Why do people not backtest their systems?

[ashis_ch]

Backtesting should be followed by a good forward testing

 

[Vicorka]

There are many systems described here but it seems almost all are not backtested. I see nowhere some tests or performance statistics. I wonder why people trust their money to untested systems? How can someone know that his system is profitable without having backtested it over at least 4 years? A backtest is no guarantee for a profitable system of course, but not backtesting seems to be a guarantee for an unprofitable system.

Free source of back tested technical indicators:
Trading Systems Scan
Follow the link to signals to see an indicator's settings.

These guys are working in that direction. If you have something specific in mind, contact them and maybe they will add it on their website.

 

[malaguti]

Free source of back tested technical indicators:
Trading Systems Scan
Follow the link to signals to see an indicator's settings.

These guys are working in that direction. If you have something specific in mind, contact them and maybe they will add it on their website.

great link I think. thanks Vicorka. I wish somebody could backtest a P&F strategy or even a renko strategy. Just cant be done though

 

[BeginnerJoe]

Backtesting should be followed by a good forward testing

Then followed by a real trading statement demonstrating it's not a looser.

 

[spaceage]

Why bother back testing when you can forward test on demo ? But both approaches are like testing punches against a dummy that doesn't punch back. Why people believe the market is a dummy that likes to give you money as if it's an ATM I will never know. But let me give you a secret: the market is alive, and not only does it punches, it also kicks. Its kicks are always 100% accurate and straight in the b*lls. Now, you still want to come use the ATM do you ?

Very well put. The only way around this seems to be using a small live account, and then gradually increasing the account size...

 

[JimTru]

Oh, definitely backtest your idea...if you can. But you won't like what you find. You'll find: 1) The idea does poorly and now its back to the drawing board, or 2) I can't consistently quantify why I do what I do.

Good luck!

 

[spaceage]

Oh, definitely backtest your idea...if you can. But you won't like what you find. You'll find: 1) The idea does poorly and now its back to the drawing board, or 2) I can't consistently quantify why I do what I do.

Good luck!

Are you saying your method doesn't actually have any particular parameters?

 

[spaceage]

Possibly, but an average intelligent person can learn those software skills in short time, like 2 weeks.

I agree, but it need not be that complicated at all. Depending on which platform a trader uses, it can be as simple as scrolling back through the charts. It's a lot of grunt work, & that's what (in my opinion) prevents most people from backtesting their systems. People associate more pain to it than the pleasure, not taking into accounts the huge benefits of going through the hard labour...

 

[Vicorka]

I agree, but it need not be that complicated at all. Depending on which platform a trader uses, it can be as simple as scrolling back through the charts. It's a lot of grunt work, & that's what (in my opinion) prevents most people from backtesting their systems. People associate more pain to it than the pleasure, not taking into accounts the huge benefits of going through the hard labour...

Scrolling a chart back in the history gives you a visual back testing. Usually people see what they want to see. Plus, it will take you hours to scroll a chart and test one combination. Real back testing should give the performance numbers. You should be able to scan 100K system's settings combination. In addition, you should be able to scan 1K stocks (500 as minimum) which will put additional pressure on your scanner. At his point of time such scans cannot be put ion he web due to the time consuming costs. If you have raw data and if you have knowledge of programming you can do it.

However, when there is a demand there will be a supplier. You already may see simple free system's scanners for number of technical indicators at
free trading system scanner
To speed-up the process of these scanner improvements, traders have to contact such companies and help them by letting them know what exactly traders are looking for.

 

[spaceage]

Scrolling a chart back in the history gives you a visual back testing. Usually people see what they want to see. Plus, it will take you hours to scroll a chart and test one combination. Real back testing should give the performance numbers. You should be able to scan 100K system's settings combination. In addition, you should be able to scan 1K stocks (500 as minimum) which will put additional pressure on your scanner. At his point of time such scans cannot be put ion he web due to the time consuming costs. If you have raw data and if you have knowledge of programming you can do it.

Not necessarily, but then again, I'm talking in terms of well defined systems, i.e. with defined entry and exit rules. If people see what they want to see, then clearly the rules are not defined enough. For that form of discretionary trading, you can't really backtest anything - because the trader will have nothing to backtest!

 

[JimTru]

Are you saying your method doesn't actually have any particular parameters?

I'm saying that the marvelous human brain has more parameters than one can know consciously, and thus code in a backtest.

 

[spaceage]

I'm saying that the marvelous human brain has more parameters than one can know consciously, and thus code in a backtest.

Backtesting is not about testing discretionary methods, but rather for testing systems with clearly defined rules.

 

[spaceage]

Seems to me that you've really reaped the rewards of backtesting. Had you not backtested that system, you wouldn't really know the system's statistical success rate. & that rarely ever is a good idea... Trading is a business, & it's all about having as much edge as possible.

 

[Hittfeld]

When I started trading, I read several trading books where systems were described with clear defined rules, suited for beginners. Almost always the author explained how he had tested his system with "looking through charts" and "life trading". The thing is, I have backtested some of those system and almost all were unprofitable.

A good example is Thomas Carr, "Micro Trend Trading". In this book he describes 8 systems and recommends them to beginners. All 8 of his systems lose money already in a simple 1-year backtest, some only a little, some really big time.

This tells me that looking through charts or life trading is useless for testing systems. And it's really dangerous to beginners to follow some system from a book or a trading forum because most are not tested.

And its useless to backtest with crappy data, crappy programs and crappy algos. But you had those three validated and confirmed - hadn`t you?

Regards

 

[spaceage]

Well, when you get a crappy result in a backtest, then it's usually not the data, nor the program. It's the algo that is crappy. Which is good to know when you considered trading it.

When you however get a good result in a backtest, then you might start to contemplate the quality of your data or your test method. Systems with a good backtest can still fail in real trading, but systems with a bad backtest will certainly not succeed in real trading.

Well said

 

[tradern]

agreed