The DIBS Method... No Free Lunch continues | Page 457 | 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

  • #9,121
  • Jan 10, 2010 7:36am Jan 10, 2010 7:36am
hey if you have already done your tests and it works for you all the power to you. I just believe in doing as little work as possible My goal in life from now on is to produce nothing of value.........

I will say this though, when I trade the smaller time frames on equities, my window of trading is quite small, typically about 1.5 hours of the trading day. When taking into account this small trading window, my win loss ratio is much higher trading equities than FX. In the equities market place, the same pattern repeats itself day after day (it may in the FX as well, it has been a year since i looked at the m1 chart on fx)............the open is insane, i still cant figure out why from a logical point of view, then direction gets picked, a solid move is usually made, then flat for like 3 to 4 hours, and then some action in the last hour.

you could however probably throw up all your work and very few people would take advantage of it and those that do, would probably be able to return the favor in kind. I have offered up my approach to numerous new and clueless traders at work, guys that see me make my day on one or two trades and they will not do it, simply because their head isn't in the right spot and they have a very poor understanding of the way markets move. When they screw up they use excuses like "i can't read the market" or "my head just isn't there". I try to explain to them that thinking like that is like trying to explain why your number didnt come up on a roulette wheel..............


Quoting sinner-
Err I already said I've done backtests, no I don't really want to share my work for free.

Can't we just think this through using examples? As one example:

You trade daily DIBS. Your expectancy ratio will be X minus the cost of trading (which is in the form of spread plus carry and possibly plus commission). Let's call the cost of trading C and the number of lots you've traded to be L.

So over a long enough timeline, the nth trade profit P = X * Ln - Cn

Fair enough you say, no worries. C is never large enough...
  • #9,122
  • Jan 10, 2010 7:51am Jan 10, 2010 7:51am
1.82 - that is insane....


backtesting
First find out if your method has a positive expectation

1. get a chart for a particular time frame
2. download the historical data from the chart in excel/csv -date/o/h/l/c
3. determine your entry and exit method. Remember to include the practicalities of life when picking this method. You may need to sleep, eat or work.
4. Scroll to the beginning of the chart and then progress forward. look for your entry signs. Make a note of this on the spread sheet, yes the same spread sheet that has the date/o/h/l/c. i'll be even more specific, look for the exact time and price you would have entered
5. Keep scrolling forward until you get an exit. note the exit price on the spread sheet - now figure out the difference between your entry price

6. Repeat steps 4 and 5 over and over until you have gone thru all your data.

If the system is not positive, then dont bother moving on to the second phase.

Phase 2. MM
1. If you have a small account, say 500. use 1 or 2 microlots per position until you have some confidence in what you are doing and double an account

2. size your positions so ever position is risking 1% of your account. the smaller the risk, the more contracts you put on



Quoting james996
Now I don't know about FSP but I know for a fact that trading b/o's of IB's with proper mm is very profitable. Those that say other wise just did not do a proper test or don't know what there doing. I'm not trying to point anyone out but these types of blanket statements are pure BS.

As promised before note the following: 1/1/09 nav=100. 12/31/2009 nav=net of all fees and management incentives 182. Oh and these are real numbers backed by real audited statements. All other talk is just that.

For numerous reason that go way beyond the scope of...
  • #9,123
  • Jan 10, 2010 3:00pm Jan 10, 2010 3:00pm
Backtesting by hand is certainly a valid way to test a strategy and has it's advantages. However, I think that automating the strategy has even more advantages - these days you can get some pretty good backtesting software - if you are prepared to write code.

It is perhaps worth pointing out that automatically backtesting a trading system is not a trivial task. This is even more true when simulating systems that trade more frequently (backtesting long-term trend following systems on futures is easier in some respects as you can reasonably use daily data and these prices are fixed by the exchanges) as then you need to move to intraday data and timing issues then become more significant.

Additionally, you still need to step through a good number of trades to verify that your system is performing as it should. It is very easy to make mistakes and over a large number of trades the mistake will get compounded. The great thing is that once you get your system working you can then easily try it over different time periods, different parameters, data from different providers, etc...

It took me a good number of months to get an implementation of First Strike Plus (I am currently still refining my version of DIBS) that I was happy with - that matched my trades fairly closely and those that Joel posted on his blog. The particular difficulties encountered will depend on the backtesting platform but in my case it was the "One Cancels Other" orders and the fact that if you have an existing (e.g. long) trade from the previous week and the entry point for the short trade is above the existing long trades stoploss, then the long trades stoploss will be amended.

Despite all this I get quite different results depending on the hourly intraday data that I use (I have tested using FXCM, Barclays and Oanda data). My next task is to do a clean-up task on my data by comparing them where there are significant differences - not a task I'm particularly looking forward to.

Quote
Phase 2. MM
1. If you have a small account, say 500. use 1 or 2 microlots per position until you have some confidence in what you are doing and double an account

2. size your positions so ever position is risking 1% of your account. the smaller the risk, the more contracts you put onhttp://cdn.forexfactory.com/images/s.../yim/happy.gif

The second stage above is commonly called Fixed Fractional Money Management and is pretty effective too.

Good Trading to you all,
Timek
  • #9,124
  • Jan 10, 2010 4:18pm Jan 10, 2010 4:18pm
I am too much of an idiot to write software code.....by the time i figured that out, I would have been done the manual back test

One thing I forgot to mention on the back test routine is that sometimes, when price hits both a entry point and stop loss in close relation to each other, you need to drill down to a smaller time frame to see if your entry got hit before or after the stop level.....




Quoting Timek
Backtesting by hand is certainly a valid way to test a strategy and has it's advantages. However, I think that automating the strategy has even more advantages - these days you can get some pretty good backtesting software - if you are prepared to write code.

It is perhaps worth pointing out that automatically backtesting a trading system is not a trivial task. This is even more true when simulating systems that trade more frequently (backtesting long-term trend following systems on futures is easier in some respects as you can reasonably use...
  • #9,125
  • Jan 11, 2010 1:27am Jan 11, 2010 1:27am
Quote
I am too much of an idiot to write software code.....by the time i figured that out, I would have been done the manual back test
I'm right there with you on that one. Even if I took the time to set up all the codes I still would go over the results trade by trade manually in the end. I don't know of one large trader that doesn't do the same. Also I serious doubt most of the folks here have the required programming/ tech skills to pull it off. I would like to meet one trader that does this completely by computer and is trading a relatively medium account(1 million or more) . I've been at this for a long time and I've never seen it.

Hey Razor how the prop shop working out for you? I bet you get to see a lot of people doing some really dumb things...lol Hope your year is going well for you.

Quote
these days you can get some pretty good backtesting software - if you are prepared to write code.
If you don't mind me asking- what software are you talking about?
  • #9,126
  • Jan 11, 2010 4:45am Jan 11, 2010 4:45am
yes i agree.
backtesting should be done manually with an excell sheet. trade by trade.

the input is time.
the output is a better understanding of the pros and cons of your stategy AND ideas for new strategies.
  • #9,127
  • Jan 11, 2010 7:40am Jan 11, 2010 7:40am
Regarding the backtesting software, i use a combination of both, excel and software.
I use Ami for charts, and for each system i want to test, i write some very simple lines of code just to make the entries and exits visual in the chart.
Next i export the entry and exits to excel, and then manually (and visually) check each entry and exit.
All i have to do in excel is note wich trades are winners, losers or BE.
It is still a lot of work (i like to backtest on as much years of data i could get my hands on) but this way i dont have to stare at the charts to much for where the entry and the exits are.
For me this is the fastest and still the most reliable way to get some good results.
I have never been able to get the same results with the built-inn backtester from the software i use, but then i'm not a coder.
Persistence will get it. Consistency will keep it.
  • #9,128
  • Jan 11, 2010 8:27am Jan 11, 2010 8:27am
One of the best investments ( I call it an investment because I use it and can come back to it as often as I like and it helps me fine tune my thinking) was paying a small amount for a program 'Forex Strategy Tester ver 2'. It has data that stretches back to late 2000, about 20 or so currency pairs (all the majors and some fiddly ones) and you can manually test or code a test to run in it.

Once you have updated your charts (much the same way as updating MT4's tick data) you can open up test mode, select a range that you want to test, import the pairs (USD/JPY etc) and the timeframe you want to focus on and then test away. Like I said you can test manually by going from bar to bar, or from a selected tick size. Along the way you can enter trades that you would take (IBs for example) and the manage them. You can have a profit chart along side that automatically measures your current equity, drawdown balance etc.

You can save your progress as you go so if your manually testing IBs on EUR/USD from 2002 on an hourly chart you can simply save your work and come back to it later on (because you are not going to manually test some 56000 candles overnight).

I have found it handy just even to test some random ideas or to help refine a plan for how to tackle a strategy. I found it less time consuming than an excel spreadsheet and less confusing trying to set up.

You can download the demo from the following link. It only gives you 1 month of data and/or 30 mins of testing at a time but for under $200 it may save you blowing your account.

Razor

http://www.forextester.com/
  • #9,129
  • Jan 11, 2010 2:23pm Jan 11, 2010 2:23pm
Trading 4hr inside bars between Jan 2001 - May 2003

Total trades 169

Loss trades 156

Profit trades 13

That would take some bottle to trade through that!
  • #9,130
  • Jan 11, 2010 3:46pm Jan 11, 2010 3:46pm
Hi guys, I've been trading IB's on the major currencies for 2-3 months now, with pretty promising results. However there are so many trades that are ending up being losers/BEs and I often end up emotionally drained for days due to the losses, I know as a trader that losses happens to everyone and this is something one must accept.
But the latest week i've been checking IB's on the stock market on the 1h tf, what i've seen is that it is much fewer IB's that are ending up in losses/be.

What do you guys think about IB's may be more accurate on stocks?
  • #9,131
  • Jan 11, 2010 4:34pm Jan 11, 2010 4:34pm
Quoting james996
If you don't mind me asking- what software are you talking about?
Hi James,

No problem - I use Trading Blox Builder edition, which as you probably know, was initially developed by a former Turtle. It originally was aimed at Long Term Trend Following Futures traders (as you would expect) using daily data but now works well with intraday FX too. I've used it for over four years now and am very pleased - it has never left me wanting to look for alternative software.

It allows backtesting against a portfolio of markets with the attendent position sizing and risk management. I've found it to be highly customisable and if something is not possible the company will release an updated version very quickly.

I've found your posts very enjoyable/useful and totally agree that it is easy to end up endlessly tweaking things (adding filters etc) to get the "perfect system", which simply does not exist. The more you add the more the system becomes tied to the market at the time you do the research and development.

Less is more when it comes to trading (within reason) and it's all about the execution...

Cheers,
Timek
  • #9,132
  • Jan 11, 2010 4:44pm Jan 11, 2010 4:44pm
Quoting DemoForever
Trading 4hr inside bars between Jan 2001 - May 2003

Total trades 169

Loss trades 156

Profit trades 13

That would take some bottle to trade through that!
Of the loss trades how many came out BE and how did you trade them. Remember that with the 4 hourly TF your trade is taken with a weekly view, that is short if the week is down and long if it is up. Also what pair was that?

Razor
  • #9,133
  • Jan 11, 2010 4:48pm Jan 11, 2010 4:48pm
Quoting Razor_trader
Of the loss trades how many came out BE and how did you trade them. Remember that with the 4 hourly TF your trade is taken with a weekly view, that is short if the week is down and long if it is up. Also what pair was that?

Razor
BTW that is PCs recommendations which leaves the door wide open for traders do tackle these IBs in a different manner, I definatly take more into consideration that just weekly direction but that is a good starting point.

Razor
  • #9,134
  • Jan 12, 2010 12:57am Jan 12, 2010 12:57am
Hi razor

Sorry about that its the eur/usd,buying above the weekly open and selling below.Exiting on a 40 ema.Simple works.
  • #9,135
  • Jan 12, 2010 1:56am Jan 12, 2010 1:56am
Quoting DemoForever
Hi razor

Sorry about that its the eur/usd,buying above the weekly open and selling below.Exiting on a 40 ema.Simple works.
13 winning/156 losing...but this is trading all about.
if i understand u right, u get 13 winning trades which make the profits over the year.
the losing trades - i guess - half of them are BE-trades?
  • #9,136
  • Jan 12, 2010 2:30am Jan 12, 2010 2:30am
Hi russiatrader

Im not taking half off at b/e.The test im doing is trying to find the most pain i can handle from large up and down equity/margin fluctuation.If i think i can cope with this ,taking half off at b/e should be a lot less stressfull.
  • #9,137
  • Jan 12, 2010 2:52am Jan 12, 2010 2:52am
13 winnings / 156 losses - that's a recipe for disaster. One win for every 12 losses, you can't recover the drawdowns. And risking only 0.3% per trade you'll have periods with 20% drawdown on the long run.
It is as you would have 60% drawdown by risking 1% on every trade. You must more than double your account just to be at breakeven.
  • #9,138
  • Jan 12, 2010 3:02am Jan 12, 2010 3:02am
[
Quote
quote=Razor_trader;3355830]One of the best investments ( I call it an investment because I use it and can come back to it as often as I like and it helps me fine tune my thinking) was paying a small amount for a program 'Forex Strategy Tester ver 2'. It has data that stretches back to late 2000, about 20 or so currency pairs (all the majors and some fiddly ones) and you can manually test or code a test to run in it.
I agree with Razor here entirely, I have also made the same investment and this has helped my trading no-end. When demo trading, the time wasted waiting day by day for the market to give you info for testing is very time consuming and if testing a strategy I think it is a bit unrealistic to demo trade for 5 years to test a strategy especially if this method then turns out to be unprofitable or requires tweaking as such.

I say 5 years because this would be the absolute minimum I would consider testing any Idea.

It has been discussed here a lot lately the necessity to back test your method and is something I have become absolutely pedantic about. This testing gives you the confidence to trade your plan as you have done it successfully time after time or you should have if you did the required work.

You also walk away with the knowledge of previous and I do say previous maximum draw down for your method as well as consecutive losing trades. This does not guarantee that this is what you will experience in the future but does give you an Idea what you can expect and thus allows you to manage your risk accordingly.

I have completed huge amounts of back testing and still spend a lot of my spare time testing different ideas due to my obsession with trading. This has given me confidence to trade my systems and stick with them even when losing which is always hard. As also mentioned this is extremely time consuming and hard work but you tell me what isn't when trying to excel at any given profession.

FSP has been mentioned a lot on this thread so I will share some of my testing with you all. I have exported it to excel so you can see the trades and the equity curve as without the software you wouldn't be able to read it. In this test I have tested the EUR/USD pair on five years of data and I risked 1% of capital on every trade, I used 100k as equity for a nominal amount.

Equity curve
http://i47.tinypic.com/255kis2.jpg

I I have attached the excel spreadsheet for any that are interested.

Trade well

Perry
Attached File(s)
File Type: xls euro fsp ons CONVERTED risk is 1 percent of capital jan 2005 to sept 2009.xls   133 KB | 670 downloads
  • #9,139
  • Jan 12, 2010 3:09am Jan 12, 2010 3:09am
Quoting Sauron
13 winnings / 156 losses - that's a recipe for disaster. One win for every 12 losses, you can't recover the drawdowns. And risking only 0.3% per trade you'll have periods with 20% drawdown on the long run.
It is as you would have 60% drawdown by risking 1% on every trade. You must more than double your account just to be at breakeven.
i dont agree.
if u have for example 156 losses, u will close out about 78 trades 1:1 , and the other half (78) with a 1:1 loss.

the 13 winning trades are trailed with a 40 EMA and will be huge, i guess around 10:1 or bigger.

so u have 10 multiplied by 13 winners= 130

and losses 78 multiplied by 1= 78.

78 are breakeven (or u call them 1:1 trades)

you end up with 52 positiv numbers, comm.&slippage not included.

u should consider to sell the first half at 1:1, like PC is doing it.
  • #9,140
  • Jan 12, 2010 3:26am Jan 12, 2010 3:26am
Well, let's talk about numbers (and the fallacies) .

Quote
the 13 winning trades are trailed with a 40 EMA and will be huge, i guess around 10:1 or bigger.

I suppose that every winning trade will give you 10:1. But your profit will not be 10:1 but 5.5:1 because you take half of the profits after 1:1. .
So 5.5 * 13 = 71.5 which is smaller than 78, the usual loss.
And that's without considering spread and slippage.

Quote
the 13 winning trades are trailed with a 40 EMA and will be huge, i guess around 10:1 or bigger.
The trailing to EMA or SMA is not a good idea anyway. You will get out every time after the retracement instead of profiting from it. It is like buying on support and selling at resistance.
2 traders viewing now