The DIBS Method... No Free Lunch continues | Page 447 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

The DIBS Method - No Free Lunch Returns 13 replies

No Free Lunch but all the Free Coffee you can drink 736 replies

Peter Crown's DIBS method revisited 14 replies

DIBS Method EA 11 replies

Problem using Dibs EA 4 replies

Attachments: The DIBS Method... No Free Lunch continues
Exit Attachments

The DIBS Method... No Free Lunch continues

  • #8,921
  • Nov 22, 2009 6:51am Nov 22, 2009 6:51am
The right side of the moving average.... when price is above moving average go long..........below, go short

as for the moving average..........build your own model and find out



Quoting Sauron
What moving average did you use? And what is the "right side" of a moving average?
  • #8,922
  • Nov 22, 2009 8:28am Nov 22, 2009 8:28am
Quoting razorboy
The reason I picked daily data is that the truth of the matter is that I can't be at a computer 24 hours a day and I found that when doing some primary testing on 1 hour charts, that is really what you need to extract the most out of this approach - not saying that you can't do it well on the one hour, just that I cant - I found that I missed additional entry points and exit points.
Just one additional note to readers of the above post. Watch out for the "composite time" issue (ie. when your "day" starts and ends).

I wish all persistence in testing. It pays off

Bluefish
  • #8,923
  • Nov 22, 2009 9:05am Nov 22, 2009 9:05am
Quoting razorboy
I just finished a backtest (bar by bar) of 10 years of daily data on the GU - I used hourly data to drill down to determine the movement of price when I needed to determine if an entry point got hit before or after a SL got hit.

Summary
originally about 2900 days of data - about 300 to 325 positions taken over the course
No entries using IB's that were formed on Sundays
1/2 off @1:1
Two trails were tested - one relatively short, the other 2.5 times as long.
Both produced postive results. one better than the other. Both seemed to exceed the performance...
Nice work on the testing. The long tail of 8500 pips, to put some perspective on it would be the same as taking G/J now and holding it past 220.00. Considering that G/J was well above this point at some time gives confidence that if you can long a position and hold it, along with a few others at this point, ride out the losses and BE trades then you may just be able to in a few years time look back and say wow I did it. The wait is a very tough process and as Joel mentioned earlier that he held a position valued initially at $550 and held out for a 250K payday on that 1 position. $550 can be a bit to wager for some without over gearing in a major way, however it gives a few clear signs.

I am holding AUD from 8821, Silver from around 16.30 and these 2 trades have done me wonders so far. I have had to watch them run up and down the block a few times, AUD was over 600 pips at one stage and now is lower with the small pullback, but that is what you must weather to get the money. I had held position along the way from 6500 but sold them out for short profits as I was, and am still, building my tolerance to "long tails".

While my current positions are not IBs, essentially there is something that everyone can take from this thread and that is the beauty in long term maturity and the difference it can make to your account and to you on a more personal level. You will learn more about yourself and your limitations through this than most other methods, unless your an avid gambler and the constant in and outs of a 1 min scalp give you the rush you need to get by.

I like that you have taken the time to really backtest and share your thoughts on the results.

Take Care, hope you all enjoy the rest of your weekend.

Razor
  • #8,924
  • Nov 22, 2009 10:59am Nov 22, 2009 10:59am
I am in the midst of putting together a small portfolio for my son - he is six, so I want to include a few pairs that should have little correlation between them. I suspect that if I put together a portfolio of G/U E/U and US/CHF, I would probably see a fair amount of correlation between their performance, so playing them all at the same time would amount to a triple over-leveraging, although if you are playing for the long term, with good risk control, it may not actually matter as long as you generate positive results.

IB's are not the be all and end all, just an easy mechanism to generate entry and kill points for simpletons, like me.

Will spend the next week building a model for two more pair and then start putting it into action


Quoting Razor_trader
Nice work on the testing. The long tail of 8500 pips, to put some perspective on it would be the same as taking G/J now and holding it past 220.00. Considering that G/J was well above this point at some time gives confidence that if you can long a position and hold it, along with a few others at this point, ride out the losses and BE trades then you may just be able to in a few years time look back and say wow I did it. The wait is a very tough process and as Joel mentioned earlier that he held a position valued initially at $550 and held out for...
  • #8,925
  • Nov 22, 2009 1:46pm Nov 22, 2009 1:46pm
Razorboy said

Quote
..After further review of the data, and not closing a 1/2 position at 1:1, the end result was an increase in overall pips by about 62% over the previous best results - not for the weak of heart though. You essentially have a break the bank win every 2 years
The gift giving by PC was the 1:1 matrix for half off. Makes the roller coaster easier to handle. The rest was simple trend following methodology. Ride the winners and cut the losers short. Well done.

Quote
There's also the general point that Peter made -- he knew the difference between
an IB that was likely to fizzle out and an IB that would lead to a substantial
breakout.
Maybe the answer was simply trading during the most active time period during the 24 hour period with the daily trend. Hard to believe by most but thats all it was. If you doubt that go do a proper back test. Than tell me I'm wrong. All other opinions are useless. By the waY I was wondering if anybody that posts here regularly and actively trades was going to post this year performance gain or lost for the year? After all said and done this is the only thing that really matters. What you made or lost for the year.
  • #8,926
  • Nov 23, 2009 5:05am Nov 23, 2009 5:05am
Quoting razorboy
The right side of the moving average.... when price is above moving average go long..........below, go short

as for the moving average..........build your own model and find out
I'm interested as to why you'd use this as a filter. Say you are using a 36 or 45 EMA, or even one as low as 20, and you get an entry the right side of the line and for whatever reason (news most likely) you get a big move in your favour that takes you through the MA line. Say this is the start of a nice big trend, such as GBPUSD in the summer of 2008, and you have loads of entries with the trend as each entry goes free and you can add to your position. All these additional entries are on the wrong side of the MA according to this rule.

I bring this up as it seems directly contrary to PCs method of adding to winning positions; for me the summer 08 on this pair nearly doubled my equity with all the entries with the trend (over 30:1 on the combined positions before my sell EMA was triggered). My other big equity-increasing trades also exhibited this type of adding to winning positions.
  • #8,927
  • Nov 23, 2009 11:11am Nov 23, 2009 11:11am
Good work Razorboy! I am trying to figure out how to backtest beyond 3-4 years on MT4 myself buy can't seem to get back more than 3-4 years of hourly data....any idea how to do this? I have acccounts with both FxPro and FXDD and each one goes back different periods but neither seems to go beyond 2006!?
Thanks,
J













Quoting razorboy
Inside bars when the moon is in the seventh house and Jupiter aligns with Mars you will never lose.


It is really amazing how much people are focusing on co-incidental and possibly meaningless variable such as an upclose or downclose or upside down zig zig triple Fibonacci close of an inside bar..

you take every inside bar on the right side of the moving average you are using for a trail and you trail it.- that is it

I just finished a backtest (bar by bar) of 10 years of daily data on the GU - I used hourly data to drill down to determine...
  • #8,928
  • Nov 23, 2009 5:07pm Nov 23, 2009 5:07pm
Quoting JamesBranson
Good work Razorboy! I am trying to figure out how to backtest beyond 3-4 years on MT4 myself buy can't seem to get back more than 3-4 years of hourly data....any idea how to do this? I have acccounts with both FxPro and FXDD and each one goes back different periods but neither seems to go beyond 2006!?
Thanks,
J
Have you updated your historical data.

If not heres how: go to tools and click history center, this will bring up the side list of pairs and timeframes. Double Click on which ever you want. For example GBP/JPY 1 hourly has information go back to late 1999. Click download and it will download that set.

I am pretty sure this should help, the only thing im not sure about is whether it will give you this info on your chart or if it only is there for when your testing an EA strategy.

Razor
  • #8,929
  • Nov 23, 2009 5:10pm Nov 23, 2009 5:10pm
Quoting Razor_trader
Have you updated your historical data.

If not heres how: go to tools and click history center, this will bring up the side list of pairs and timeframes. Double Click on which ever you want. For example GBP/JPY 1 hourly has information go back to late 1999. Click download and it will download that set.

I am pretty sure this should help, the only thing im not sure about is whether it will give you this info on your chart or if it only is there for when your testing an EA strategy.

Razor
THE above works, just checked it. You need to open a chart in offline mode though to scroll back (the online charts auto correct when the current bar changes).

Good luck

Razor
  • #8,930
  • Nov 24, 2009 5:35am Nov 24, 2009 5:35am
Say I am playing using a 20 day MA and using daily charts only looking for inside bars, I basically use the MA as a buy/sell line in addition to my trailing stop. It makes sure I buy a rising market and sell a falling market. I will then look to add to this position when additional inside bars form. So if my first position is long, i then add on to it with additional longs when I see new IB's form - this is PC's logic

Alternatively, I could use different periods for the buy/sell line and the trailing stop, the only problem I see from a logical point of view is this sort of situation:

Say you use a 20 MA as a trailing stop and a 35 MA as a buy sell line.

-you get an inside bar - a buy signal, that is above the 35 MA, so you enter long but your entry bar closes below the 20 MA - your trailing stop. What moving average do you use to trail the position? you are already below your trailing stop for a long position.

For every MA you use as a buy/sell line, there are going to be trades that fall outside your parameters that turn out to be good trades

The situation you describe is exactly what I am looking for, it may not have come across in my previous post


Quoting PinMan
I'm interested as to why you'd use this as a filter. Say you are using a 36 or 45 EMA, or even one as low as 20, and you get an entry the right side of the line and for whatever reason (news most likely) you get a big move in your favour that takes you through the MA line. Say this is the start of a nice big trend, such as GBPUSD in the summer of 2008, and you have loads of entries with the trend as each entry goes free and you can add to your position. All these additional entries are on the wrong side of the MA according to this rule.

I bring...
  • #8,931
  • Nov 24, 2009 6:51am Nov 24, 2009 6:51am
Thanks Mate, worked perfectly!
  • #8,932
  • Nov 24, 2009 7:25am Nov 24, 2009 7:25am
Quoting JamesBranson
Thanks Mate, worked perfectly!
  • #8,933
  • Nov 24, 2009 8:55am Nov 24, 2009 8:55am
Every single MA I add simply adds a single more problem to my vision.
Every MA has it's strong point and it's weakness. I've tried to eliminate one MA problem by adding another - the result: you need to see how beautiful my artwork was.
Now I stick to H1 62EMA and M30 89SMA (smoothed) as my base, and pivot levels as some floating restriction, i.e I won't take any IB trade if it's nearer to any PP level, I'll wait for a break.
This is little thoughts from a 3 year old newbie.

Kennedy
Mystery Is The Final Product Of Ignorance. Hey! What's ma chart doin'
  • #8,934
  • Nov 24, 2009 4:01pm Nov 24, 2009 4:01pm
make sure the data is intact - I have noticed small chunks of data missing

Quoting Razor_trader
  • #8,935
  • Nov 24, 2009 9:49pm Nov 24, 2009 9:49pm
about trivial things like the size of the trailing stop to use........the longer the trail, the larger but fewer wins you will see, or " finding the IB bar that can't fail", one of the major issues I see, and it applies to all trading methods I suppose, is which pairs to play in tandem with one another

The Dibs approach works on all pairs obviously, but should you confine yourself to the major pairs, which are all based on the USD or would you be better to say, play one major( GBP/USD, EUR/USD), a cross (say E/J or G/J) and then maybe something exotic or a commodity based pair - the goal being to avoid, as much as possible, positive correlations between your your positions.

If your capital base is large enough and your positions small enough, the correlations probably wont matter, as long as you can handle the psychological aspect of draw down. But if your capital is limited, this might matter

Just a thought
  • #8,936
  • Nov 25, 2009 4:22am Nov 25, 2009 4:22am
Hormones, incentive, experience "make best traders"

LONDON (Reuters) - British scientists say a perfect combination of testosterone, experience and a hunger for a share of profits can produce financial traders who consistently outperform the market -- even during a crisis.

http://www.reuters.com/article/newsO...5AO02K20091125
  • #8,937
  • Edited 9:50am Nov 25, 2009 8:22am | Edited 9:50am
I thought I'd share a bit of information based on my early trading experiences. It's based on research I did to understand why I nearly caned my account the first time, hence my focus in earlier postings here on psychology rather than method. I understood that understanding 'how the market works' wasn't going to make me profitable, after all, it is just the actions of several thousand irrational human beings reacting to various things (and a few automated trading systems). What I needed to understand was how to make me a better trader and how to react to my emotions. It's also (partly) why I trade the 4H candle rather than the hourly - aside from me not being able to do the hourly owing to my job, I reckoned the time requirement any candle is about 15 minutes - I could therefore trade for 75 minutes a day to cover the five candles rather than double that, and have usable free time between each candle. It doesn't feel like 'real work' so I don't feel like the market owes me anything.

Firstly I created a 'trade expectation calculator' in Excel, which sequences trades and gives a profit or loss by a certain R:R amount based on a random number (This is something you'll see reference to if you read Taleb's work). It's very useful to sit there, hitting F9, and seeing the possible range of outcomes to (for example) having your trade hit stop-loss 50% of the time, free-trade 25% of the time, and for the remaining 25% varying levels of R:R profit, compounded over several thousand trades. It's easy to write one if you have a basic knowledge of Excel functions.

Second was a bit of research which made me realise why the 'long tails'/'letting the winners run' can make you so much money. Set up a profile in MT4 (or any other platform) covering a few major pairs (eg the main dollar pairs plus GBP/EUR) on 4H, and apply a 45 EMA and a 360 EMA. Save this profile. Now change to a weekly timeframe, get rid of the 45EMA, and change the 360EMA to 12EMA (30 4H periods in 1 trading week, so the line will be roughly the same on both timeframes). Save this weekly profile.

So you trade off the 4H chart and you take the trades in the direction of the 45EMA line, if it is almost flat you take either break. A lot of the time your trade makes a bit then retreats, either leaving you with a free-trade or not. But if you catch that 'winner' at the start of a big trend, and keep adding to it as you get further IBs in that direction, you can realise returns of better than 50:1, probably on about 1 in 100 trades. That's probably only a few trades a year, on average, that will give you this level of return, the rest not making much at all. Switching between the 4H and weekly profiles shows you where the 360EMA exit line is taking you, you've taken very small risks, 40 or 50 pips on each trade, to return thousands of pips.

Theoretically you could trade this size EMA, but it would very difficult psychologically (I couldn't do it, I use a much tighter MA). Putting this in practice and taking the repeated losses is hard and it takes a strong stomach to tolerate the small but repeated drawdowns. But the main thing you can see from this is that eliminating a slightly more frequent occurrence of losses doesn't improve your bottom line, if anything it worsens it if you try to overcomplicate it. Just go with the short-term trend and you'll get lots of free trades; if that is already, or turns into, the long term trend you are more likely to win big. Simple. But it's nothing to do with being a clever trader or anything else, I believe that this isn't a black art, but the psychology of trading is very difficult to master.
  • #8,938
  • Nov 25, 2009 11:18am Nov 25, 2009 11:18am
Shut up........you are giving away the secrets!!!

Spread the rumors, keep the truths......let your losses run and always aim to make 15 pips each and every day

Maybe it is just me, but I have found very few trading books that actually talk about trading based on probabilities, positive expected values and payoff ratios (lots of garbage about Fib lines, Support and Resistance and other silly chart patterns). Even Mark Douglas, when he alludes to the casino analogy didn't really explain it that well. It took my first 8 months of trading to really figure out that trading is much like playing blackjack while counting cards.

The only thing that really drive me "nuts" is when you get an inside bar when the market is flat - do you enter once in either direction in the case of false breaks - currently building this into a model

how much data did you crunch?

Quoting PinMan
I thought I'd share a bit of information based on my early trading experiences. It's based on research I did to understand why I nearly caned my account the first time, hence my focus in earlier postings here on psychology rather than method. I understood that understanding 'how the market works' wasn't going to make me profitable, after all, it is just the actions of several thousand irrational human beings reacting to various things (and a few automated trading systems). What I needed to understand was how to make me a better trader and how to...
  • #8,939
  • Nov 30, 2009 4:38am Nov 30, 2009 4:38am
Quoting Razor_trader
THE above works, just checked it. You need to open a chart in offline mode though to scroll back (the online charts auto correct when the current bar changes).

Good luck

Razor
You can switch off the auto scroll feature. If you hover over the various bars you get a tooltip, one of them is for a button called 'autoscroll' just click it on or off as needed.

Ensign has a very useful feature where you can replay data, so in effect load your history file, and then play it forward and it simulates the trade bar y bar, Quite handy to assess how you might have responded in real time.
  • #8,940
  • Nov 30, 2009 8:04am Nov 30, 2009 8:04am
I'm not generally in favour of filtering DIBs, although I keep an open mind about it. Here's an example from Thursday's Dubai inspired risk asset sell-off/USD rally.

The chart is EUR and there were two DIBs formed in the same conjestion area. The first broke by a few pips and was stopped the second was a clean break i.e. if you took it you had 0 drawdown and ran for 220 pips or 11x an initial risk of 20 pips.

If you use a longer timeframe to filter trades I don't see how you would filter the losing one without also filtering the winning one.
Attached Image (click to enlarge)
Click to Enlarge Name: eurh1.gif Size: 21 KB
3 traders viewing now