Value of Backtesting and Stops

Well...you can go ahead and be the best backtester....I will just be the best trader. I am not saying backtesting isn't important (for those of you who jump), just that it's a minor part of being profitable...at least for me...trading profitably for 7 years now.

This is just my experience. I backtest things and go...yeah...well...okay. I guess I am different from all the profitable traders here, so be it.

If you are running an automated system...okay..do you backtest and program your stuff.

I used to go to these trader meetings in Orange County CA. Full of guys who were backtesters (backtestors) and were the most stressed out guys you ever saw. They were losing everything or at least not making money with their backtested systems.

I am skeptical of people who put a great deal of time into backtesting. Many times, and intuitive test is adequate. Your trading skill is much more important. At least mine has been for a good number of years so far.
 
How do you know you are even close to "the best trader". What if you met a backtester who blows away your returns year after year?

There are lots of crappy backtesters out there, just like traders.

Backtesting *IS* a trading skill.



Quote from pairsarb:

Well...you can go ahead and be the best backtester....I will just be the best trader.
More...
 
Quote from traderdragon:

How do you know you are even close to "the best trader". What if you met a backtester who blows away your returns year after year?

There are lots of crappy backtesters out there, just like traders.
More...

Adjusting your settings based on historical data (yesterday, last week, last month, last year) and trading with those settings is like driving a car while looking in the rearview mirror. (in other words: you may get somewhere but it will be awfully slow and you cannot compete with the guy who looks out to the front)

There is absolutely no substitute for a system which does not require adjustment, no historical data, works on all timeframes, in all markets and gives you real time the status of the markets.

This is the stuff why Jesse Livermore was able to average 150% / week. Read his book and read how he scoffed about these "scientific methods". Read how he wrote about the "sucker rallies". The games that are being played are all slightly different, this is why no mechanical system can trade all markets, all timeframes. (this is the same as at the poker table or in the pool room: all the sharks have a slightly different game.)

This is called "reading the tape" and it is not necessarily restricted to the short timeframes only.

Holmes
 
Hi traderdragon
I read from your posts that you're being successful with your mechanical, backtested systems so I congratulate you for that.

I have a couple of questions, if you don't mind.

How long has your systems "survived" on average since they went live?

How do you decide when it's time to turn a system off for reviewing or replacing?

Are your systems mainly trend following or countertrend?

Be well
GS
 
Quote from Holmes:

Adjusting your settings based on historical data (yesterday, last week, last month, last year) and trading with those settings is like driving a car while looking in the rearview mirror. (in other words: you may get somewhere but it will be awfully slow and you cannot compete with the guy who looks out to the front)
More...

More evidence you have no idea how to backtest properly. A good system will test against a very very large set of parameters. Only inexperienced backtesters develop systems that are perfectly curve fitted to a tiny set of parameters which will only work on historical data.

There is absolutely no substitute for a system which does not require adjustment, no historical data, works on all timeframes, in all markets and gives you real time the status of the markets.
More...

A worthless statement. Can you statistically prove such a beast exists?


This is the stuff why Jesse Livermore was able to average 150% / week. Read his book and read how he scoffed about these "scientific methods". Read how he wrote about the "sucker rallies". The games that are being played are all slightly different, this is why no mechanical system can trade all markets, all timeframes.

(this is the same as at the poker table or in the pool room: all the sharks have a slightly different game.)
More...

Oh, you mean like those poker tables online where backtested poker robots take peoples money? :D

This is called "reading the tape" and it is not necessarily restricted to the short timeframes only.

Holmes
More...

Jessie died broke. Enough said.

No mechanical system can trade all markets and timeframes? No kidding really? Does it need to? No. I have systems that work great on NAZ, great on NYSE, but not great on both. They are different markets.
 
Quote from inCom:
Hi traderdragon
I read from your posts that you're being successful with your mechanical, backtested systems so I congratulate you for that.

I have a couple of questions, if you don't mind.

How long has your systems "survived" on average since they went live?
More...

My first system lasted about 5 years and faded. The rest are going 5+ years now with no sign of slowing down. They back test well for 15+ years straight. Probably more, but I dont have enough data to go further back.


How do you decide when it's time to turn a system off for reviewing or replacing?
More...

I track its statistical behavior. win%, profit factor, draw downs, average win per trade, all slowly get worse over a long period of time as systems fade. After some time, they reach a point where im no longer interested in trading them, although they still have "edge" left in them.


Are your systems mainly trend following or countertrend?
More...

No comment. :cool:
Closely guarded secrets.
 
Quote from traderdragon:
....

Jessie died broke. Enough said.

.....
More...

No need to respond to the rest of your worthless dribble.

Having 5 Million US in a trust account and getting an allowance from that in 1940 I would not exactly call being broke.

Keep going on with your popular and misinformed delusions.

Old Chinese proverb: Those who say that it cannot be done ought not to interrupt those who are doing it.

Moonbat button applied, no need to respond!

Holmes over and out.
 
When someone knows they are beat, and cant defend their position, he often just runs away like a child :D

Actually, youre wrong, he only had 1 million in the trust. So yeah, he didnt exactly die broke, he only lost the majority of his fortune.

But more importantly, you didnt bother to address any of the real points I made, because you cant.

People who dont posses the skill to backtest properly often scoff at it and try to convince themselves its worthless. The blind trying to convince themselves that being blind is ok.




Quote from Holmes:

No need to respond to the rest of your worthless dribble.

Having 5 Million US in a trust account and getting an allowance from that in 1940 I would not exactly call being broke.

Keep going on with your popular and misinformed delusions.

Old Chinese proverb: Those who say that it cannot be done ought not to interrupt those who are doing it.

Moonbat button applied, no need to respond!

Holmes over and out.
More...
:D
 
Many traders have an idea of what will make money. Often they find a developer to build their secret kick ass trading strategy or worse yet build it themselfs. If they servive the development cycle they launch this system into the live world. Hopefully they understood all the events that can happen with the actual execution of such a strategy. Such as routing provider issues, Client connection closes out, or a major market event. I would say backtesting is a great idea to proof that your indicator or indicators is a valid tell on which way the market is going to go. Using this backtesting code you can simulate buying power restrictions, profit targets, stop preferences, etc. Where testing this with real money could discourage you and cause you to give up on what could have been a briliant idea. ..greg
 
You must log in or register to reply here.
Back
Top