Daybreak

all pairs have decided to give it a go today:

g/j short @ 122.99
a/u long @ 1.0213
e/u long @ 1.2324
u/j short @ 79.45

double check my work, please

Confirmed here

Daybreak version 1.1 attached.

Changes:

  1. Close open positions at midnight, i.e. candle close
  2. Fixed bug where some orders were placed intraday instead of midnight

Not perfect but getting there.

today’s results:

g/j short @ 122.99, close 123.10, -11
a/u long @ 1.0213, close 1.0185, -28
e/u long @ 1.2324, close 1.2254, -70
u/j short @ 79.45, close 79.34, +11
total -98

since today was a pretty stiff loss, thought it might be a good time to look at july overall:

7-2, -28, +151, net +123
7-3, -49, +5, net -44
7-5, -100, +137, net +37
7-6, -7, +158, net +151
7-9, -48, +44, net -4
7-10, -109, +5, net -104

if i did that right (check my work), we were up +159, lost -98 today, up for the month +61

Read more: 301 Moved Permanently

spirit, your continued efforts are appreciated. ran it on june data. looks like it is now nailing the long entries. no short entries showing up and it is still waiting until the end of the data set to close. long positions appear to be opened correctly during the month, but none were closed until the 29th. then, all were closed at the same time. still profitable, though. lol.

Fix for JPY pairs where EA could not open trades due to Delta being set to 0.0001. Should be Delta * Point but I’ve set it to 0 for now as a quick fix.

Work in progress :8:

pipwoof,
I am aware of disadventages of spreadsheet backtest. It is hard work. On spreadsheet, system looks very good. So, I hope of a working EA for this. I have few ideas to modify the system and to see what works best - to play with SL, and to optimize time of a day to close. But, I have no time and intention to work this throught spreadsheet, too much work.

B2spirit, it is great work with your EA programming. But, not working for my platform (FXCM).

There is a forum for FXCM users that can demand programming request for automated strategies. It is LUA programm language. I made my request, but it can took weeks until it’s done.

happy pips.

thank you so much for all the work you are doing
are you live with this system?
Do you find its working now?
Do you think its risky?
What are you overall results since first trying it?
Thank you so much I really appreciate it

And everyone else, are you liking the system?

Hi Everyone,

I just got done reading this whole discussion and wanted to say keep up the good work.

I just have a question. Does anyone have the count of the max number of loses in a row?

If I can get this information, I may have an idea that could enhance this system. (it is not a trend idea lol)

Also can I get the win % of the data you have now?

Thanks,
Stosh

  1. Fixed another order close bug
  2. No trades on Sundays
  3. Orders on new candles only

Lines 70-73 should look like the code below, else the StopLoss and TakeProfit values will be incorrect.


   if (BuyPrice > Ask)
      OrderSend(Symbol(), OP_BUYSTOP, Lots, BuyPrice, 1, BuyPrice-StopLoss*Poin, BuyPrice+TakeProfit*Poin, "Daybreak", 0, 0, CLR_NONE);            
   if (SellPrice < Bid)
      OrderSend(Symbol(), OP_SELLSTOP, Lots, SellPrice, 1, SellPrice+StopLoss*Poin, SellPrice-TakeProfit*Poin, "Daybreak", 0, 0, CLR_NONE);

spirit, i get openings about right. they are actually right at the high, should be +1. but, closings appear to be coming on the open of next day. still no shorts.

if i calculated yesterday right:

g/j, long @ 123.56, close 123.43, -13
e/u, short @1.2233, close 1.2243, -10
a/u, long @1.0244, close 1.0246, +2
u/j, short @79.19, close 79.59, -40
total, -61

and today:
g/j long @ 123.98
a/u short @ 1.0177
e/u short @ 1.2211
u/j long @ 79.78

i have not started trading this one live.
“working” meaning making lots o’ pips, no.
we are in an arena where some suppose our chances of succeeding less than 10%. less risk at a roulette wheel.
overall results from jan-07 thru jun-2012 are in post 1. i will publish july results after the month.

as offered in post 146, these are our june results for gbp/jpy:

first column after the date is long trade result, then short, then total, then eliminating sunday trades from the total. last column is equity for the pair. ex, on june 4 we got both long and short signals. the spreadsheet will take both. we made +42 on the long trade, lost -83 on the short, for a net -41. it was not a sunday, so total carried over to next-to-last column. equity on the pair ended may at +24,741 and shows as of the end of june at +24,216, for a loss -525. some of the other pairs did better, but still loss for the month.

06/01/12 0 9 9 9 24751
06/03/12 0 0 0 0 24751
06/04/12 42 (83) (41) (41) 24710
06/05/12 44 (137) (93) (93) 24617
06/06/12 140 0 140 140 24757
06/07/12 72 0 72 72 24829
06/08/12 0 (47) (47) (47) 24782
06/10/12 (15) 0 (15) 0 24782
06/11/12 (102) 51 (51) (51) 24731
06/12/12 (29) (123) (152) (152) 24579
06/13/12 (91) 0 (91) (91) 24489
06/14/12 0 (49) (49) (49) 24440
06/15/12 (4) (75) (79) (79) 24361
06/17/12 62 0 62 0 24361
06/18/12 (51) 5 (46) (46) 24315
06/19/12 0 (92) (92) (92) 24223
06/20/12 22 0 22 22 24246
06/21/12 (54) 0 (54) (54) 24192
06/22/12 0 0 0 0 24192
06/24/12 0 0 0 0 24192
06/25/12 0 141 141 141 24333
06/26/12 0 (64) (64) (64) 24268
06/27/12 (41) 0 (41) (41) 24227
06/28/12 0 81 81 81 24308
06/29/12 108 (200) (92) (92) 24216

Read more: 301 Moved Permanently

results today:

g/j long @ 123.98, close 122.32, -146
a/u short @ 1.0177, close 1.0129, +48
e/u short @ 1.2211, close 1.2187, +24
u/j long @ 79.78, close 7931, -47
total: -121

unfortunately, both our yen trades pulled the trick of roaring up to trigger long positions, then heading the other way for losses. maybe tomorrow…

Read more: 301 Moved Permanently

In this system, when we open and close position in single day, it is very risky to just keep positions that are moving against you - and just let them to close on end of day. There is no time in just 1 day to recover from example -146 pips.
From my backtest, ST should be at 50-60 pips.
I know you are aware of that… just to emphasize.

Fixed TakeProfit calculation.

gk, point taken. thank you. i have been concerned about the spreadsheet limitations and keeping the stop wide enough that it wouldn’t attract too many double entries. i’m now thinking we would prefer the doubles over the increased risk. i will apply some varying stop losses in studies today and post when i have results.