according to my numbers, june came in this way:
gbp/jpy -511
aud/usd +73
eur/usd +248
usd/jpy +56
Portfolio -134
one emergency stop was hit in june, gbp/jpy on 6-29
thanks for your efforts. i do understand both limited time and limited experience. maybe your good start will motivate someone else to help out. i attached the ea today and tried some backtesting. as i typically do, just took a few days out of 2007. the ea is opening the trades spot on, but they are not closing. i assume we are using daily data. the platform i am testing on is ibfx, so close is 23:03. i’ve tried plugging a variety of numbers into the timetostop. no luck. is it working for you? any help appreciated.
Hi Pipwoof! I really like this strategy!! Looks awesome! I started today to test it in a practice account, hopefully I’ll use it in a few weeks with real money 
Just a question here. Do you have a recommended entry hour? Because I’ll close all the trades at the end of the day, but I think that sometimes, at that time will be risky to enter the market, because the market could be near of the yesterday’s high or low. When this happens… Any recommendation? Or we have to figure it out?
I think that it isn’t a big deal If the market is near the yesterday’s low or high and the traders have enough power to break the resistance or support than your order will be triggered but during the time of consolidation the market will usually bounce before triggering your order.I think that a little bit precaution can be taken by setting order 5-7 pips away from the exterme of the yesterday’s candlestick!
Hi pipwoof!
I’m excited with your strategy and I just opened a demo to test it.
Two doubts. Which GMT time you consider the daily close? My broker uses GMT 0, but I know some which uses other GMT and it means H4 and D1 candles different.
Should we consider GMT 0 as the daily close and then look for entry of the next day?
AND…
In case we break lower, we short… If it goes up and break up, close short and buy or let short runing?
joan, i place orders at the beginning of the platform day. i am in central time and currently trade with oanda (edt) and ibfx (gmt), meaning if i am trading ibfx i will place entry stops at 7:00 pm and if oanda 11:00 pm. if your computer is close enough to your bedroom, you can also set price alerts to get you up as price approaches an entry point. you might also consider using fxalert to send a message to your cell phone.
i am currently working on trend studies and believe we will want to use a trend filter, trading with the prevailing movement. when we get that down, we will be entering only one side. that should simplify things a bit.
for the sake of those who might have to set it up and go to bed or work, for testing purposes, and for the space available on bp, i kept these rules pretty static. in real trading, if we are watching the screen, i don’t think we can or should avoid some discretion. an example would be when you see the close of the day coming at or near the high of the day. if you are already in a long position, why not just leave it open with a tight stop as we enter the new day? or, as jj suggests below, require a few additional pips movement before entering. i have seen times where the pair closes near the high/low and just doesn’t have the oomph to continue next day. sometimes it does.
another discretionary decision would come when we have seen a good run in our favor. then, the market enters that afternoon sleepy time where hourly ranges get small and we see a drift off our high point. close, bank the money, go to starbucks. hope it doesn’t hit another big run which would have been in your favor while you’re gone.
discretionary decisions are hard to teach, often wrong, and just come with experience. if you are the type to panic, move stops, curse the market, then you need to keep it mechanical. if you can keep your head, make some judgment calls. in any case, maintain a trading journal, so you can look back and see what you did and what on earth you were thinking at the time.
i am not trading this one live yet. i wanted to wait for the bp feedback and tweak a bit. we should be able to wrap that up in a few days. be sure you read my post on historical drawdowns. don’t know if you are able to watch the markets during the day, but will be introducing a variation on this method for next month’s contest which requires more attention.
thanks for your participation. best wishes for your trading.
baruch, i think what i just wrote to joan may answer your questions. i use hour bars. right click on your chart, go to properties. then, common tab. show period separators. this draws a line for you at 00, beginning of the platform day. i don’t mean to insult your intelligence if you already knew that, but i never know how “newbie” a “newbie” is.
hahahahah not a problem friend!
i’m not a newbie, a trade for some good time and I a scalper looking for something less stressfull.
Having a strategy where you place one order per day and give good results sounds nice.
And my question was before your answer to joan, and I read it just after my post.
Ok, I will try it and post my statement to all. I will modify the entry, instead 1 I will wait 5 or 10 pips (I’m thinking about it yet)
and about the question when both sides are broken? what you usually do? close the first trade and open second, or let the first runing?
Hi Crisscross1983,
I just ran a test on this EA for the period btw Jul-2000 to Jul-2012 and after 2396 trades, the account went from $10,000 to $0. I’m not a programmer and hence can’t read code. I’m not sure if there is anything from the EA that need fixing or I make some mistake running the test in MT4.
Regards,
i have a little more time this morning before we get out for the day, so will update you on looking into using a trend filter. it would seem, like that guy in college who wanted to hang with you, we just can’t get away from our friend the trend.
i approached this using data from 1-2007 thru mid-june 2012, two pairs, gbp/jpy and eur/usd. first, looked at our basic results without regard to trend. next, used a loose definition of trend. finally, tightened up trend definition.
as always, subject to error, check my work.
without regard to trend, the gbp/jpy gave me 642 losses, 597 wins, for a total of 1,239 trades. losses -45,115, wins +69,476, net +24,361. average loss -70, average win +116.
gbp/jpy, loose definition of trend gave me 372 losses, 444 wins, total number of trades 816. losses -27,021, wins +48,194, net +21,173. average loss -73, average win +108.
gbp/jpy, tighter definition of trend gave me 375 losses, 465 wins, total number of trades 840. losses -27,826, wins +51,324, net +23,498, average loss -74, average win +110.
eur/usd, no trend consideration gave 678 losses, 576 wins, total 1,254. losses -32,710, wins +40,105, net +7,395. average loss -48, average win +70.
eur/usd, loose definition of trend. 398 losses, 508 wins, total 906. losses -21,186, wins +31,649, net +10,463. average loss -53, average win +62.
eur/usd, tighten up the trend. losses 401, wins 513, total 914. losses -20,127, wins +32,701, net +12,574. average loss -50, average win +64.
we can talk more later about how we might want to define trend. for now, it does seem that the filter would be helpful. perhaps more so with some pairs than others. using the filter would resolve the issue of entering both sides and would reduce the number of trades taken, thereby reducing transaction costs.
enjoy the holiday!
apqn, the ea is not working properly at this time. i ran a test from 2007, 1 full contract, and tripled the account. but, it entered more than one trade per day. also, the stops don’t seem to be working correctly. i did find that you had to set tp high enough to allow the trade to close, but can’t do anything about the other problems at this time. bottom line, the ea needs work.
which filter you used in this case?
okay, wife not ready to get out for the day. have a little more time. baruch, i didn’t use a mechanical definition of trend for these studies, so they are subjective. i wanted a general screen to see if there was merit to the trend idea. it seems there was, so we will need to do some work if we want a more precise definition of trend. i pulled a weekly chart, put a 5/20 ema on it, looked for pivots, and made some judgements. example: gbp/jpy from 1-07 thru 07-07, UP, from 8-07 thru 3-08, DN, etc.
when we get around to this trend thing, we are going to have all the familiar problems. if we lag, we come in too late, if we anticipate, we come in too early. you know. i think it’s going to help us to use a fairly broad trend definition just to stay out of the way of the proverbial oncoming train.
maybe others have some trend indicators they want to suggest. i have this weird thing called zcomfxao direction. i don’t know how it works, don’t remember where i got it, and it won’t let me access the code to even try to understand it, but it puts an arrow on the screen, presumably a trend indicator. if someone can say something positive about it, i will include it.
i dont know how to attach a file!! :31:
but i guess i have a good indi for it…
im trying to attach an file… it says invalid file
baruch, i like your style! we are the blind leading the blind, but we’re gonna get it. i don’t know if this is the best way to do it, but i took the indi file and compressed it using the available windows software. right click on the indicator file, choose send to, which brings up the option to go to a “compressed zip folder.” that compresses the file and changes the extension to one the bp attachment manager will deal with.
nice
it worked
i tried .rar, but its accepts only .zip
Thanks for your answer pipwoof! I’ll be following the updates here :), and I’ll let you know if I find out interesting tweaks.
Regards
And there’s the rub… identifying when the trend changes, but not too early and not too late, but juuuust right. Guess we need Goldy Locks, lol.
A couple of the most common trend indicators is probably a good place to start, here are some suggestions:
RSI(14) - Long if RSI(14)>50, short if RSI(14)<50
EMA(n) - If previous day’s close>EMA(n) then only go long, if previous day’s close<EMA(n) then only go short. The best value for n would need to be derived through testing, I would think. Eyeballing the charts makes me think around 20 would be a good area.
Nick