Snow Roller - An Avalanche of Profits | Page 9 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

MT4 (Metatrader) installation instruction for Snow Leopard 10.6 Mac 102 replies

The Avalanche 13 replies

adjusted avalanche //tremor ea --- feedback 0 replies

Winning on the emotional roller coaster 6 replies

Stocks Markets Avalanche prelude to USD rally 0 replies

Attachments: Snow Roller - An Avalanche of Profits
Exit Attachments

Snow Roller - An Avalanche of Profits

  • #161
  • Apr 6, 2011 12:21am Apr 6, 2011 12:21am
Check the usual location for the newly created SnowRoller_1_002. This version has some unique features and builds upon the base of 1_001.

You will notice that the default values for TrendLookbackMins & TrendLookbackMins2 are now 5 and 100. This is from testing on the 1 minute time frame multiple symbols with the other changes made.

Instead of exiting completely when the trend condition changes, the EA scales out as is done in SnowRoller_Classic. The settings OOTB are tuned for 1 minute and testing was done on multiple symbols separately and they seem to do well so unlike version 1.001 where I have the USD pairs running 200/0 and JPY INVERTED running 200/1500, hopefully one size will fit all with this EA.

I have both running and will evaluate things tomorrow. Depending on which version works better I'll add the time of day code to that EA. I have 1.000 on 5 min with USD & JPY inverted as described above, 1.001 on 1 min and 5 min with USD & JPY inverted, and 1.002 running with OOTB.
  • #162
  • Apr 6, 2011 1:56am Apr 6, 2011 1:56am
  •  sym
  • Joined Apr 2011 | Status: Working... (too much) :) | 356 Posts
Quoting FXEZ
Check the usual location for the newly created SnowRoller_1_002. This version has some unique features and builds upon the base of 1_001.

You will notice that the default values for TrendLookbackMins & TrendLookbackMins2 are now 5 and 100. This is from testing on the 1 minute time frame multiple symbols with the other changes made.

Instead of exiting completely when the trend condition changes, the EA scales out as is done in SnowRoller_Classic. The settings OOTB are tuned for 1...
Hi FXEZ
I want to thank you for your work.
There are few minor details I don't like in this EA, but never mind, it is great.

Have a nice day
--- nothing to say ---
  • #163
  • Apr 6, 2011 2:23am Apr 6, 2011 2:23am
Hi, I started to test your EA two days ago. OOTB on several pairs. I experienced bailout twice (equite fell by 30% and EA closed all trades)

My observations are:
- it looks like it is counter trading although you are trying to implement trend trading. It ends up as classic grid with extra dangerous error that it keeps opening counter trend trades at similar levels. You can see in the picture below that although trend in all pairs is obviously up, the EA keeps opening short positions in most cases.

- you have logical error in your closing routine. Once the bailout occurs the EA keeps opening and closing trades right away. It is because you compare MaxEquity with current equity. Current equity after bailout is 30% less than before so once the EA opens new trade it closes it right away. You can see that on the picture below where after bailout there is a long flat line going down. It is opening and closing trades making series of small losses.

PHP Code
 if (equity - MaxEquity >= MaxEquity * (CloseAllProfitPercent/100.0) || equity - MaxEquity <= -MaxEquity * (CloseAllLossPercent/100.0)) {
      ClosingAll = true; 
      CloseAll();
Imagine max equity before bailout was $10000 and after the equity=7000, then this condion is always true, since
equity - MaxEquity is -3010 (because we have open trade, which is immediately in loss)
-MaxEquity * (CloseAllLossPercent/100.0)) is -3000 and it is >= than -3010

I think it should be calculated using balance and current equity like this
PHP Code
 if (equity - AccountBalance() >= AccountBalance() * (CloseAllProfitPercent/100.0) || equity - AccountBalance() <= -AccountBalance() * (CloseAllLossPercent/100.0)) {
Then it reflects current situation. I am not sure what has MaxEquity to do with it.

This code
PHP Code
if (AccountEquity() > MaxEquity) GlobalVariableSet("SRMaxEquity", AccountEquity());
Does not apply after bailout, so MaxEquity remains the same high number.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: SnowRoller.jpg Size: 232 KB
Click to Enlarge Name: statement.gif Size: 5 KB
  • #164
  • Apr 6, 2011 4:33am Apr 6, 2011 4:33am
Backtesting V1.02 with OOB settings over €$ from 01-01-2011 till today, results in a complete bailout.

Just opened a new demo account with 10K€ and forwardtesting it over the 11 pairs of post #1 to see what happens in next 24h
  • #165
  • Apr 6, 2011 8:21am Apr 6, 2011 8:21am
Version 2 result has more better see ....

Can you give me best set file?

bye
ajentha
Attached File(s)
File Type: doc SnowrollerV2.doc   169 KB | 299 downloads
  • #166
  • Apr 6, 2011 9:16am Apr 6, 2011 9:16am
From +2039€ to -2341€, since yesterday 19.00 GMT till now, it's a huge DD, because it was "adding" to looser trades and now that it close them ... Account balance goes down.

Still testing .. now also V1.02, let see.
  • #167
  • Apr 6, 2011 9:43am Apr 6, 2011 9:43am
Quoting rbetancor
From +2039€ to -2341€, since yesterday 19.00 GMT till now, it's a huge DD, because it was "adding" to looser trades and now that it close them ... Account balance goes down.

Still testing .. now also V1.02, let see.
Agree the same in my 2 acocunts.

Niki
  • #168
  • Apr 6, 2011 10:06am Apr 6, 2011 10:06am
3 out of my 4 demos are in profit. Not much profit but at least not floating with large losses. I just watched v 1.002 close out a basket as the JPY pairs were in trend. So far changing the JPY pairs on 1.00 and 1.001 to 200/1500 TrendLookbackMins/TrendLookbackMins2 seems to be working better.
  • #169
  • Apr 6, 2011 10:08am Apr 6, 2011 10:08am
Quoting ajentha
Version 2 result has more better see ....

Can you give me best set file?

bye
ajentha
I'm just using the defaults. And I like what I see (so far) on v1.002.
  • #170
  • Apr 6, 2011 10:46am Apr 6, 2011 10:46am
Thank you for your best effort, try and try we can get best,

Now your EA result is revers, see


bye
ajentha
Attached File(s)
File Type: doc SnowrollerV2.doc   95 KB | 316 downloads
  • #171
  • Apr 6, 2011 10:50am Apr 6, 2011 10:50am
I just started a new demo for 1.002 in addition to the ones I'm currently running. I'm going to try something a bit different:

time frame = 5 minutes
doOpposite = true: EURJPY, GBPUSD, AUDJPY, USDJPY, EURCHF
doOpposite = false: EURUSD, GBPJPY, AUDUSD, CHFJPY, USDCHF, GBPCHF, EURGBP, USDCAD

The theory is that by creating offsetting positions in AUD, GBP, EUR, CHF between USD and JPY the initial losses taken when positions are initiated will be less. That's the theory. Once the snow roller effect kicks in, the side that has the trend going should add more positions and we get accumulation on the profitable positions while the losing positions are position locked.

It's not a perfect hedge but for 13 pair the net outright position with this setup is according to my rough calculations: 0 USD, 1 JPY, 0 CHF, 0 AUD, 0 EUR, 1 CAD, 0 GBP.
  • #172
  • Apr 6, 2011 10:58am Apr 6, 2011 10:58am
Quoting FXEZ
I'm just using the defaults. And I like what I see (so far) on v1.002.
Interestingly enough, I don't seem to be making much headway with v 002. I've had it on since last night and just seems to stay flat or a little negative. I'll keep it running to see if it picks up steam soon.
v 1.00 seems to be cranking out some profit for the last 24 hrs tho, all that's changed is the 200/1500 for jpy pairs.

Will keep em both running and see what becomes of it.
  • #173
  • Apr 6, 2011 11:27am Apr 6, 2011 11:27am
Quoting tazmaan
Interestingly enough, I don't seem to be making much headway with v 002. I've had it on since last night and just seems to stay flat or a little negative. I'll keep it running to see if it picks up steam soon.
v 1.00 seems to be cranking out some profit for the last 24 hrs tho, all that's changed is the 200/1500 for jpy pairs.

Will keep em both running and see what becomes of it.
After I posted that, the equity had just popped, then new positions opened up and it went into a fairly good DD right off the bat. Thus the new demo with mostly hedged pairing to prevent the initial giveback when opening positions and determining the trend direction.
  • #174
  • Apr 6, 2011 12:14pm Apr 6, 2011 12:14pm
Quoting hotstorm
- you have logical error in your closing routine. Once the bailout occurs the EA keeps opening and closing trades right away. It is because you compare MaxEquity with current equity. Current equity after bailout is 30% less than before so once the EA opens new trade it closes it right away. You can see that on the picture below where after bailout there is a long flat line going down. It is opening and closing trades making series of small losses.
Thanks for pointing this bug out. I'll have to fix this. It appears to affect all versions. I meant to go back and change that line but forgot. I'll have new versions by tomorrow uploaded.
  • #175
  • Apr 6, 2011 2:30pm Apr 6, 2011 2:30pm
FXEZ:

Here is my link to v 1.00 m5 with jpy pairs 200/1500, rest default.
It made it thru the first 24hrs very nicely, so figured it warranted a myfx link
Max open dd thus far just under 4%, and all trades just closed a bit ago and it's starting its cycle again.

http://www.myfxbook.com/widget?id=10...pe=1&color=red
  • #176
  • Apr 6, 2011 8:50pm Apr 6, 2011 8:50pm
I shouldn't have kept overwriting the previous versions. One of the versions was doing quite well but I overwrote it with a newer version. Pity....
  • #177
  • Apr 7, 2011 12:48am Apr 7, 2011 12:48am
Quoting tazmaan
FXEZ:

Here is my link to v 1.00 m5 with jpy pairs 200/1500, rest default.
It made it thru the first 24hrs very nicely, so figured it warranted a myfx link
Max open dd thus far just under 4%, and all trades just closed a bit ago and it's starting its cycle again.

http://www.myfxbook.com/widget?id=10...pe=1&color=red
I hope it warrants more than a single day's evaluation. So far my demo with jpy pairs 200/1500 for 1.00 m5 is hanging in with a small equity DD. My demo on m5 1.001 with the jpy pairs 200/1500 has hit the 30% CloseAll *twice*. The first time it recovered back to $5k then now it's back down again.

My M1 1.001 with the JPY 200/1500 is floating slightly under water on equity, as are my M1 and M5 with the jpy switch of version 1.002.

It will be very interesting to see these in action when the markets start to trend. 1.002 should not act like a martingale. Right now it's acting like (as are some of the others) anti-martingale with small losses. A bit frustrating but at least not blowout bad.
  • #178
  • Apr 7, 2011 12:56am Apr 7, 2011 12:56am
I'm anticipating slowing down on the updates and doing more evaluation & mixing of symbols & other variables, but as this is a bug fix, and a feature add (that has been widely requested) I went ahead and did it for all 8 files (4 EA versions of SnowRoller).

The bug fix is the one that effects the CloseAll routine when a loss is taken. Thanks to hotstorm for pointing this out. I'm still using equity for the pops otherwise you can end up with balance pops that never get you ahead, if the pop is based on a fixed % of balance (which is declining due to the nature of the anti-grid).

The feature addition is the ability to set start and stop times. I did a full writeup of the update here.
  • #179
  • Apr 7, 2011 5:49am Apr 7, 2011 5:49am
Thank a lot, very quick updation

bye
ajentha
  • #180
  • Edited 6:26am Apr 7, 2011 6:02am | Edited 6:26am
Quoting FXEZ
So far so good... http://www.forexfactory.com/attachme...1&d=1301378309

Is this a screenshot of your snow roller?

If it is then it has *NOTHING* in common with my snowball and instead just doing exactly the opposite. The philisophy of snowball is to trade an ANTI-grid (in trend direction!) and what you are showing here seems to be a grid (against the trend), the exact thing I am avoiding with snowball.

Snowball would have opened 4 long positions in the above example and would be in profit while your EA seems to be betting high stakes against the trend and risking ruin!

Your EA is averaging down, adding to losers, Snowball instead will do the exact opposite and add to winners and cut the losers immediately.

Please don't associate this ordinary grid gamble ea with my ANTI-grid EA, since this is doing the exact opposite! If you had read my website then you would have seen what I have to say about grid EAs and averaging down.

You cannot claim to be similar to my snowball EA when you completely negate its philosophy and turn it into the exact opposite and create just yet another grid EA.

 I hereby dissociate myself and snowball from this in the strongest possible terms!

I am not amused!
0 traders viewing now