Oh, then you really "arb"? now that is something different. but anyways, don't worry, i am not even close to playing millisecond- games. i am amazed that a single person can do that profitably. do well ... PS: acts 9:7, "The men who traveled with him stood speechless, hearing the voice but seeing no one."
TSGannGalt - You formulate in detail what I am in the middle of changing my overall system development process into. Great post(s). Thank you. BTW: I look at price/volume only trading select ETFs and non dividend yielding / highly liquid common stocks - my target is to hold positions for 2-10 days.
There's no point of PM replies for some stuff so... As mentioned a few posts back, I use discretion as part my process from my point of my skills. I am able to develop a model and trade them without adding any discretion, profitably. Simply, you can take any crap system and make it look profitable by using top-notch technology and algos on paper. Once you implement it, the added value (top tech and other perks) will not keep up with real life trading. First thing required is to have the "mind" to DUM (in acrary terms) "or" expose/extract an edge out of the market. You're simply carrying over that "mind" on a constant/monitoring basis as the discretion factor. I just hope some Tradestation user doesn't try copying me, because they won't be able to pull it off.
definitely. i have never seen a system that matches the backtest. you gave me some brain food. yet i think i am not good in predicting when our systems do well. always think i can, but if i am taking away positiveMemoryBias i guess i am pretty much at random ... nevertheless, the whole issue seems to me like regime change modeling and i do believe humans can beat machines in that. ... well, certain humans ...
10:12 very interesting. i would have thought the price people would exceed by far, then it was just the exogenous and now we are at about even. i am tempted to start a thread on which time series the exogenous guys do use, but that is probably asking too much ... gets into fishing ... thnx everyone for participating.
hi mbatrader, do you use matlab2ib? we've considered this interface too, but then decided to give twslink a try. what are your experiences in terms of speed, reliability etc...? what do you think are the problems / disadvantages of twslink? thx!