Manual Backtesting

Prophet,

I'm not sure what two points you're referring to. In any case, I wasn't referring to an automated system that requires minimum trading input, and that can be easily coded and backtested by software. I'm talking about a trading method with rules that have to be closely followed and implemented by the trader. Not discretionary, yet still too complex to completely code (not that it CAN'T be coded, i.e. it is mechanical and not discretionary, rather that it might be complex to do so. That is the case with my method and I imagine with many others too).

A method like that requires the manual backtesting for the trader to get the confidence in the method that is so absolutely crucial in trading...

all the best,

Phil
 
Quote from splphil:
I'm talking about a trading method with rules that have to be closely followed and implemented by the trader. Not discretionary, yet still too complex to completely code (not that it CAN'T be coded, i.e. it is mechanical and not discretionary, rather that it might be complex to do so. That is the case with my method and I imagine with many others too
More...
Ok. I understand now. Since your method is complex and can’t easily be automated, the trader must have enough discipline to follow the rules exactly instead of modifying them based on recent results. I hope you may find some automation or assistance to allow you to deploy more of your methods and/or on more markets.

My systems are similar in a way. Usually when a system has big losing day it will recover at least 50% of the loss the next day, sometimes over 100%. Thus they can’t be second-guessed over a period less than a few weeks. They must be traded consistently and size must always be diversified inter-system. My account size, leverage (max drawdown) preferences and contract sizing for ES, NQ and 6E limit me to running 7 to 10 systems when there are 30 to 100 I could now run simultaneously on a sufficiently large account. I also have to adhere to rules governing what to do in cases of data interruptions. Automation turned out to require more discipline than I used to believe.
 
Quote from prophet:

Ok. I understand now. Since your method is complex and can’t easily be automated, the trader must have enough discipline to follow the rules exactly instead of modifying them based on recent results. I hope you may find some automation or assistance to allow you to deploy more of your methods and/or on more markets.
More...

Well, complexity has many definitons. The actual rules of the method are pretty simple. My difficulty with coding them is just because I have several different entries, several different exits, which each have their place depending on what's happening in the market. About discipline; my method is actually quite adjustable in many respects, leaving room for a trader to adjust it to his own personality, thus helping a lot with being able to trade it with that needed discipline.

Quote from prophet:

Automation turned out to require more discipline than I used to believe.
More...

Too right on that. It also entails some emotional difficulties of it's own, like worrying that the system was actually automated %100 correctly, and won't start buying and selling where you didn't mean it to. That was always one of the things that kept me away from trading that way. But maybe I'm just paranoid :).

All the best,

Phil
 
This is so 2004.

I love it,

The benefit of manually back testing is that one can use other tools like drawings, Yes they rely on number but I do not thing that is a programmable system that can draw channels. If so let me know.
This allows the use of an additional tool, It is simple but hard to implement which makes it less used.

If some thing is less used people assume it is batter, It may be true, I mean Trading with moving averages and stochs and what every just don't work on their own. putting them together is optimising not to say it is not a good way to go about it ,

Manual back testing also allows one to look at bar formation and develop new ideas.

I used Bollinger bands just cause it seemed right I had no idea what I was doing, I would simply buy when a bar was pushed up.
Only through manual back tests did I understand how to use them throughout a trade.

Here is a interesting site alot can be learned just playing around,

http://chartgame.com/play.cgi?duq4rx-10
 
I am wondering how you guys manually back test.

I have a issue that I cant play with different time frames when i manually back test.

I always end up accidently looking ahead. There is just some thing about changing time frames that is too emotional.

Some times to see the trend that a trend is in one can see it on a 5 min chart, other times it can only be seen on a daily ..

I'm wondering if you guys have this problem and how you overcome it.
I just cant seem to make a rule for this .
 
You must log in or register to reply here.
Back
Top