Pitfalls of system development.

[MarkBrown]

Underlying themes I see repeatedly in the world of automated trading. There are many reasons others I'm sure could add to this basic list, but below are some of the ones I run across on a consistent basis.

1. There is no perfected predefined expectation of outcome.
2. Data itself is flawed and full of deception, many details on this.
3. Markets are often split in half treating up/down moves the same.
4. Cost of business is often omitted because of unrealistic expectations.
5. No forethought of catastrophic events and actions to take in event.
6. One size fits all approaches are far too often the go to aproach.
7. Needless to say over optimization on too little amounts of data.

 

[Simples]

Underlying themes I see repeatedly in the world of automated trading. There are many reasons others I'm sure could add to this basic list, but below are some of the ones I run across on a consistent basis.

1. There is no perfected predefined expectation of outcome.
2. Data itself is flawed and full of deception, many details on this.
3. Markets are often split in half treating up/down moves the same.
4. Cost of business is often omitted because of unrealistic expectations.
5. No forethought of catastrophic events and actions to take in event.
6. One size fits all approaches are far too often the go to aproach.
7. Needless to say over optimization on too little amounts of data.

Doesn't same pitfalls apply to discretionary trading too, or why wouldn't that be?

 

[GhostOf144]

The biggest problem people have is trying to develop a system without knowing anything about the way markets move, execution, or trading. The numerous books and internet nonsense promote this as though one should develop a system, then learn to trade.

Cart before the horse...

 

[Sprout]

Underlying themes I see repeatedly in the world of automated trading. There are many reasons others I'm sure could add to this basic list, but below are some of the ones I run across on a consistent basis.

1. There is no perfected predefined expectation of outcome.
2. Data itself is flawed and full of deception, many details on this.
3. Markets are often split in half treating up/down moves the same.
4. Cost of business is often omitted because of unrealistic expectations.
5. No forethought of catastrophic events and actions to take in event.
6. One size fits all approaches are far too often the go to aproach.
7. Needless to say over optimization on too little amounts of data.

What would you define as workarounds or ways of avoiding/mitigating the pitfalls you've outlined?

 

[lawrence-lugar]

What would you define as workarounds or ways of avoiding/mitigating the pitfalls you've outlined?

You can't avoid the inherent dangers that linger in the market. That's partly why it's so difficult to succeed, or succeed fruitfully, trading. If you want safeness, put your money in a savings account generating 0.10% yearly.

That's like asking...can you guarantee to me that another car won't hit me today when I leave the house and neighborhood. (of course not)

All automated trading systems and formulas...always reminds me of Long Term Capital Management of the late 90's imploding.

I hate sounding like a broken record, but...Trading is part art, part science. Facts vs random variables. Happyness and greed vs fear. etc etc
The market is incredibly complex, it's anything but linear...to predict, and trade and manage the future.

 

[bashatrader]

Good article here about backtesting pitfalls. Must read.