Strength for one week, No Charts Attached ;) | Page 10 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Charts, Charts, Charts! 1 reply

Hallett Norman - Mental Fitness For Traders (Attached) 2 replies

EA works attached to single chart, messes up with multiple charts 2 replies

The Pocket Book Of Economic Indicators (Attached) 2 replies

One More Zero - How To Trade FX Like A Pro (Attached) 0 replies

Attachments: Strength for one week, No Charts Attached ;)
Exit Attachments

Strength for one week, No Charts Attached ;)

  • #181
  • Edited 11:57am Aug 11, 2011 10:52am | Edited 11:57am
Quoting zznbrm
Here is a summary of 5-month automated backtest. Results are in US dollars, trading on a FxPro standard account using 0.1 lots (10,000 units) per trade. FxPro is on Eastern European time, so I adjusted the times accordingly. Account started with $10,000 and, as you can see, ended with $13,655.63. I used very conservative spreads, but did NOT include swap.

During the week of 5/9/2011, drawdown got to about 33%.

If anyone is interested in the full results, I will post the entire result set. Like I said previously, my only deviation from Post 1...
These results seem logical. When taking random trades, the result should stay around BE longer term.
As previous strength says nothing about next direction, we may call these random trades, in fact.
I still like the idea of trading the strength, and possibly it would work with hedging correlated pairs.

My today's basket is also in big loss. Yesterday's profit gone, account almost at initial state now. Will keep it until at least 18 GMT, I guess.
  • #182
  • Aug 11, 2011 11:07am Aug 11, 2011 11:07am
  •  adeforex
  • Joined Aug 2008 | Status: Trading Live as a Professional Do | 363 Posts | Online Now
Quoting Indrek
If there is big intervention or sth (like today) then your account will be gone on weekly trades too if your lot sizes are too big.
Of course, when taking weekly trades you can make five times less mistakes And you do not have to buy CHF all the time at the time when it is falling.
Let us see what guys will have for this week's result. CHF is very very bad now...
I think MM will be very important with this method.What about using DAILY/weekly/monthly ATR/ ADR for tp/sl to avoid large loss. I know FB want to leave his trading naked to maximise profit but multiple losses can be fatal. Now, the system can be profitable in the long run. I prefare closing my account daily just like you are demoing it. so that one will know his/her risk per day/week/ month. The task is setting the tp/sl in the long run. Thats my suggestion anyway
We Win Some, We Loose Some, But Let your Win Rates be More to be Successful
  • #183
  • Aug 11, 2011 11:09am Aug 11, 2011 11:09am
Quoting zznbrm
Here is a summary of 5-month automated backtest. Results are in US dollars, trading on a FxPro standard account using 0.1 lots (10,000 units) per trade. FxPro is on Eastern European time, so I adjusted the times accordingly. Account started with $10,000 and, as you can see, ended with $13,655.63. I used very conservative spreads, but did NOT include swap.

During the week of 5/9/2011, drawdown got to about 33%.

If anyone is interested in the full results, I will post the entire result set. Like I said previously, my only deviation from Post...
Thanks for doing this. I think thats all the info we need
  • #184
  • Aug 11, 2011 11:26am Aug 11, 2011 11:26am
Quoting pilotui
Thanks for doing this. I think thats all the info we need
Yes, thank you for your efford! But I call 5 months of trading using weekly setups hardly enough evidence... If you look at the weekly chart you'll see that 5 months or about 20 weeks is not much on this timeframe. Market conditions change over time and 5 months are hardly representive of possible market conditions...

To know if a certain system is long term profitable you need atleast 10 years of backtesting. I know this seems very long... but as this system uses 10 week strength calculation (which is about 2,5 months) 10 years might be a reasonable amount of time to test.

This will probably be very time consuming but may be worthwile the effort!
  • #185
  • Aug 11, 2011 11:42am Aug 11, 2011 11:42am
great stuff, ZZ. Thanks for posting these.

Could you please post all of ur findings? I'd definitely provide a shocking top secret on how to improve(or at least change) your tests dramatically

c'mon, i'm eager to see them.

Quoting zznbrm
If anyone is interested in the full results, I will post the entire result set. Like I said previously, my only deviation from Post 1 rules was the currency percentage change calculation.
  • #186
  • Aug 11, 2011 12:01pm Aug 11, 2011 12:01pm
Hi,

Come back from a beach day and.....wow, demo account 10% down, really happy for that, the 3 day of test and the correction day, really early, now we have to think how to avoid that.......maybe a hard SL and a trailing SL for each trade .........
  • #187
  • Edited 1:54pm Aug 11, 2011 12:38pm | Edited 1:54pm
Quoting charvo
great stuff, ZZ. Thanks for posting these.

Could you please post all of ur findings? I'd definitely provide a shocking top secret on how to improve(or at least change) your tests dramatically
Hi charvo!

Attached are detailed results for 5-month test. I've also attached summary results for 2010-2011.

There really is a lot of data. I've done my best to verify that my test script is correct and the resulting data is correct. It would be nice to see what others come up with.

Edit: I did a couple other tests for the Jan 2010 - Aug 2011 timeframe:

Everything the same except I used 1 week look back period:

$30,000 went below $0 in July 2011.

Everything the same except I reversed buy/sell (when 10-week trend is up, I sell; when 10-week trend is down, I buy):

$30,000 got as low as $15,000 and closed at the end of last week at $21,000.
Attached File(s)
File Type: xls 52-FBForex2-Summary.xls   22 KB | 395 downloads
File Type: xls FBForexMarchAug2011.xls   355 KB | 421 downloads
  • #188
  • Aug 11, 2011 1:08pm Aug 11, 2011 1:08pm
Thanks a lot, ZZ. as boasted , if interested, you try to allow 0.03 lot of each winning position running instead of being closed on Friday afternoon (i assume that 0.1 lot is used).

yes, let them running for free, since you have 5 months only, not quite a long time. Or take profit at 2000 pips? the equity curve should be shocking. that's my small secret.

btw, of coz u close all if by Friday u have no winners.

edited: wow, the 1 year result looks pathetic, pondering the logic again, so is 2010 a year without trends...

Quoting zznbrm
Hi charvo!

Attached are detailed results for 5-month test. I've also attached summary results for 2010-2011.

There really is a lot of data. I've done my best to verify that my test script is correct and the resulting data is correct. It would be nice to see what others come up with.
  • #189
  • Aug 11, 2011 2:32pm Aug 11, 2011 2:32pm
Quoting zznbrm
Hi charvo!

Attached are detailed results for 5-month test. I've also attached summary results for 2010-2011.

There really is a lot of data. I've done my best to verify that my test script is correct and the resulting data is correct. It would be nice to see what others come up with.

Edit: I did a couple other tests for the Jan 2010 - Aug 2011 timeframe:

Everything the same except I used 1 week look back period:

$30,000 went below $0 in July 2011.

[b]Everything the same except I reversed buy/sell (when 10-week trend is up, I sell;...
Thanks for the extensive testing... not really encouriging... holy grail still remains to be found

Your test demonstrates that for a start, lot sizes are too high... as mentioned before 0.01 lots per 10.000 seems more appropriate...

But still... seems the system as originally thought off won't just work out of the box...

Will have to look into it...
  • #190
  • Aug 11, 2011 4:11pm Aug 11, 2011 4:11pm
Quoting FBForex
But still... seems the system as originally thought off won't just work out of the box...

Will have to look into it...
The problem is that risks are not balanced, we bet all our money on one horse.
I still think that it is necessary to bring in elements of hedging using correlation between pairs. For example, if there is long EURUSD in the basket we would like to have either short EURGBP or short GBPUSD involved to hedge the risk.
  • #191
  • Aug 11, 2011 9:03pm Aug 11, 2011 9:03pm
Quoting Indrek
The problem is that risks are not balanced, we bet all our money on one horse.
I still think that it is necessary to bring in elements of hedging using correlation between pairs. For example, if there is long EURUSD in the basket we would like to have either short EURGBP or short GBPUSD involved to hedge the risk.
I'll save you all alot of time right now... hedging will not change a thing. It will only prolong the process of reaching profit X or account blow up. You will only spin your wheels if you start coming up with hedges. Sure drawdown will be lower, but so will profits. Net effect will be the same. Reward is proportional to risk. Hedging is not the solution you are looking for.
  • #192
  • Aug 12, 2011 3:03am Aug 12, 2011 3:03am
Will keep posting daily results... just to be sure...

Open profit has dropped as CHF dropped against all currencies. As 2+2=4FX pointed out, hedging is not an option. Better solution would be to adjust lot sizes to lower the risk.
Attached Image (click to enlarge)
Click to Enlarge Name: Mt4.jpg Size: 165 KB
  • #193
  • Aug 12, 2011 4:20am Aug 12, 2011 4:20am
Just tested an idea.

I took zznbrm's 5 month results and decided that if total profit for a running week would become negative, all trades are closed and trading will start only next week. As you can see result improves a lot.

This idea seems logical, since if total position turns negative, currencies are not trading according to strength and it makes no sense keeping positions open.

@zznbrm: Would it be possible to post your results from 1-1-2010 till now applying this idea?
Attached Image
  • #194
  • Aug 12, 2011 4:32am Aug 12, 2011 4:32am
I also wonder how it's possible that on my Demo account last week had a profit of € 2767.58 and your results show only $ 259,64 profit? This seems very odd????
Attached Image (click to enlarge)
Click to Enlarge Name: Mt4.jpg Size: 239 KB
  • #195
  • Aug 12, 2011 5:32am Aug 12, 2011 5:32am
Quoting FBForex
Just tested an idea.

I took zznbrm's 5 month results and decided that if total profit for a running week would become negative, all trades are closed and trading will start only next week. As you can see result improves a lot.

This idea seems logical, since if total position turns negative, currencies are not trading according to strength and it makes no sense keeping positions open.

@zznbrm: Would it be possible to post your results from 1-1-2010 till now applying this idea?
Can you please explain in a little more detail please? I thought you closed all positions at the end of the week irrespective of whether it is in profit or loss??
Now trading live
  • #196
  • Aug 12, 2011 5:35am Aug 12, 2011 5:35am
Quoting FBForex
Just tested an idea.

I took zznbrm's 5 month results and decided that if total profit for a running week would become negative, all trades are closed and trading will start only next week. As you can see result improves a lot.
Good idea, ie. wait out the storm, don't ride it to the end of the week.

Another thing: I notice your largest few winners in your table are always a good multiple in size of the largest winners (very good RR ratio). Prob. also a function of your cutting half the losers' size daily.

But also, once you get big winners, like you did for half of this week from 10-wk strength asserting itself, with the rush to safety (ie. into CHF and JPY) and out of stocks, can you put a wide stop on the largest winners? It can be pretty far away, say enough to grab half the dollar amount (or EUR amount in your case) of the winners (still gives you good RR because your winners are so much larger than your losers). This way you don't get so hit by a risk reversal (back into stocks, out of CHF and JPY) half way through the week, as happened this week and the rumour of a EURCHF peg (negative for CHF strength, as it implies a fixed link with a weak EUR).

The point is: cutting half the dollar (or EUR, in your case) losses/losers daily, and putting a wide stop on the big winners may be still give you a good/robust RR on balance or in net terms, because your winners are so much larger than your losers (a big multiple of the losers in size). Then again, this won't help if your prev. strongest 10-wk major, the CHF, goes into the tank from Monday am onward or tues etc. (the whole week retraces from previous CHF strength). Negative from the start of the week would take you out of the week entirely on Mon., but strength could then reassert itself mid-week (the opposite of what happened this week). Life ain't easy, LOL.
  • #197
  • Aug 12, 2011 5:45am Aug 12, 2011 5:45am
Quoting FBForex
Just tested an idea.

I took zznbrm's 5 month results and decided that if total profit for a running week would become negative, all trades are closed and trading will start only next week. As you can see result improves a lot.

This idea seems logical, since if total position turns negative, currencies are not trading according to strength and it makes no sense keeping positions open.

@zznbrm: Would it be possible to post your results from 1-1-2010 till now applying this idea?
When did you determine when to close for a loss? Each day at a specific time, like when you cut losers?

Also, all your winning weeks seem to match the original test. That tells me at no time during those winning weeks the account ever went into negative? Because if it did you would have to close the whole basket. Im finding it little hard to believe that in all those weeks of testing each day finished with a positive balance. Just a thought.
  • #198
  • Aug 12, 2011 6:01am Aug 12, 2011 6:01am
Quoting thorn
When did you determine when to close for a loss? Each day at a specific time, like when you cut losers?

Also, all your winning weeks seem to match the original test. That tells me at no time during those winning weeks the account ever went into negative? Because if it did you would have to close the whole basket. Im finding it little hard to believe that in all those weeks of testing each day finished with a positive balance. Just a thought.
You're correct. According to zz's test results in positive weeks at no time during the week (at least at the end of each day) the total position for that week (including closed trades) went negative. It stayed in profit for the whole week.

Would be great to get the test data from zz for 2010 and 2011 to verify if this method is solid. Also another point I've been thinking of during the past week is what results would be if we close all losers at the end of the day (at 20.00 hours Central European time that is, markets are likely to have settled and it's a suitable time for me to look at the positions) instead of just closing half the losing position. Data from last 5 months suggest that if a position goes negative it's likely to become more negative during the week or if it turns around it does just marginally.
  • #199
  • Aug 12, 2011 8:18am Aug 12, 2011 8:18am
Attached is a test for the last 5 months, again using data provided by zz with closing all negative positions at the end of the day (first two columns) and with closing all negative positions at the end of the day and if total profit of the week turns negative (two yellow columns).

Seems that closing all negative positions at the end of the day isn't a good idea. It seems that positions that are marginally negative can turn around for some nice profit at the end of the week.

If zz could provide all data of all positions (including negative positions) untill the end of each week (currently data is erased when position falls in largest negative positions of half the negative amount) we could see if just keep the negatives till the end of the week is even better.

Hope we can figure this out...
Attached Image
  • #200
  • Aug 12, 2011 8:42am Aug 12, 2011 8:42am
Quoting FBForex
I also wonder how it's possible that on my Demo account last week had a profit of € 2767.58 and your results show only $ 259,64 profit? This seems very odd????
- I closed trades each day at 20:00 Central European time, you did not. Pay particular attention to Friday of last week. Almost every trade retraced from the time you closed until 20:00.

- As stated, I used a different (although, I would have to assume similar) currency percentage calculation. For last week, all of my trade directions were the same as yours except EURCAD.

- I'm backtesting using H1 data, fixed spreads, and no swap.

Why H1 data? Because the strategy calls for all trading decisions to be made at the top of the hour (08:00, 20:00). So H1 data is sufficient for these purposes.

Why fixed spreads? Because MT4 does not store historic ask prices. So, I chose "conservative" spreads for calculations. I could have used historic tick data from a different source, but that adds complexity to the code, takes longer to process, and is not necessary for these tests.

Why no swap? Because it is really not necessary at this stage (trying to get a general idea of past performance). If these backtests start to show more positive results, then further backtests could be "fine tuned" to be more precise. This system is not at that point yet.

Backtesting is not perfect and has never been advertised as such. However, it serves a very important purpose - to judge system performance under different circumstances in a very fast manner. Forward testing of the same system would take much, much longer.
0 traders viewing now