Supernova System Journal | Page 7 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Another Trading Journal - Yuhu's Journal 17 replies

Supernova GBP/JPY Mini Trend Catcher 6,405 replies

My Personal journal up 18% in 2 months going to start a journal 8 replies

mphpopular supernova manual journal 194 replies

Indicators:SuperNova !!! and Nova! 2 replies

Attachments: Supernova System Journal
Exit Attachments

Supernova System Journal

  • #121
  • Sep 28, 2007 3:16pm Sep 28, 2007 3:16pm
Quoting Gnat
Victims of circumstances (not that I am a Barclay James Harvest decipel in any way) ... my mother still does not understand that the boy who never did his homework and connected one sparetime to the other became practically a work-o-holic.

But I hope that you really do not mind me putting that forward on a Friday. And someting else I realized... I HAVE NO BEER IN DA HOUSE!! Jaul!!

Your test actually brought me to check the charts and see the difference between the 2 worlds. So it is your fault in the most positive way here!

Have a good evening and beautiful weekend. Lyrical leadress of the army of bots

Good Night and good testing! I will try it myself again here.
Gnat, The Fly

bsssssssssssssssssssssssss
G'night to u2 and see ya tomorrow
  • #122
  • Sep 28, 2007 3:21pm Sep 28, 2007 3:21pm
Quoting Qpip
G'night to u2 and see ya tomorrow
Bye bye... see ya tomorrow
  • #123
  • Sep 28, 2007 3:24pm Sep 28, 2007 3:24pm
Here is my testing preparation. As per the document provided by cryten, I installed yet another instance of MT4 (I now have 5 LOL). I deleted the original account, set the #bars to 99999999 and 9999999 and deleted the *hst files from the history folder.

Then the instructions called for getting the data from Alparai, but the link didn't work and I noticed that "download" wasn't available at the time of writing (sept 2006) whereas it is available in the most recent version of MT4.

So I simply downloaded the 1 min (took about 10 minutes) and didn't need to convert anything (again as per the instructions) since as you can see from the data in my subdirectory, everything is all there (lookat da size of the 1 min data Ma Ma Mia).

I will then proceed in the same manner for all the other pairs. Then will start the actual testing. I've got data dating back from 1999 and I wonder if this is useful.... Maybe I'll just start with jan 1 2007 - aug 31 2007 like I did last week-end and see if things differ/correlate. Then I'll go back in time (with what time I have left up to sunday evening !!!).
Attached Image
  • #124
  • Sep 28, 2007 6:52pm Sep 28, 2007 6:52pm
Alright, tests not working at all. Dunno what the problem is. I start the test and then (I watch the visual) and it gets stuck and with the price zipping up down up down up down endlessly.

I cut a few nines from my 9999999 setting, thinking maybe that was too much, but then it goes through the test with results = zero.

Will look into it tomorrow with a clearer head. If anyone else is performing these tests over the week-end, please advise if it’s working out right or not, THANKS!
  • #125
  • Sep 29, 2007 8:42am Sep 29, 2007 8:42am
OK everything is working now, rebooted, reinstalled, restarted, re-everything, and now the tester works.

So I'm now proceeding with the tests as mentioned.

nb: Gnat - you mention weekly tf and from what I've downloaded, it seems I've got weekly and monthly data. However, the MT4 tester only goes to daily. Any way around this?

Am currently testing 9 pairs + 7 tf (1) with filters (2) dot to dot in 2 "manual" directions (if I have the patience )

Will post results as soon as (1) is completed.
  • #126
  • Sep 29, 2007 9:40am Sep 29, 2007 9:40am
Ma'am Sahib!

I hope it did not spoil your evening yesterday...
I have real problems to test the Supernova.ex4 also. There are so many bugs reported, that the longest file in the end is the journal!

I have the Glenn EA on test right now (4H). It needs the attached indicator as it seems. Not really needed for this concept, but the EA earns some money in a 55% test.

The take profit target and the dimension of SL is a big problem as it seems. The test are running around me and I see all these pips go to ... someone else! Hmpf!

I will send over some reports and settings as soon as I have something finished. Visual testing is lasting too long, but it is good to see conceptual flaws.

Prost!
Gnat
Attached Image
Attached File(s)
File Type: ex4 Damiani_volatmeter v3.2.ex4   5 KB | 287 downloads
  • #127
  • Sep 29, 2007 10:36am Sep 29, 2007 10:36am
I'm halfway through my testing. Everything is going very VERY well. 90% accuracy almost across the board.

And this time, the dot-to-dot approach actually works perfectly, with trades being taken on both sides.

Should be able to post final results within an hour or 2.

Arrivédertshy (sp),
  • #128
  • Sep 29, 2007 10:38am Sep 29, 2007 10:38am
Quoting Gnat
Ma'am Sahib!

I hope it did not spoil your evening yesterday...
I have real problems to test the Supernova.ex4 also. There are so many bugs reported, that the longest file in the end is the journal!

I have the Glenn EA on test right now (4H). It needs the attached indicator as it seems. Not really needed for this concept, but the EA earns some money in a 55% test.

The take profit target and the dimension of SL is a big problem as it seems. The test are running around me and I see all these pips go to ... someone else! Hmpf!

I will send over some reports and settings as soon as I have something finished. Visual testing is lasting too long, but it is good to see conceptual flaws.

Prost!
Gnat

Hola amigo Gnatto,

Is the Glenn EA the same as the one cryten posted on post.... hmmmm can't remember but that's the one I've been using ever since the beginning.

If they're different, it'll be interesting to compare our results.

See you in a bit,
  • #129
  • Sep 29, 2007 10:42am Sep 29, 2007 10:42am
Holla Infantin!

I am testing this last EA from winsteadglenn on a 4H TF and it is showing rather mixed results. It seems that a parabolic trailing stop would really be a good idea - but I am not yet sure.

Really looking forward to your findings!!

See you in a bit then!

Gnatjetto
  • #130
  • Sep 29, 2007 12:59pm Sep 29, 2007 12:59pm
OK the results are now compiled. My computer has SMOKE coming out of the back heh.

First of all, this time around as I mentioned earlier, the dot-to-dot strategy worked perfectly and trades were taken in both directions. I did some visual checking and both long and short trades were taken after closing the previous long/short trade. This wasn’t the case with last week’s tests where only trades in 1 direction were taken. Maybe this is due to having more ticks tested? It seems that some people this week experienced the latter results, ie trades not being taken after closing (using dot-to-dot). We’ll need to forward test this very thoroughly to see if in fact the EA does work ok in that respect.

Second point is that modelling quality (mq) was 90% across the board except for a few exceptions, namely the 1 mins.

You can go here if you want to compare with last week’s results#77

As you can see, although numbers vary, in 7 out of 9 cases for dot-to-dot and 5 out of 9 for filters, the optimal time frame remains the same. I put a box around the cells which had a different “best tf” in the last tests. Most were 1 tf beside the new results.

Clearly higher timeframes have much better results than lower time frames. And again, filters have a better performance than dot-to-dot on slightly shorter time frames. However, the results for 5m and 1m (this one has 25% accuracy) are so BAD that one might think of reversing the strategy in order to get those amazing results but backwards !

I put in bold which results between dot-to-dot and filter were the best final results. One may want to reflect on the fact that 7 out of 9 pairs have better results without filters. And if we eliminate EURUSD which had negative results, that’s 8 out of 9 with better performance without filters. I’m not real happy with this fact, as filters are supposed to act as confirmation / reassurance. Hmmm.

The actual monthly # pips are higher than what we had for the last tests (which is good ) and so are the TFs in general (many 4H and 1 daily).

Finally, in the 2nd table, I’ve put up the pairs ranked in order of performance, with best tf and filters/no filters and monthly pips. Please refer to post 78 for comparison purposes. Last week I hadn’t added all the potential monthly pips combined for all the pairs, but the total is 1,295. The results for this week (you may let out a tonitruant (loud) MA MA MIA!) is 1,949 .


http://i22.tinypic.com/30vc2sg.jpg

http://i21.tinypic.com/30icbv5.jpg
  • #131
  • Sep 29, 2007 1:17pm Sep 29, 2007 1:17pm
Hallo and already good evening from old Europe!

Testing the GBPJPY on a 4H basis drove me a little nuts today, as there is no such thing as a valid exit strategy by now. I will go to the winning settings once more on visual mode (this takes hours!) in order to find out what the EA misses and where.
A fact is (to me), that until now you have to beat the bots with up to no effort. But maybe my settings are too radical there. No SL or large=no SL does far better, then a safe guard strategy.
A trailing stop seems inevitable, but how about the dimensions? If it is too tight, your winners stay small and the loosers, see above, will melt down 10 winners with a grin.
Unfortunately the dot-to-dot strategy, as clear as it may seem on screen, seems to be too hard to handle for the EAs (Glenn 1_1, original Supernova, uconns supernova 1_1) - the only one left is the newest Spernova 1.2! What exactly do you use at the moment?
I will attach a test that ended positive with some comments on it. Could you send out a setting file for your EA used in the above tests?

Working on
See ya then!
Gnat

80% of the losing trades are SELLS!
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Test Glenn 1-1 4H Sep29.jpg Size: 148 KB
Click to Enlarge Name: Test Glenn 1-1 4H Sep29 pic.jpg Size: 51 KB
  • #132
  • Sep 29, 2007 1:41pm Sep 29, 2007 1:41pm
Well... what can I say...

due to optimization ... the best setting is NO SETTING.

Still - there must be a way to avoid the losing trades. If we could somehow eliminate those, the pippage will easily double.

See the enclosed

Greetings
Nick Gnatterton
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Test Glenn 1-1 4H Sept29 No limits.jpg Size: 97 KB
Click to Enlarge Name: Test Glenn 1-1 4H Sept29 No limits pic.jpg Size: 34 KB
  • #133
  • Edited 2:50pm Sep 29, 2007 2:01pm | Edited 2:50pm
Gnat! Quit fiddling about and go see my results in post 273. I got better results than you gna, gna : (or gnat, gnat?). If instead of backtests, you and me would have actually traded using those settings, we’d both be :

(1949 * 8) * 0.87 = 13,565.04 $ USD richer using mini lots or
(1949 * 8) * 8.7 = 135,650.40 $ USD richer using standard lots

I know, this is theoretical, doesn’t take spreads into account, proper capitalization and all that, but still... quite nice.

I’ll post the report for GBPJPY on the 4H in a second so you can see the difference.

EDIT: it's post 130, not 273 (the journal doesn't even go that far LOL)
  • #134
  • Sep 29, 2007 2:06pm Sep 29, 2007 2:06pm
GBPJPY 4H

http://i23.tinypic.com/nq8i8k.jpg

http://i20.tinypic.com/j8ds9e.jpg
  • #135
  • Sep 29, 2007 2:22pm Sep 29, 2007 2:22pm
Quoting Qpip
GBPJPY 4H

http://i23.tinypic.com/nq8i8k.jpg

http://i20.tinypic.com/j8ds9e.jpg
This really looks nice... This is representing the same timeframe as my test with the winsteadglenn Glenn1-1 EA?

Fiddling about... fiddle about - fiddle about
I'm your wicked Uncle Ernie (do you know the song from "Tommy" by the Who?) Came to my mind. I have an exceptionally strange mind , but I guess you are more than equal!

Well - there is a concept called "soft landing". If this one could take care of our losses, there will be even more than the nice looking $$$ you printed above. I will get on the phone with someone in Japan I know, who may be able to tell me, whether this may match here or not. Guess by tomorrow.

Still a combination of the heap of currencies that qualify for this concept makes this already a very profitable ... reason to sack whatever is a pain in the #++!!

273 in the Supernova thread shows Moneda bowing to Cryten. LOL ... no, I surely know that tableu you referred to. 73 here... it is sticking on my wall here.

OK. This works over the years and is doing better in higher volatility. Real money now??

Erstmal!
Gnat
  • #136
  • Sep 29, 2007 2:45pm Sep 29, 2007 2:45pm
Quoting Gnat
This really looks nice... This is representing the same timeframe as my test with the winsteadglenn Glenn1-1 EA?
Yes, I made sure we were both covering the same period, ie jan1 to aug 31, 2007.

Quoting Gnat
Well - there is a concept called "soft landing". If this one could take care of our losses, there will be even more than the nice looking $$$ you printed above. I will get on the phone with someone in Japan I know, who may be able to tell me, whether this may match here or not. Guess by tomorrow.
Hmmmmmmm WAZZAT? I'll google it. Sounds in----teresting.

Quoting Gnat
273 in the Supernova thread shows Moneda bowing to Cryten. LOL ... no, I surely know that tableu you referred to. 73 here... it is sticking on my wall here.
Nah I meant post #130 (73 ?!?!?! <-- all mixed up in post # and we're both deeply involved in backtesting NUMBERS Get THAT one up on your wall (new results, you skipped over that post)

Quoting Gnat
OK. This works over the years and is doing better in higher volatility. Real money now??
That IS the question, my friend.

For now, I'm compiling (I should be a program heheh) data from the best pairs/tfs/filter-or-not of my new results and will intertwine them as they happened in time (this will take some serious cut and pasting oh my god, better crack a beer open to keep me company while I slowly go insane POUARF) the objective being to see if the balance ever goes down the toilet or if it just constantly goes up (I'll make a graph and post it). I also want to check by doing this how much capital would actually be needed (the minimum that is).
  • #137
  • Sep 29, 2007 3:08pm Sep 29, 2007 3:08pm
You know what I'm thinking, I think what I'll do next, now that TFs and filter/no filter is out of the way, I'll test the SLs, trailing stops and best combination of both. I may even add TPs in there at some point. Something along those lines, filling in the table below:
Attached Image (click to enlarge)
Click to Enlarge Name: stops etc.jpg Size: 31 KB
  • #138
  • Sep 29, 2007 3:12pm Sep 29, 2007 3:12pm
Quoting Qpip
Yes, I made sure we were both covering the same period, ie jan1 to aug 31, 2007.



Hmmmmmmm WAZZAT? I'll google it. Sounds in----teresting.

Nah I meant post #130 (73 ?!?!?! <-- all mixed up in post # and we're both deeply involved in backtesting NUMBERS Get THAT one up on your wall (new results, you skipped over that post)

That IS the question, my friend.

For now, I'm compiling (I should be a program heheh) data from the best pairs/tfs/filter-or-not of my new results and will intertwine them as they happened in time (this will take some serious cut and pasting oh my god, better crack a beer open to keep me company while I slowly go insane POUARF) the objective being to see if the balance ever goes down the toilet or if it just constantly goes up (I'll make a graph and post it). I also want to check by doing this how much capital would actually be needed (the minimum that is).
I guess the soft landing is somehow connected to the Turtles. But it may as well be something I picked up on the stock market. In simple terms it is a sort of retracement stoploss. If the price reaches a certain lever against your position, you will close it after XX% of retracement. So you acknowledge the loss, but the trick is, you do not take it at the peek. The concept seems to be deeply rooted in the portfolios of banks and insurances.

#73, #130, ... Hihi... : ... that one would have replaced it, if I had not been too lazy to store it on the USB Stick and carry it over to the Email and Printing composter. I keep such things away from my tradingmachines.

I still do not have beer here... another induced thirst!! As for the money, I guess even building an account for a few months and then do the sacking step would do the job ...

Patiently waiting for your excellent compilations - I guess we should get it started. Not that we end up as your footer leads us ...

Every year is getting shorter never seem to find the time.
Plans that either come to naught or half a page of scribbled lines

But NOT THIS TIME!
See ya!
Gnat

And as soon as there is good money we rent some place in London or New York or even Edmonton (LOL) and have a members party! Apart from the golddust the plan sounded really promising!!
  • #139
  • Sep 29, 2007 3:15pm Sep 29, 2007 3:15pm
Hey imrainsait!

Have you been searching for your own name and found it here? Nice short term concept you run there.
Do you like ours? Only we left the short term to you by now.

Very welcome!
Gnat
  • #140
  • Sep 29, 2007 3:36pm Sep 29, 2007 3:36pm
Quoting Qpip
You know what I'm thinking, I think what I'll do next, now that TFs and filter/no filter is out of the way, I'll test the SLs, trailing stops and best combination of both. I may even add TPs in there at some point. Something along those lines, filling in the table below:

You post something and I do not see it for a full 20 minutes... I swear that my two last posts were directly under the one I cited. Then your post comes in and it is 12 minutes younger than mine.

AMAZING !! Maybe I sneezed out my brain...

Going after the SLs and trailers may actually round up the thing. I guess I will try it too. (22:34 CET) I will surprise you with the result for breakfast then!

Bye Birdie!
Gnat the frettled grindlewordl
1 trader viewing now