Donchian / Turtle breakouts query

J

jonboy123

Well-known member
Messages
400
Likes
3
hi, reading the turtles book.

for example: 20 days breakout (4 weeks)

can we look at weekly candle increment chart? or are daily bars the way to go?
i guess daily right? as during the 4th week candle...you cant look into a full 4 weeks, i hope you understand what im saying.....we should look at daily candles right?
(sorry im also thinking aloud here)

and if you are looking at daily candles, say you are watching todays candle, are you comparing the previous 20 candles? or previous 19 candles?
if you are comparing the previous 20 candles, so then thats 21 candles in total you are concerned with....?

And the Exit strategy: 10 day low for long position.
does this mean 10 consecutive days where the price closed lower than the day before?
or does it mean you just look at the 10 previous days and find the lowest price of the 10 days, and if the price falls below that lowest low, then you exit....?

10 day low: so you look at 10 candles in total including todays, or 11?
11 meaning todays, and the previous 10 candles.

Thanks to any help given.
 
jonboy123 said:
hi, reading the turtles book.

for example: 20 days breakout (4 weeks)

can we look at weekly candle increment chart? or are daily bars the way to go?
i guess daily right? as during the 4th week candle...you cant look into a full 4 weeks, i hope you understand what im saying.....we should look at daily candles right?
(sorry im also thinking aloud here)

and if you are looking at daily candles, say you are watching todays candle, are you comparing the previous 20 candles? or previous 19 candles?
if you are comparing the previous 20 candles, so then thats 21 candles in total you are concerned with....?
Click to expand...

I think they would enter the trade on the 21st candle, long or short, as long as there's a breakout from the high or the low on the 20st candle.

And the Exit strategy: 10 day low for long position.
does this mean 10 consecutive days where the price closed lower than the day before?
or does it mean you just look at the 10 previous days and find the lowest price of the 10 days, and if the price falls below that lowest low, then you exit....?
Click to expand...

You look at the 10 previous candles and calculte the lowest low. As soon as that one is penetrated you exit (on that candle, or on the candle of the next day if you want). You can consider it a moving low, just like a moving average.
 
straddle's right.
turtles looked at high/low of last 20 trading days (or 50d) but there we're no charts like today, so today's bar/candle/animated representation of daily price action :p was not available to them online.

however - and this is important.
the system is robust enough that no matter if you look at 20d high/low or 19d high/low or 21d high/low, you'd still get similar results
 
thankyou
so if im looking at 20 day breakout, im looking at a time frame of 21 candles, right?

if im going to back test this, is there trading testing software to do this?
or i have to do this manually? seems fidgity if im going to do it manually, eg keep moving a 20 day section rather than keep counting 20 bar candle by candle.

how do people back test? ive read somewhre there is software to do this for you.

(there were charts in them days, in their newspapers, they were spotting highs lows visually on their charts)

foreward testing is easier to monitor as you can draw a horizontal level for low and high and just wait to see if current price breaks through it...

can someone help me understand how to backtest this ?

cheers

found a website with expert advisor indicator for donchian breakout, but when i downloaded the pack it was not in there.

sorry, so 10 day low, means you are in effect looking at 11 candles? todays plus the previoius 10 days?
 
try looking for a "Donchian Channel" indicator, it should plot them directly on the graph.

FTSE example attached.
 

Attachments

  • goldbubble1.png
    goldbubble1.png
    16.6 KB · Views: 989
it stopped working lol. If you like trend trading just make your own system :)
 
amnonco said:
straddle's right.
turtles looked at high/low of last 20 trading days (or 50d) but there we're no charts like today, so today's bar/candle/animated representation of daily price action :p was not available to them online.

however - and this is important.
the system is robust enough that no matter if you look at 20d high/low or 19d high/low or 21d high/low, you'd still get similar results
Click to expand...

I think the Turtle System is an excellent example of how a system should be built. To be more precise, the turtle system also had a filter which stated you cannot enter a trade if the previous one was a 2N loser/winner (or something in that way). This filter seems to suggest there was some form of serial correlation between their winners/losers. None of the Turtle books goes deeper into this filter ..
They also were restricted to loosely correlated futures.

Also, I don't think they made millions because of this (very) simple strategy, but mostly from their (agressive) position sizing algorithms.
 
The Turtle System doesn't work now. It was an example of a good system in that it thought about entry, exit, position size, risk etc and was the first proper trading system a lot of people had encountered (compare with most of the rubbish you read in trading books). However it had one flaw, it was curve fit, and after a time stopped working completely. Part of that was trader effects, as it got known. However another part I think was the curve fit complexity - the don't enter if previous trade a loser is one example of a filter rule based on specific market behaviour which reduced the robustness of the system.
 
id rather play with it myslef and judge for myself if it works or not, rather than take someones word for it.

yeah it shows you what you should do and the kind of decisions/rules you need to consider for a system/strategy. I dont think it was overly curve fit though, as he does speak about keeping it simple to make it more universal rather than too specific and therefore not robust. So much detail in the book thought some of it went over my head.
 
Hi Bevok,
Where did you get this chart from?
And why have you used a log scale?
 
amnonco said:
second that.
Click to expand...

Hi there

Its from here: http://www.tradingblox.com/forum/vi...e+system&sid=f472a1b33ff2c59da850128eaece8e10

I wasn't really intending to be controversial, I (wrongly) assumed it was fairly common knowledge that the original system doesn't work now. When I say that, I mean with all the original rules (last trade winner rule, failsafe entry etc) on a typical diverse futures portfolio. If you run it on Trading Blox (or obviously any other platform with the rules implemented and portfolio level results) you might find that the results since about 1993have been poor.

Obviously this and any other idea needs to be tested by the trader, both to see if I'm telling the truth, and to ascertain what might work and not work about the system's components. My opinion is that it was a bit too curve fit to the way markets performed in the 70s with some of the specific rules, and also that 'common knowledge' (trader effects) diluted its edge. Thats not to say that it isn't worth studying. It is a superb example of a fully encompassing trading system - in fact for a lot of traders it was the first time they'd seen something like this, rather than the simple 'entry setup' ideas that obsess most trading books. Additionally some components of the system work extremely well when implemented a bit differently - the Donchian channel can be very powerful and effective.
 
jonboy123 said:
id rather play with it myslef and judge for myself if it works or not, rather than take someones word for it.
Click to expand...

I couldn't agree more !!!! Too much is posted on the internet and taken at face-value.....

Good Luck and you may be surprised especially if you "think outside the box" a bit with it....

Chorlton
 
You must log in or register to reply here.

Similar threads

Pipsaholic
Moving Pivot Point Indicator
Replies
2
Views
2K
TEAMTRADER
T
N
  • Poll
My Eur/Jpy Trade journal
2 3
Replies
57
Views
8K
ndnole
N
P
VTL Client Script: CarlPlayer/KendoFX
Replies
0
Views
2K
ProgtorHAL
P
J
4H Box Breakout
Replies
8
Views
12K
nubchai
N
A
Please review my STOP AND REVERSE Method ( UK)
Replies
10
Views
2K
D70
D
Top