method is reasonably diversified, and involves trading approximately 10 non correlated sub systems.

My returns over the past 8 year have been fairly consistant. However, when you look at each sub system, the variation in returns are all over the place. The system that returned 190% last year actually lost 15% in the previous year. The reason for those differences was nothing other than random chance. You may not be prepared to accept that, but its the reality of trading with any kind of an edge.

Years ago when I first started trading, it was very common to see dozens (and sometimes hundreds) of people on forums collaborating to develop trading methods. Such collaboration is a complete waste of time, but one thing that it did teach me was the distribution in returns from any method are completey random. Its not uncommon to see 100 people testing the same EA with the same parameter set, and to see a massive distribution in results. That didtribution is due to random chance, or luck.

It just takes one trader somewhere in the world to click a button and move the market, and that can impact my trade in a positive or negative fashion. I cant control that, I cant predict that. FFS the trader may actually hit the wrong buton and buy when he wanted to sell.

To be profitable, traders give up the ridiculous idea that they can predict the actions of hundreds of thousands of other market participants (on the basis of zero information), and they need to start focussing on whats happening NOW.

Of course support and resistance works, trading with the trend works, cutting losses short, allowing profits to run works, and TA works, but its a probability based outcome, and one where realising when you have luck on your side (or against you) can have a massive impact on your bottom line.

Thats why I say dont underestimate the role of luck.

 

With luck you may can make a few pips but in the long run "luck" alone will be not enough.

 

Hi Guys

This is my first post and I have yet to start trading in earnest. The trouble is that over the last year or so I have been backtesting lots of my own ideas which seem to work for a while but then don't do so well from then on. I understand about risk/reward and win/loss and looking at the bigger picture etc. I also realise the importance of placing stops at the right place and the impact they have on the ratios mentioned. However, I feel that I have exhausted my imagination for new strategies and need some help with new ideas that the classic books on TA don't give you. Kevinator offered a while back to start development of a system but has gone quiet since; Zupcon offered to share a system with Hamy but now refuses. Would either of you gentlemen consider taking me on as a student? I am willing to put the effort in (the sheer amount of learning I have realised in programming Excel and ProRealBacktest over the last year has surprised even me!). For the record, I believe that trading is not a matter of luck for the dedicated practitioner - but trading without a sound strategy is just gambling and therefore based on luck. I eagerly await a response from either or both of those members who might still be willing to pass on their much appreciated knowledge.

 

out of interest do you base your ideas on indicators or historical data mining?

 
Hooya:
out of interest do you base your ideas on indicators or historical data mining?

Hi Hooya, I am assuming that you are asking me that question. Actually, I use both. You would not believe the number of hours (hundreds!) I have spent transcribing OHLC data from my trading platform for a given instrument and time frame in order to backtest a particular strategy. I have also downloaded daily data from other websites but found that they often contain errors or do not reflect the price structure of the trading platforms I use (IG Index, CMC Markets). I generally go back at least one thousand candles for my backtesting (about 5 years for daily, 3 months for hourly).

My experiments with strategies have involved combining indicators - for example, one idea looked at the daily moving average (using various periods and types of MA, EMA, Wilder etc) to determine a trend, and then using RSI convergence/divergence to enter an appropriate trade. Stops were determined by trial and error for the best outcome (either points or %) - exiting with profits based on RSI value or Fib or some other indicator.

As an aside, I multiply my Reward/Risk ratio by my Win/Lose ratio to get a value which reflects the profitability of my system.

A recent system which showed some promise was based on measuring the gradient of a moving average to try and hit the peaks and troughs more accurately. Coupled with something like a MACD or RSI to determine if the instrument was overbought/sold gave some good results - for a time! But then there would be a really awful period which undid all the good work.

I have even tried trading against the trend which also works for a while but the optimum period for the trend is just too dynamic and never stays static long enough to exploit it profitably.

Sorry for such a long-winded response to a simple question but I suspect you are interested in my thought processes and so I wanted to pre-empt any questions in that respect.

All the best.

 

I know where you are coming from. I haven't traded for the last few months because I haven't had the time to put in the work.

Some people would have still traded but I like to have the numbers in my favour and to do that I feel you have to put in the work.

They key for me when I design any system is knowing how many trades it will take me to start breaking even.

I have traded before where I could only win 45% of my trades and not lose over a basket of trades. Sure eventually it went in to breakeven and I stopped trading it. I am sure had I stuck with it it would have made money again. BUt you can't aways wait for systems to come good...especially if you are trying to live off your profits.

You were asking before for help from a couple of posters. Why not offer something first? Post up your best system and maybe people can help with where you went wrong. Or maybe may have a suggestion.

 

Sure, why not! If what goes around comes around then I have nothing to lose. Here is a system which is very profitable on the GBPUSD Hourly chart:

Indicators: 49 period simple moving average (sma), 10 period ADX (adx)

Rules: Open a long trade on close of bar if the sma > sma[1] and sma[1] 30 where sma[1] is the sma of the previous bar etc. Open a short trade on close of bar if sma sma[2] and adx > 30. Exit a long trade when close > 1% higher than sma. Exit short trade when close < 1% lower than sma. Set a large stop of around 5% (about 800 points).

Over the last 3 months (as far back as my chart package will go on the hourly) there are 4 winning trades and one losing trade with an average profit of 250 points and the one loss of 131 points.

The system works on the turning point of the sma to get in as close as possible to the top or bottom of the chart and works quite well on the GBPUSD. The ADX filter improves the performance although why I cannot say since it is being used in an illogical way to its conventional use.

There is also a significant profit on the 30 minute chart but for other time frames and instruments it fails abysmally.

For info, the 30 minute chart gave 5 wins and one loss for an average of 127 points win and 52 point loss over the last 3 weeks (the limit of my backtest capability on the 30 minute chart).

Comments please.

 

I agree that it could be luck but some system might work as well, anyways something or someone watching the trades if the system consistency makes profitable then something happens to it if you know what I mean!!and u dont make profit no more so its hard to not lose!

 
hamy:
... if the system consistency makes profitable then something happens to it ...

I do know what you mean - but if the system is good then I think that is just short term variability and in the long run the bad trades should 'iron out' to give an overall positive result.

My problem is that to have confidence in a system it needs to work for me in any Forex pair and any timescale. Just having a profitable system for one pair as I have already described above and for only one or 2 timescales does not give me enough justification to trade for real with it. After all, it may only be profitable because I have found a suitable 'curve' to fit the data (with no guarantee that it will continue to do so).

Once again, I implore Kevinator or Zupcon or anyone who is willing to take on an eager student to pass on their knowledge of developing a winning system.

Kind regards to all

 

Feightdog..I will take a look at what you posted when I get a mo. Busy on something else at the moment.