How do you know you have an edge

[killthesunshine]

Quote from Nexen:

You have an edge when you consistently make money over a long period of time. A period embarking different market scenarios allowing you to test yourself during different weather conditions.

right, the proof in the pudding is in the eating but you can lose a lot of money as you go from strategy to strategy letting them play out

obviously the sooner you know the better. is there an easy way to know what you have sooner?

we know backtesting is nothing more than curve fitting.

 

[intradaybill]

Quote from ivanbaj:

Testing for fair coin is well documented. For example

http://www.aiaccess.net/English/Glossaries/GlosMod/e_gm_test_1.htm

But how this applies to trading results? Perhaps there is no difference...

Good question. very good.

At the end of the day, if you are making money, nobody that accuses you of trading a random system will make any sense.

I had a dispute here once with a self-proclaimed expert of nothingness obviously, who made the totally absurd claim that a trading system is random if a large collection of random systems averaged better results. This "brain" is lurking in these forums, warning!

I never presented him with a formal proof of why he was wrong because he did not deserve it due to his arrogance and hostile attitude.

 

[intradaybill]

Quote from killthesunshine:

we know backtesting is nothing more than curve fitting.

Nope, backtesting is not always curve fitting.

How do you define curve fitting?

 

[killthesunshine]

Quote from intradaybill:

Nope, backtesting is not always curve fitting.

How do you define curve fitting?

how long till you become confident past tested condition hasn't changed in the present sense till you walk it forward with real money?

 

[intradaybill]

Quote from killthesunshine:

how long till you become confident past tested condition hasn't changed in the present sense till you walk it forward with real money?

If your system depends on market conditions then it will lose money when they change. It must be market condition invariant.

 

[walterjennings]

Quote from achilles28:

With the VIX this high, everything works. False positives galore. The real test is when the VIX drops and the market ranges tight. Newbs drop like flies. Backtest thoroughly over a period with the VIX under 20. If it turns a profit, and you're okay now, *maybe*. If not, be wary.

I've neglected VIX in my research. Correct me if I'm wrong, but isn't the iPath S&P 500 VIX Short-Term nearing a two year low? Maybe I'm reading it incorrectly.

 

[ivanbaj]

Quote from intradaybill:

Nope, backtesting is not always curve fitting.

How do you define curve fitting?

I also believe that curve fitting is ok. After reading few quant sites and books I realized that the evil curve fitting is not that evil and many times necessary and a goal.

 

[walterjennings]

Quote from ivanbaj:

I also believe that curve fitting is ok. After reading few quant sites and books I realised that the evel curve fitting is not that evel and many times necessary and a goal.

Curve fitting is good as long as the curve you are trying to fit to is a good representation of what you are trying to predict, as well as having a small enough grammar as to not be mapping directly to the sample set.

 

[BSAM]

Quote from ivanbaj:

I also believe that curve fitting is ok......

Hmm...Me an' you musta gone to different schools. I have found curve fitting to be bad. Very bad.

 

[killthesunshine]

Quote from walterjennings:

Curve fitting is good as long as the curve you are trying to fit to is a good representation of what you are trying to predict, as well as having a small enough grammar as to not be mapping directly to the sample set.

you still don't know till you walk it forward with real-time trades..or are you saying you know for sure your historical result will prove out tomorrow?