if it were clear, there would be no discussion. you were obviously hounding the kid, i'm not the only one who noticed it. he pointed to a thread on stat-arb, which has everything to do with hft. it's obvious that's what he's doing, he's mentioned it more than a few times in this thread even going so far as to post some convergence charts. your frustration is due to your ignorance and his unwillingness to give you details. i don't remember lolatency outing himself to stealing his employer's code. that's just bullshit and is a pretty damn hefty accusation. this is where you're being a VERY big prick. as far as ideas go... so you're saying if you could trade your prop desk strategies independently or somehow use the ideas, you wouldn't right? that would make you a bad person, right? asiaprop, the mother teresa of finance... give me a break... and now back to the real world. you showed atticus your limits, lola showed you his, no difference. you think you're justified, atticus thought he was too. the worst you got was being called a dilettante though, your victim didn't fare as well. cut the shit asiaprop. regardless of how much 'factual data and logic' you've applied, you're behaving like a nutjob here. you don't accuse people of stealing code because you're 'frustrated', you don't judge a trader for gleaning ideas from their employer, and you sure as hell don't bully a young kid with a $50/day goal. wtf is wrong with you?
But markets boom and bust frequently. Market crashes are pretty common. It's rare to have a year where one country, asset class, or sector is not imploding or skyrocketing. How can something be a black swan if it happens every 3 or 4 years, and most experienced traders know how to handle it?
There's no mystery behind HF trading. If I said what I was doing even in the most general manner, that'd be the end of it. Most new ideas are simply twists of old ideas. By the private messages I am getting, I've come to the conclusion that other people already know or have guessed what I'm doing as well.
Fair enough and if you re-read you would not once see a single request on my side for you to expose your strategy, disagree? Also, have I ever sent you a PM asking you about anything? I have not. I merely asked about some general ideas of what HF guys are looking at. TsGannGalt already answered, that was all I was looking for, thanks.
I consider Weinstein's approach high-frequency, but different from me. He clearly has an edge. He looks at stocks where there's a tremendous movement and tries to capture the reverse trend over and over. The point being, he has an edge, he trades a lot intraday, and he's retail. What more did you want? He has a high-freq. edge. If he manages to make money with a 200ms+ latency, there's your answer. The guy really makes like 40k-100k a month, with the occasional down month. His tools? TradeStation Radar and Interactive Brokers. You don't get more retail than that. For the record, I never took anyone's source level code. I don't have a single line of source that was paid for by someone else. Do I know how other people's stuff works? Yes. Have I reverse engineered black box functionality? Yes. Can I recall from memory everything I wrote for them? Absolutely. But what do you expect? I wrote kernel code for a living. If you give me unmangled debug symbols, that's all I need to know what you are doing. ... As for running someone's code outright -- that's lunacy. I've never been anywhere where the code is so compact and trivial without multiple systems that it could have been stolen and just run at home, ... on some retail platform like Tradestation, or Genesis LASER no less.
Really, it's much more simple than you think it is. I know you don't believe, but 95% of the edge in HF trading comes from the technology, not the trading rules. I just happen to know [or think I know -- thus the experiment] where the technology doesn't buy anyone a real advantage.
It's a matter of speed vs. managability/extensability. If I want to have a static run-through tests it's quite simple. The issue is when I start automating and letting the AI takeover the development. One example is, when the walk-forward event is triggered, I have the AI decide what kind of walk-forward setup best suits the condition. Having a simple flat file data makes the data extraction a pain in the arse.
Appreciate the support. When I was inside the corporate walls, I used to read ET every lunch period. I wasn't allowed to post from work, because of policy. What got me was that the guys in the pair-trading thread and the Weinstein thread were doing well. Plus, where I actually worked, people actually left several ideas on the table that they didn't actually develop because they were looking at doing things that could make even more money. The fact of the matter is, ... it's not practical for large scale institutions to worry about $200-$300/day profits if they can put their manpower to use doing much much more. I don't have visions of myself making millions. I just have visions of myself managing an operation how I see fit, and that's what keeps me motivated. I don't like being told "no" when I see an idea I think can work. I just saw the fact that people left money on the table as an opportunity to break free. ... Because if I wasn't interested in institutional profits so much as just living out in the boonies working on models to collect the crumbs the lions and tigers left behind for me in the market. As for stealing, I really think that guy was way off on that. For me to have stolen my old employer's code would have required a lot of tools and components I just couldn't afford -- I would've had to steal all the licenses, certs, and then on top of that, I would've had no support for the third party components they used. Not to mention, the IT infrastructure in big banks that can afford it is very tight. People get fired for so much as printing code, USB is disabled, etc, etc. If I did steal, I'd have been arrested and prosecuted already. ... Reverse engineering quant libraries? I was just commenting on how absurd it is to protect optimizations on commodity-off-the-shelf math functions and act as if they were secretive. My point was that the overall institution was failing on account of the fact that it was restricting information it didn't need to restrict.