Blackstick
Member
- Messages
- 89
- Likes
- 0
SL 40 hit - sank like a stone!
Adamus
Experienced member
- Messages
- 1,898
- Likes
- 97
Let's discuss!
I have run an optimization process from 01.01.2010 to 17.12.2010, then I took the sets with better results and backtested them.
In the attached file you have the best sets for 2010 I have found.
Please consider the model quality is 90% and metaquotes data have some gaps.
It seems the set 55SL 43TP 20BE did not do better than 65SL 43TP 20BE, but there is a interesting set which I like it: 50SL 40TP 20BE, it has the small Gross Loss, the best Profit Factor, and a reduced DD.
Which set do you like and why? Everyone is invited
I think 1 year for an optimisation is too short. I would only be happy with at least 5 years. I used 10 years. That way I get the best average settings for many years and the chances that the results will hold up in the future are much greater. Any of those last 10 years could repeat themselves, or we could see a new variation of market behaviour with a new impact on FMT results.
wiseambitions
Established member
- Messages
- 764
- Likes
- 7
Ok, so if the euro started in 2002 and is now 9 years old, would you want to watch it another year til it gets to 10? Maybe it won't last that long in the form of currency as we now know it - depending who next of the PIGS (Portugal, Italy/Ireland, Greece and Spain) hit more trouble. I speak in jest
We have to be cautious with all this back testing and optimisation. Take today as an example, because some might still at 09.50 be in this trade having set 65SL on a start price of 1.6150, and conceivably could stay in the trade til 43 TP gets taken, but the next guy who went long at 1.6152 might have been knocked out and is automatically going to be 108 pips worse off at the end of the week. In an analysis of a year's performance I'd say that differences even of 250 pips are hardly of significance. The answer seems to be to find settings you are happy with, be disciplined, and stay with them. :smart:
Deatil
Member
- Messages
- 92
- Likes
- 0
Let's discuss!
I have run an optimization process from 01.01.2010 to 17.12.2010, then I took the sets with better results and backtested them.
In the attached file you have the best sets for 2010 I have found.
Please consider the model quality is 90% and metaquotes data have some gaps.
It seems the set 55SL 43TP 20BE did not do better than 65SL 43TP 20BE, but there is a interesting set which I like it: 50SL 40TP 20BE, it has the small Gross Loss, the best Profit Factor, and a reduced DD.
Which set do you like and why? Everyone is invited
Have you back tested a reversal trade of your stop at 65 when it gets hit? The reason for the question is that once the reversal takes place and a move of 65 pips has happen against the trade, you will in all likely hood see a new “mini trend”start.
The 65 pip stop will be 90% of the time outside the Asian range. Which is why is seems to work better than the 40 pip range which still keeps you just inside of the Asian range or barely just breaks it.
deserteagle
Active member
- Messages
- 205
- Likes
- 1
I agree, for example today my trade get in at 1.6148, therefore I'm still in, my SL was not hit just by 2 pips.
deserteagle
Active member
- Messages
- 205
- Likes
- 1
Some people say backtesting is an art, so there are many point of views on this, I do not backtest so long in the past, as maximum I do 3 years back to find a range of profitability, then I look in the previous year to get the best set inside that range.
Thanks for sharing your advice.
deserteagle
Active member
- Messages
- 205
- Likes
- 1
No, I did not backtest that, but it would be a good idea to check.
Adamus
Experienced member
- Messages
- 1,898
- Likes
- 97
Haha! I almost took you seriously. Jest aside, if it ever gets so far that the Euro breaks up, I think the mood of the world's governments will be so bad that they would also impose dire anti-speculation taxes or rules to prevent all but the big banks from trading forex.
Yes that's the main thing. Even if you get your trading rules from the entrails of fish or the tea leaves in a cup, if you are disciplined they might even work.
As DesertEagle said, if backtesting is an art form, then anyone can find anything in their backtest results, but I disagree because I believe trading discipline is needed in backtesting too - you need a definite plan and a definite goal. To make a mint from trading mechanically, you need robust, repeatable results and a focus on the kind of P&L that will accelerate your compounding, which means regular much more than profitable.
Actually optimisation is generally bad for the future. You have to try to avoid optimising a system unless it's crying out for it. The so-called 'art' of backtesting is recognising when you have optimised enough.
System parameters that can be optimised and will work well with the optimal settings in the future can be identified by a chart of the optimised parameter value against the profit. If your chart is a nice looking hill, low at the ends and rounding smoothly over a set of high values in the middle, then you can be reasonably sure it's stable to optimise that parameter. If the chart looks like a seismograph of the Los Angeles earthquake, then it's not optimisable. And of course there's everything in between and how do you know where to draw the line? You don't.
Deatil
Member
- Messages
- 92
- Likes
- 0
@deserteagle
Yes I would suggest you have a good look at this, it’s not rare to see these moves take place after the 65-69 stop loss get hit, just in the past 2 days alone take notice it has happen each time.
But just a tip, take a close look at 67 pip area on reversals, after a loss from FMT. You be able to cut your loss of 65 pip loss by around 40 pips on the first move before it will bounces to the former support area that held that marked the 67 pip area before it sells off again,
Use today and yesterday P/A as a guide for your back test.
Looks like news coming from China caused GBPUSD to drop like a rock. This is the second time in the past 30 days we've been burned by fundamentals coming out of China. Does anyone know where I can find the upcoming news/impacts for China? It looks like most of the major currency calendars don't have this information.
forexfactory.com says that yesterday was a bank holiday for China, so I don't see how there could have been news from China that affects the GBPUSD?
The news was this morning and it affected almost all major pairs because the value of the dollar went up. I don't see this news listed on their calendar page, but it is listed in their news section (after the fact).
deserteagle
Active member
- Messages
- 205
- Likes
- 1
Thanks for the info. Do you know when these news releases from China started coming out? Is this something new in 2011, or was this happening in 2010 when FMT was profiting consistently?
I'm not sure when the news releases from China started coming out, but I think it's new in 2011. Someone else mentioned it earlier in this thread but I don't feel like digging through to find it. I've been trying to find a calendar that lists China's news releases but I haven't been able to so far.
jshear
Junior member
- Messages
- 43
- Likes
- 3
The China interest rate increase was not and is never announced via a schedule. China raises rates when the feel like it. There was speculation China would raise rates in the coming days and usually they do it on holidays or sometimes on weekends. This week is Chinese new years so they did it on a Chinese holiday. You will never see in Forex Factory scheduled Interest Decisions for China.
Jeff
Jeff
Last edited:
Thanks Jeff for clearing things up.
Deatil
Member
- Messages
- 92
- Likes
- 0
I notice you wrote back on 26 Jan, on Post 5794 you stopped using FMT, http://www.trade2win.com/boards/trading-journals/101770-forexmorningtrade-journal-725.html
What made you change your mind in just a little over 10 days? or are you just on demo, trying to find the 'sweet spot" again through more back testing?
