Illumination

FetteredChinos

Veteren member
Messages
3,897
Likes
40
Ah the chinos fun factory has been at it again.

been doing some research back on the Eur/USD. and guess what.

here's another strategy.

the figures are approximate, in that i am not entirely how to code a stop onto a Stop-and-reverse signal, so i have had to assume prices move in a straight line. but without a stop, the results are , quite frankly. astounding.

please note, this is not an intraday system, and the method will be disclosed in due course...



i would appreciate it if anyone could double check this to make sure im not going mad, as this is precisely the sort of strategy i am looking for. something to be able to trade once a week from the pub.


results from Feb 2001 to present are...

Total Points 24613 (this is not a misprint)
Trades 188
Wins 113
Win % 60.1%
Max Win 964.00
Max Loss -800.00
Drawdown 1977.00 ( i know. this is a bit of a problem)
Ave Win 371.73
Ave Loss -231.91
Prof Factor 2.42
Prof per Trade 130.92


it takes 1 trade per week as the main signal. occasionally takes a second.

excel sheet to follow in due course...

FC
 
FetteredChinos said:
the figures are approximate, in that i am not entirely how to code a stop onto a Stop-and-reverse signal, so i have had to assume prices move in a straight line. but without a stop, the results are , quite frankly. astounding.
What exactly does that paragraph mean?
 
FetteredChinos said:
Drawdown 1977.00 ( i know. this is a bit of a problem)
Not with overall results like that, over nearly a 4-year period, it isn't! Tell us more ... :)
 
DD.

with regard to that paragraph..

the best way to understand it is to view the spreadsheet i am about to attach.

column X is the stop-and-reverse part of the method.

i have used a bit of artistic licence to say that if the loss is more than 400 points, then to assume that the loss was just 400 points. this is not strictly accurate, as the trade may have gone more than 400 points into the red, but then closed back above this line. without wanting to spend ages coding this up, i have just assumed that prices move in a straight line through this level and kept going.

besides, as far as i can make out, only 3 SAR trades closed at more than a 500 pt loss, so i think we are ok in this respect.


sheet to follow.

FC
 
Right. here we go...

this strat may not suit everyone. but as Chow would say, it is a "nice little retirement fund builder".

rules are.....


Based on Weekly Data.

If weekly close is greater than the balance point, score week as +1
If weekly close is less than the balance point, score week as -1
Sum score of last 5 weeks. If positive, then go long. If negative, then go short.
close position 7 weeks later
set stop at -400 pips. If that is hit, then SAR and trade in the opposite direction.
assume stop of 400 pips for this reverse. Close all SAR trades on the same 7 week close.


if you arent aware what a balance point is, it is simply an average of the high and low for that period.




the fact it is weekly, and has such a long hold period may not suit everyone. but once you are up and running, a trade is opened and closed every week.

FC
 

Attachments

  • Illumination.xls
    205 KB · Views: 535
Hi FC,

Sounds like a very interesting concept, especially if it back tests well over the last 4 yrs consistently, as I'm sure you have, like me, strategies that work great for one or 2 years but not during the longer term.

I have a few q's if I may.

I am looking to begin trading this from next week, (hypothetically), so I check the last 4 weeks and this week and see that the total is +1 so I look to go long. But where di I go long from? Close on Fri, open on Mon, 7 a.m. on Mon. Apologies if I'm thick but I don't think you specified this.

After getting long, the current week ends down 200 pts so not invoking the stop level, but the running 5 week total is now 0. Is this a hold situation, or just a stop with no reverse?

Assuming I stay long, the next week closes down 100 points from my open, so the stop is still not triggered, but the running total is now -1. Are you saying the SAR should kick in here, or only if the stop of 400 pts is hit.

Sorry for the basic quest, but this seems a great addition to EOW traders.

Thanks for sharing.

G-Man
 
Sorry FC,

Forgot one.

Assuming the 7 week timescale for each individual trade, does that mean that at any one time you could potentially have 7 open positions? If so, does the total pip count classify each trade individually?

G-Man
 
right here goes. i'll try to answer them

trades are taken from the weeks close. ie friday.

some of my data seems a bit wonky and includes some saturdays, but they shouldnt affect things too much. but you may be better off using the friday close for simplicity.


next, assuming we are long as of last friday at 1.2949. since we are using a 5-week sum. there is no way we can get a total of zero. therefore each week will trigger a trade. and we treat each week as a trade in its own right.

so when we put in our long position from last friday. we entered a stop and reverse order to sell at 400 points below. ie 1.2549.

this order stays like and due to close on the 24th december i believe.

so when we get to 24th december, we have to look back at our trade for last week and see what happened. if it got stopped out and reversed, we close it. if it didnt get stopped out, and hit a healthy profit. we close it.

in between now and then, there are 6 more weeks of trades. so this friday, we work out what the trade is, and set the orders accordingly, irrespective of what happened last week. and this weeks trade is due to close out on the 31st december. and so on.

it will require a bit of monitoring on a piece of paper to work out what trades are to close when etc, but each trade is treated individually.

and yes at any one time, you could have 7 positions open, i believe, but the pip count treats each trade individually. so the points in this sheet, are the points you would have got.

obviously there could be some weeks where we are loaded to the long side, and it takes a dive south, but that is the same with any strategy. sometimes it works in our favour, sometimes not. as a result, we just have to assume that those instances balance out over time.

hope this helps.

FC
 
Absolutely fascinating, FC ... at the risk of sounding even more stupid than usual, it's something like a more sophisticated 5-period RSI on a weekly chart, isn't it.?You're using whether the close was above or below the balance, instead of using "up" or "down" for the week ... but there's a similarity in the concept, I think? Or have I misunderstood how RSI's work?

At tomorrow's close, then, regardless of how this week closes, you'd be opening a long trade, presumably, since the last 4 weeks all scored a +1 on this system?
 
yup, thats the fella.

i quite like it in that we are trading with the trend. hence the possibility of 1,000 point wins..

the wide stop allows us not to get triggered by the fun and games on big US data days, as after the huge spike, the price then reverts slowly back to where it was.

it only changes direction when the trend seems to change. the SAR gets us into the new trade before it is too late.

it works to some extent on Cable, but not quite as profitably. about 75pts per trade if i recall. but with greater drawdown. (yes even larger than 2000 points)


havent tried other currencies, but if it works on Eur/USD , then there is a fair chance it will work on swissy.

the only problem i see with it, is that even if we were to start trading it today, then we wouldnt see results until next year! but then after the initial 7 weeks, then we start closing positions on a weekly basis. so it should be alright. im gonna paper trade this for a while to see if it works going fwd.

im not sure that 200 trades is enough of a sample to be indicative, but only time will tell.

i will look into whether using 5 period RSI adds anything worthwhile, but part of the point in me developing this was so that i could setup a regular trading method with minimal effort.

after all, it takes no time at all to see whether we close above or below a balance point etc. whole process could take maybe 10 mins a week.

in my book, that is time well spent.


gives me time to attempt to start a professional golf career...

hmmm

:)
 
Ah yes ... interesting. I was wondering about other currencies. You must be right about the Swissy.

4 years' results sounds very impressive. But as you'd say, it's "only" 200 trades ... point taken.

I was thinking of asking about whether you've looked at this on a daily as well as a weekly basis, but as you've made clear in your post above, it would raise all sorts of new issues like "news", and might turn out to be far more complicated than one would think at first glance ...

Golf? Rather you than I, in the winter. Think I'll stick to poker, backgammon and online "sports".

Thanks for such an interesting thread ...
 
Nice work.

My only thought is that 60% seems a bit too high.

You really want to see how robust this system is.

Try it with 5 weeks and 10 weeks instead of 7 weeks.

How profitable is it then, and is the win rate still around 60%

Also try it with GBP/USD with say a 500 or 600 point stop. A good robust system should work in most markets.

Assuming your figures are correct then with 1% risked per trade the system makes about 16% a year
(before slippage/spreads and commisions etc) with a max drawdown of 5%.
Thats the best case, in reality it might be more like 12% profit/ 8% drawdown.
 
interesting.

so the occurence of being both long and short at the same time can occur?

I was wondering if an improvement might be for example if you were long and got a short signal to just close one of your open long positions rather than open a new short one, because having positions open in both directions is just an expensive way of being flat since you pay the spread for both positions but each of which offsets the other.
 
ok, just tried the basic method on the daily data.

has a 15 day hold period.

result, without SAR is 12,188 points from 700 trades.

17pts per day.

obviously this can be improved. but goes to show that the method works on differing timescales...


DD i tried differing hold times, and basically, anything from 1 week upwards is profitable overall.

as such, im confident that it works. i just found that 7 weeks produced the best profit to drawdown ratio.

obviously, small data set caveats apply.

but this is why i offered the spreadsheet, as many cooks make light work in trying to find a more robust method.

the daily method may be worth further investigation.

next stop, testing on hourly bars.

could be hard work.

FC
 
also, with regard to the % return per annum..

using leverage. and sensible posn sizing:-

maximum theoretical drawdown:- 1977 points.

treble that gives approx 6000 points..

so we can trade at £1 per point per £6000 account balance.

24000 points over 4 years is 6000 points per annum average.

so i make that 100% per annum return, expected drawdown 33%

better than a savings account at any rate.

FC
 
FetteredChinos said:
also, with regard to the % return per annum..

using leverage. and sensible posn sizing:-

...

so i make that 100% per annum return, expected drawdown 33%

better than a savings account at any rate.

FC
To get 100% you need to risk 7% per trade. Thats not sensible. 1% or max 2% is sensible.
 
FetteredChinos said:
ok, just tried the basic method on the daily data. has a 15 day hold period. result, without SAR is 12,188 points from 700 trades. 17pts per day. obviously this can be improved. but goes to show that the method works on differing timescales...
Yes, it does indeed. And that speaks highly of it. That's very interesting. I know how irritating it can be when you produce something really original and so obviously with great potential, and people immediately start asking a lot of "subsidiary" questions about something different ... but (there had to be a "but", didn't there?) I was wondering why you took a 15-day hold period for looking at the daily. I'm asking because in theory this would mean holding 15 different positions simultaneously (not necessarily in practice, I realise ...). I was wondering if something like a 5-day or 7-day hold period might work out? :)
 
Top