Donchian Channel Trading

Yeah I have and basically my conclusion is they are not the best settings. Take these results 301 Moved Permanently and 301 Moved Permanently
same system as described, much more interesting results.

Now here’s where I play my get out of jail card. I happy to share the tools, the hard work is up to you. It’s your system at the end of the day. Testing for variable settings is one thing. Sitting there doing visual testing is another. Just because MT4 isn’t ones trade terminal, doesn’t mean you can’t use it for testing. Strategy tester is a very good tool.I have my testing MT4 platform and all my data on a separate partition on my hard drive. This prevents data loss or contamination from my broker.

I will have a look at your links, thanks…

I appreciate your backtesting and sharing your bots very much…in no moment i was intending to make you the official backtester of the thread…

I hope you will continue posting here (no backtests anymore :))…

I downloaded calgo. after a couple hours still confused so im afraid I cant contribute to the thread. Its all gibberish to me Im more confused than when I downloaded it.

Hi

the thread is meant to be a discussion about donchian channels…which settings are used, how people handle the stops, and so on…

so, you can contribute in any way, be it practical (backtests, …) or theoretical (new ideas, old ideas, …)…

Hi Bob

wanted to chat about the two rules below…

  • rule about not taking the next trade after a win trade, i think is important rule for good reason in the strategy. idk about piece of pie, and i sure don’t know about virtual trades, but i think there could be ways and means to implement this, and was hopeing to discuss the possibilities with you without clogging up the thread.

-my understanding about the limited exposure rules was related to when applying the strategy to many pairs / instruments, so is not applicable when just trading just one pair.

LOL I’m happy to be the back testing *****!!!

It’s good fun but I’m the one getting the reward out of it cause I have many pairs I can test and then get to choose my baskets.

So this is what I observed on back testing across several pairs on the one hour chart. Using a multiple of ATR as a trailing stop was indeed a viable exit strategy. A multiple of 1 was the most profitable but also occurred the greatest draw down. When it wins it wins big but losses are frequent. At multiple of 2, returns where greatly reduced by still signification. Draw down was reduced but probably of an excessive size especially if you only trade one or two pairs. At multiple of 3 returns again are reduced returns but draw down was the best

Of course there are other factors, I look heavily at expectancy and total trades. And of course know what time of day it is and where about in the monthly news cycle you are. With that said IMO I would stick to the multiple of 2 x ATR. And trade as many pairs as possible. 10 pairs at 0.2% risk is a far better option than 4 pairs at 0.5%.

One just has to look at the statistics. There are far to many net losses directly following a 20 day break to make it viable to re-entry the market. But then would have the EURUSD have done better or worse if we skipped that next trade.

-my understanding about the limited exposure rules was related to when applying the strategy to many pairs / instruments, so is not applicable when just trading just one pair.

Trading donchian channels is about diversification. And that reduces risk. Trade the strategy across as many pairs, indexes, metals and commodities as possible. This keeps risk to a minimum while maintaining strong growth. But with that all said trade your way not mine

I think the 2x ATR trailing stop looks quite good…might try it also on the 15min TF…
i trade the 28 pairs of all combinations of the 8 major currencies…that and trading 3 different TF at the same time is my diversification right now…

ok, just throwing down some thoughts,

skip a trade rule, i see or understand more in terms of trend following. when your winner closes the market may well be consolidating or taking profit, not necessarily a reversal of the bigger trend. no matter, instument is same but account maybe not, would be interesting to test the difference. ea could detect when first run, calc win or lose, then trade or not. could keep track internally of win and loss trades and could write this to file. or manually run the previous period and manually tell ea what last trade was. have you looked at the classic turtle indi code on the mql5 site? i bet it don’t follow this rule.

more than one pair… ea could do that or could have a trade manager ea. dunno if u could test a trade manager. hmm. a course i did said also keep longs and shorts equal number, idk if that is original turtle rule.

brain cell had thought about donchian channels on renko chart, probably parallel renko would be best. then n is x amount of change in price, and block size can be set to amount of atr even. can be backtested?

still looking for a way i can feel trust in. with a system like this, at least, it’s more about following all the rules right, and a bit less about emotion.

thx for code and for covering system that appears to have stood the test of time. cheers.

bob, one more question to your DC 20 - 2x ATR backtests…i guess the test took every valid signal that came across without any restriction of day of the week or time of the day…as you mentioned these factors in other posts, i just wanted to clarify that in your backtests that factors weren’t included…

G’day Mike. Yep these tests are unfiltered and take each and every trade. And yes after analyzing the results I added a time filter which produced even better results in terms of yield. However it didn’t affect draw down. Conclusion, this allows us to reduce our risk per trade just by trading at the right times.

Mondays I’m very wary of trading. Seems like the myth of Friday trading has been busted. The first couple of hours of Asian trading are profitable so too Europe. 10am to 10:30 GMT is a favorite time of mine to scalp the EURUSD as a reversal in direction occurs on a regular basis as the morning traders finish up their business and head for meetings. This is reflected in the channels with price action retracing deep back inside and taking out stops. Note this is only an observation on the Eurusd as it is my beloved pair.

Then from lunchtime London until midmorning New York is on boots n all, for hell and leather. Again but after that the business end of the day is done. The banks have balanced up their books and its only the leftovers to profit from. And there’s a whole lot of everyone trying to get their hands on the left overs. Oh and then there news days.

Now these obs are on the 1 hour chart for the EUR, GBP, CAD, CHF and USD. The AUD and NZD play out a little different so to the JPY.

Trading is easy, quick, simple and boring as. The real fun and hard work is behinds the scenes doing the research, collecting the data that quantifies our trade decisions. If ones happy to accept large periods of draw down then all one has to do is take each n every trade and that can be done via a bot. Although I highly recommend not using mine.

However a bit of hard work and one can filter out the bad trades, reduce risk and improve return. Next post some more results.

Just smash this out on the EURUSD. It’s getting late here so just the results and a bit of light discussion. All tests on the 1 hour chart at 1% risk unless stated otherwise.

First chart break set at 20, sl at 10 to trail immediately


Next a tweak to 24 for the break and 9 sl set to trail immediately


Next again a break of 24, but this time back to sl set at 2x ATR. Trailing stop set at the 9 channel and trails once the level has passed the stop loss.


Now this looks by far the most impressive until one realizes that the draw down on chart 3 is 26% compared to 19% on chart 2. Out of interest chart 1 has a 30% draw down. So for chart 4 I used the same settings as chart 3 but reduce risk to 0.7%.


This brings it back on par with the 2nd chart settings. So we see another variable being introduced. Money Management. So at this point I’m reminded of two things. First the turtles had 7 processes in their system. Entries and exits only account for two. Secondly there’s an old saying, cut your losses short, let your profits run. Worth a though

Old m8 pabloroacho, haven’t forgotten about you. Just going to sleep on your obs and think about them tomorrow. I don’t see any reason why renko charts wouldn’t work. But donchian trading is about breakouts. IMO there are only two things that matter, price and time. Unfortunately with renko charts time can be lost and price is misrepresented. I can see pips being lost due to late signals on renko - but I could be wrong. So I think about it a bit more.

g’morning bob,

yes, we’ll have to consider money management seriously as part of the system…
in my live account i am just risking 0,25% per trade right now…

the problem will be minimizing the drawdown while still maintaining the profits…one can’t have it all :wink:

for now, i would discard the 20-10 settings as it’s clearly the weakest system (smallest profit and big drawdown)…

Morning Mike.

Like what your doing. Over 28 pair maybe even 0.25% a little high. I’m set at 0.5 on the 10 pairs I trade on the 4 hr. Got a bit more testing to do on the 1 hr but trading the 24 hr break seems to have the greatest potential. 2 x ATR seems the perfect SL. TS does seem to fluctuate from 6 to 12 maybe even 15. Thats just going to come down to the trader and maybe discretion at this point is better. Hope to set up a demo by the end of the month running a bot and we can do some forward testing.

For now I’ll stick to this project before I shift focus to other time frames.

over 28 pairs it would be 7% total risk…maybe i recalculate to come closer to 5% total risk (if all 28 pairs are getting stopped out)…i am trading it on 15min, 1h and 4h TFs…this morning i started thinking about trying also the 5min…just putting entry orders right before the london and ny session…after ny session i would cancel the remaining entry orders and start again the next day…maybe i’ll give it a try…

i like your idea of the demo with the bot, let’s see what result the forward testing brings…

ummm, what is the volume spike in the middle of the backtest? what time & date was that trade? just curious, some news event?

A picture paints a thousand words


Thats just a gap in my data. 2 days out of 2000 trading days. Hard to think that would effect the results too significantly.

G’day lads. I have set up a demo account running two bots and linked into myfxbook for all to see.

First of all the all import link Turtle Traders System by forexbob70 | Myfxbook

Now about the bots. First bot trades the 1 hr chart with a time filter. Logic is if no trades exists, open a buystop and sellstop at the upper and lower donchain break channel. SL is set a a multiple of the ATR. No TP. Orders are then cancelled, recalculated and refilled every hour (during trading) until one order gets filled. OCO. Trade management is by a stepped TS. First the TS trails immediately at the ATR multiple from the breakout DC. Second it then locks in BE at open price until the trailing DC level passes this point. The stop then trails at this level until it is hit, the order closed and it starts all over again.

2nd bot trades the 4 hr chart. Works the same trade logic but uses fib levels to define inside channel levels instead of ATR. I’ve simplified it from the one mention in an earlier post about this strategy. It has a fixed TP but now has a active TS set at these inside channel levels.

Risk is % Account Balance. For ego but also because of back testing I’ve set for 1% to keep a constant. Personally I’m expecting draw down to be 30 to 35% at this level and could easily last for 4 to 6 mths. As more pairs get added this %risk is expected to be reduced.

Settings for DC levels, trading times, ATR etc are the results of extensive back testing over the past 10 years. They’ll be a reflection of my skill as a tester. Which probably is not very good :). The past is no guarantee of the future. Time will determine the results. Initially we’ll trade the EURUSD, GBPUSD and USDJPY. I’ll add more pairs over time.

So maybe in a years time will be discussing a 200% growth in the demo account. Maybe in 6 weeks we’ll be thankful that it wasn’t a real account. Who knows? Who cares! The journey will be fun.

As normal I will make these bot available but this time I’m not giving them away. You’ll have to PM me. And don’t be disappointed if you don’t get a response. If you have contributed to this thread or the forum in general I will be happy to share. If not better piece together a good argument.

Very Interesting! DC is as old as the hills. Turtles played with this so I guess it has some merit. Only in trending markets. That being said, good money management should save the drawdown a bit. Definitely not a bad long term idea.

Good luck BB