It's only that good because its HISTORY based on optimization. It's like looking at every single MA cross over after its already happened then building a set of rules that says if/then. The problem with that is you do not see the future and all the cross overs that uncrossed 2 bars later. Therefore it is totally based on hindsight and most likely will not work real time. My ex gf was a total doll till 3 months later and all her claws and flaws came out but if I married her before I saw reality I would be miserable now and that's what your sort of asking OP to do.
It depends on the type of system. Proving your system is robust through at least one market cycle is best.
I believe these are two different issues. I'm referring to qualifying a system. I think you are referring to qualifying a trader. I trader can have a superior system and still blow it because of the human element. I would agree with your time frame regarding testing a trader's ability to control himself.
NO, please do not trancuate posts to reflect your misconceptions. I wrote: "This PI is the inverse of the well-known coefficient of variation in Statistics" and then I wrote that "which [meaning the PI] is also called the Signal to Noise Ratio, and then multiplied by sqrt(trades). " if y = 1/x then x = 1/y this is what inverse means. The coefficinet of variation is the inverse of SNR means the SNR is the inverse os the CoV. Why do simple things like that confuse some people like you?
I did forward test all three symbols on my IB paper trading account. The results of the forward test are almost same as the backtesting. The system traded on market order, executions time were less than 500 ms. The slippage was very small, once or twice a day, one tick only if it happened. Slippage is not always bad. It happened in two directyions, some times in your favour, some times against you. I also traded ES and NQ on my live account for the last three months. As I didn't have enough capital to trade both ES and NQ all of the times, so I switched trading between ES and NQ. Most time I traded one contract, few times two contracts. Believe it or not, I always picked the wrong symbol and doubled the size on the bad days. I attached a zip file that includes the performance reports for the three backtestings mentioned in my first post. The test period started from 2009/07/20 as I don't have any test data earlier than that and IB only provides historical market data not more than one year old. Please take a look if you are interested. BTW, how can I insert the jpg images into the message body here?