Once again thanks for the interesting** trading strategy. I’m beginning to do my own number crunching in Excel on this program and came across what seems like an important concern…
Note: I understand in your original posts you disclaim that no spread fees, etc. were taken into account in the overall profitability, but I just want to make this point aware to other readers in the thread who are considering jumping right in with real money.
… If you do a ROUGH estimate on spread cost alone -I assumed an average of 3 pips per trade and 1000 trades per year across the 4 pairs- The total pip count decreases by approx 15,000 pips. In my personal analysis on FXCM (just using a year’s worth of back testing = I’m too lazy for more) my trades would have made 1,550 pips but then lost nearly 3,000 on spread costs. This testing was with just the original premise, before any stop losses or other modifications were added.
I am still new to Forex and I’ve been up all week with finals, so it’s possible I’m missing something very obvious. :34:
Are you or any other readers as concerned as I am with the actual effect of fees on the profitability of the system?
lucky, i appreciate your concerns and taking time to post. geez, i certainly hope no one is jumping in with real money. you can hardly read a paragraph on this site without being warned about researching and demoing before you go live. but, i guess it happens.
there is a notice on post 1 advising traders to be sure they see post 52. that post is about periods of time when this approach simply doesn’t work. markets change. when they do, they may cripple methods that have previously had great track records or bring back into play methods that had quit working. in my personal trading experience, i have never found anything that consistently worked all the time. i find that i have to be nimble and willing to change. some of my best help in making those changes is the equity curve. it is just a way of looking at whether a system has quit working and for how long. i have a trader friend who says he doesn’t need an equity curve, he just watches his account balance.
daybreak showed positive results for the most recent six and a half year test period. prior to that, there were periods where you would want to trade it with terrorist money, i.e., it lost. i don’t know for sure what pair(s) you were testing or over what time period. i can only conclude that if the trading results did not pay expenses, you had to be testing in a losing period. i did note what seems to be a problem with the math for your rough estimate. three pips costs per transaction seems fair enough. i pay three for g/j and between one and two for the others. but, there is slippage, mistakes, and surprises to cover. so, that part seems okay.
i do think i see an error in that you are supposing four pairs would trade an average of 250 times per year each for a total of 1,000 trades. if that were true and the average cost per trade was three pips, you would have 3,000 pips in transaction costs per year, not 15,000. maybe you meant for the entire test period and that would be about right.
actually, i am showing about +48,000 gross pips for the total four-pair portfolio over the test period. you are correct that this does not include transaction costs. to get these results over a 78 month period, each pair traded about 1,250 times, except for the usd/jpy which traded about 800 times. let’s be generous to the transaction costs and say 5,000 trades X 3 pips per trade = -15,000. not as pretty as the original +48,000, but will leave us with net +33,000. the lowest performer was eur/usd with around +8,500 gross and found it necessary to trade about 1,250 times to get that. by our estimates, transaction costs would have been roughly -3,750, leaving +4,750. this nets about +60 pips a month and won’t do much to help pay the bills unless you’re trading $100 a pip. if you were able to trade all four pairs, you might have made just over +400 a month.
i have and will continue to warn readers that this system has limitations both in the research methodology and the performance aspects. it was tested with a spreadsheet and daily bars. the limitations of that approach have been discussed from the beginning and throughout the thread. it is a “set it and forget it” trading method and that implies certain performance limitations. for example, suppose you set it at 7:00 pm, check it before you go to bed, check it again in the morning, and go to work or school. when you get home that evening, you learn that there was a major rate announcement sure to bear on the currency markets. you find that prices blew right through your stops and you suffered a serious loss. might you have done more about that had you been able to keep up with the news, watch your trades, and close out as that event approached? of course. but, that’s what we call a full time trader, not someone who has responsibilities at work, school, etc.
now, i hope i didn’t miss something very obvious! heh. best wishes for your trading and good luck with your finals.
thought of one more thing and will add it as an edit here. you mentioned a profit of some 1,500 pips with tranactions costs of 3,000. per our estimates, that implies 1,000 trades and an average profit per trade of 1 1/2 pips. bottom line, whatever approach is doing that is just not adequate.
First of all, I really do appreciate all of the thought put towards responding to my post!
For my data, I looked at the year 8/1/11 to 8/1/12 on the four pairs you’ve mentioned (EUR/USD, AUD/USD, GBP/JPY, USD/JPY) I had no stops, and entered and exited trades at 17:00 EST each day. (That returned the roughly 1,500 pips before transaction costs), which resulted in a -1354.4 net loss after spreads.
After posting, however, I did experiment with emergency stops on each pair not too far from the ones you suggested:
EUR - 150
AUD - 115
GBP - 150
USD - 75
I came to these numbers by averaging the ATRs throughout the year prior to my backtest ie 8/1/10-8/1/11
With this, I was able to profit 437.2 pips after spread costs, which as you mentioned, isn’t all that great.
I also found that excluding Monday trading due to the uncertainty and relatively low volume compared to the other days of the week actually hurt results, making only around 250 pips net, even with the smaller amount of executions AKA less spread costs. My next point to explore is whether trading from say 0:00 - 17:00 EST is more profitable than having a trade running 24 hours a day. I’m stuck with how to quantify this in Excel though…
If you or anyone in the thread knows how to go about this, I’d be very grateful! So far I’ve pulled hourly data for 8/1/11-8/1/12 and filtered out hours 18:00 - 23:00, and I’m at a standstill trying to condense the 17 hourly bars into one daily data point…
tyler, my data for 8-1-10 to 8-1-11 shows a portfolio equity growth from 39,973 to 46,176 = +6,203. assuming about 1,200 trades X 3 transaction costs = -3,500. net +2,703, +225 per month.
8-1-11 to 8-1-12 did not do as well, growing from 46,176 to 49,267 = +3,091. assuming similar transaction costs, we would have a loss about -400.
this highlights the importance of portfolio management. it is important to watch the equity curves for the portfolio overall and for each pair individually. as markets change, we will always need to be making adjustments. i haven’t studied this situation in depth, but i can see that by the beginning of this test period, u/j had become a weak link and wasn’t pulling weight. there may have been others, but this one stands out. had we dropped that pair from the portfolio and put in eur/jpy, our growth would have been to +5,547, less transaction costs of -3,500 = +2,047. not outstanding, but still about +170 a month.
how would we have known to make those substitutions? by looking at the equity curve when the method is applied to a particular pair. when we see that curve start to fail, we get on notice and start looking around for a performer. that’s why we also have to look at pairs we do not have in the portfolio to determine if we want to add them. if there isn’t a good one, we may just have to cut our portfolio size. look at it like players in football. a tight end was doing well, but now seems tired or maybe injured. he may need a rest. this guy looked good in practice, let’s put him in and pull the other guy.
while we’re doing that for pairs within a system, we also need to be doing the same thing for our systems as a whole. diversifying means trading different pairs within a system and different systems. has this system failed? it did fail for some periods prior to 2007. it may be time to bench it.
i have frequently referred to equity curves and may not have gone into much detail about how to use them. here are equity curves from five different pairs showing results from the daybreak method from 1-11 thru 8-10-12. what do you see? if you were trading daybreak, what would you do now?
i see that g/j has been a steady contributor, adding some +2,500 to the portfolio, with about a -500 pip drawdown in the middle of my test period. i would keep it online and reconsider only if it went into a drawdown of another -500 from where it is now.
a/u has been hotter than a pistol and there have been fundamentals affecting that rise. though at one point it had put +4,000 in the account, i see a drawdown more than -1,000 and would be watching it closely. if this equity curve drops below 3,000, i will be taking a/u offline and continue to monitor it on paper until we see if it starts to perform again.
u/j put +800 pips in the bank, then gave -400 of it back. then, got to +1,000 and gave -400 back. this isn’t enough reward to continue with the pair at this time. i would drop it.
e/u may look kinda wild, but we can handle that as long as those eod closes are profitable for our entry points. they have been, with the pair putting +1,500 in the bank and apparently trying to get +2,000. the important thing is that the curve works its way upward. however, we also have to note that this pair traded about 350 times at 1 to 2 pips spread. say 2 X 350 = -700 pips in costs. it may be on thin ice, but here’s another way to look at it. if you have the account to trade those that look more profitable and another one or two, why not take the few extra minutes and pick up the +1,000 or so?
e/j is one i recently added to the triple threat portfolio and i definitely think we should add it here. with the exception of that bump in the road toward the first of the year, she has been a real sweetie. that bump, btw, might have prompted us to take this pair offline with the move down through +1,250 equity and we would probably have brought it on again as it crossed back above that number.
of course, do not bring anything online (live) until you have paper-traded, demoed, and come to thoroughly understand the trading method. best wishes for your success!
Thanks yet again for a wonderful idea and even better supporting discussions. This is a little off-topic, but I have been developing your Daybreak premise further for a presentation I will be giving at school in the coming week. First of all, do you mind if I mention you/Babypips for citation purposes of where my idea originated? I would PM you, but my post count has to be at least 20 to do so. I was wondering if you wouldn’t mind messaging me your name in order to give due credit for the preliminary idea.
i appreciate that you would even be that considerate. i would like to see babypips get anything in the way of credit. they work at making a suitable site for exchange of ideas and do their best to protect us from the vultures. you can refer to me as, “pipwoof, a full-time trader from the chicago area.” once you post on a public forum, the idea is kinda owned by everyone anyway.
fyi, i am sifting through a number of ideas suggested by visitors to the thread. unfortunately, the ea’s that were submitted here do not function properly, so i am doing a lot of tedious handwork. but, it could be that one of these thoughts could make a better method than i ever thought of. if i happen to find something, stay tuned for an exciting announcement.
I appreciate your inputs. Currently I am reading this thread from the first page. I have same strategy in my mind. But I didn’t read the all pages. After reading will comment here. Hope You can help me.
I just don’t understand all the concern over spread costs. Once a trade reaches “0.00” any spread cost has been eliminated.
If you are positive 1 pip you are in profit – period. If I am trading $10/pip and it takes 3 pips to get to 0.00 – thats OK as long as I get there the subject of cost is mute. Real Cost is only a factor if I have a losing trade. 1 pip I just made $10. Right?
I didn’t make $10 minus $30 for a net of minus $20. Right? I just don’t know what I am missing. It seems like such a miniscule subject to even discuss. Sorry - don’t mean to diss anyone - just don’t understand the reasoning.
How long have you been trading…? Many traders only target 5-15 pips profit per trade, scalping, so a 3 pip spread is anywhere from 20 to 60% of those types of trades. How is this not significant?
Spread becomes less significant as the size of the tp increases but still… It is always going to play a significant role in your account growth.
First I have been trading 4+ years.
Secondly - It is relevant to the distance you have to go to get to 0 pips - but not to profit. If you target 10 pips and make the 10 pips (even tho it took you 13 pips to get there) you made 10 pips profit. Period.
BUT – If it bothers you - find a broker that has less of a spread - But you will always have some spread. It is just part of trading, always has been, always will be.
Remember - This is just my opinion – but then I am usually less concerned about things I cannot change than others may be.
If your target price movement is ten pips… Then you only get 7 pips in profit… That is significant… If you want 10 pips in profit then you have to target 13 pips in price movement… The extra 3 pips in price movement over a large frequency of trades will have a dramatic reduction in your win percentage… So yes i agree with you, spread is significant.
