In trading of financial markets, I wonder it would depend on individual trader's definition of over-optimisation when designing trading systems/strategies. http://www.elitetrader.com/vb/showthread.php?threadid=192056
I would define that a strategy is over-optimized when the optimization process reduces the life trading returns, instead of improving them. But normally you can find that out relatively easily with out of sample tests or walk forward tests. It's not 100% safe, though - when you have to tweak a strategy until the out of sample test comes out great, it's most likely still over-optimized.
Heh what a thread, didnt bothered to read all the noobs reply. With that graph of yours with A,B and A+B*2, i think you are far away to even talking about portfolios. In my own automated strats, I have at least 5-8 strats, all uncorrelated, different time frames, asset classes, and methodologies if your arguement point is that no 1 strategy is profiting. its basically a portfolio of strategies, and your point being ? At the end of day, with or without the use of leverage, all strategies must return Alpha, hence lowering of risk is to use multiple different strats, and hopefully it eeks out some Alpha. its very easy to think a strat like A and B then A+B*2 like a quanty formula, but in fact its a fooled by randomness illusion
jcl i think u shld cut. Obviously anyone and everyone reading your stuff knows you know nothing. Nobody in this shitty world, quants and HF will talk about their business models here. If you are so smart, why aren't you so rich. Why aren't you got interviewed on Bloomberg as one of few to touch 1st 100m before 30. (i touched my 1m b4 30, at least). So pls.
People come here for entertainment and the rare useful post that may benefit their work. Nothing else. The rest are posers like you, what do you have to show for? Any proof of your claims? Of course not, I am not asking for proof because I could not care less about your performance. I participated earlier because OP made factually incorrect statements. Start your own thread to invite all who salivate over pissing contests...
Gosh, a may be new to trading, but I can't imagine you can make money using a losing strategy. you would think you need a winning one... or am missing something?
keyword "risk adjusted returns". Look it up, if you spend just 15 minutes understanding what it means then you know more than 70% of people on this site. I am serious. But generally you are right. Losing strategies do not help your total return.
Thank you, will start drilling down links http://www.investopedia.com/terms/r/riskadjustedreturn.asp#axzz1xTN7dqSn
Welcome to this forum. As you can already see in this thread, we're not actually elite - we have here members of any sort of intellect and any age down to 12. Yes, you'll also need a winning strategy for making money with a losing one, and it must fulfill certain conditions, such as negative correlation. You'll understand when you read the thread. At least some of it.
keyword "risk adjusted returns". Look it up, if you spend just 15 minutes understanding what it means then you know more than 70% of people on this site. I am serious. how about just trying to learn what you can from this board? "But generally you are right. Losing strategies do not help your total return. " the goal of this thread is to show specifically how to win with losing strategies.