Discussion in 'Business, Careers & Education' started by NameBack, Oct 19, 2011.
Well, we're just testing right now. Pretty much nothing to report yet -- we won't really know if it works or not for at least six months, and frankly even that is a pretty small sample set. We're just glad it hasn't tanked.
It takes longer than 6 mo, it is often a very similar market for months on end. Ask most investors if the environment they're trading in seems similar to a few months ago and a year ago and you'll get very different answers. Most trading systems seem to be terrible at adjusting to different market conditions over long periods of time.
I agree that 6 mos is a pretty small n, but it's large enough that investors are willing to bet on you. At least that seems to be the feedback so far.
Currently we're in talks with a large national insurance company to work as an RIA under their financial advisory division. They get 35% of the profits but they do all the sales to fill up the fund and all the compliance dirty work, so it might be a good opportunity. Most likely though the deal would hinge on our test fund's performance over the next 3-6 mos.
sounds like good news
I'm just highly skeptical on an algorithm's ability to predict market movements that have been largely been dictated by the macroeconomic news in Europe, which, fundamentally, cannot be predicted by statistics, as they are purely qualitative events.
I think it's a big assumption that the events dictate the markets entirely -- how events are interpreted by investors is certainly a big part of the equation. And to the extent that how markets respond to events has to do with market trends, I think algorithmic trading can be successful even in macro-driven times. Macro events can also produce market trends that are easy to exploit -- for example crashes. I've mentioned before that our algorithm backtests quite well on crashes, because they're easy to pick up on, as they are large uni-directional swings. So if Europe produces a crash, that's a trend we can pick up on and exploit. The same goes for other trends as well. Often bad news out of Europe has produced downturns that last a week or more, and vice versa.
It's a big assumption and I'm only saying it because the recent (past couple months) market movements by and large have been dictated by macro news in the U.S. and Europe. I agree with you that it's possible to exploit macro events after-the-fact, but I'm just saying it's pretty much impossible for your algorithm to predict that the market will go tanky tank or sky high tomorrow because that kind of movement is produced by such macro news.
I can't count the number of times on my hands that the market has gone 2% one day, followed by an upswing of 3% the very day after because of some translucent shit that some politician in Europe said about the euro debt crisis while we were asleep. That kind of stuff, I'm saying, your algo cannot predict; nor can anyone else's.
Don't take this as me shitting on your algo. I think it's awesome and great, brave, even, for you to tackle such a task in this kind of market. I sincerely hope it works out for you so I can invest in it too
Yeah, I know what you mean, and we've definitely gotten fucked on some of these positions when Moody's or Draghi come out and say some bullshit. On the other hand, look at how much of what these politicians/bureaucrats/ratings agencies say that everyone already knows. Like when Draghi said that there was "substantial downside risk" and the market tanked 1%. No shit! Downside risk? In Europe? I had no fucking idea!
And while we all scratch our heads at why the market seems so stupid, it is an example of the markets interpreting the news as much, or perhaps more than, being dictated by the news.
We just went through a little slump though, so I'm not denying it's a rough market. We like big swings -- rallies, crashes, anything easy to pick up on -- or choppy up-one-day-down-the-next volatility. We can capitalize on either of those things. What we don't like is what the market has been doing recently where it'll climb for three or four days and then drop for three or four days. When the oscillation is the duration of one position in our model, we tend to lose a lot of positions. Ultimately though the dip is still shorter and shallower than the worst we've seen in backtesting, and we just won three positions in a row so we are hopeful that we've snapped back out of it.
Wow, I know nothing about this, but good luck. It's nice to see an SF member in the middle of MAKING IT
Want to wish good luck as well as this is something I've thought about for years and think has the possibility of being done with successful results. The fundamental portion of the market aspect is what scares me when it comes to a formula as such just because, well...people are stupid and react accordingly. Technically speaking, I think this can be done with great returns. Personally, I'm way more technical than fundamental, but fundamentals can take what should be a strong position technically and reverse course. Anyways, count me in with the believers that an algorithm can be used for market analysis and I wish you the best of luck with this venture.
I think companies like Zygna exist to remind us that people will bet on almost anything. Sounds like an interesting deal, I wish you well with it. I'm curious to see how it fairs in comparison to similar funds.
My strategy continues to be a value/growth blend of large cap stocks, without statistical overlays. I have in the past asked a day trader for input on an entry points for a stock, but I found there view to be so narrow that it was unnessecary.
What kind of statistics are you using? Mostly based on linear modelling I'm assuming? You can probably share a few tidbits I imagine without giving anything useful up, eh?
Well, there's price, and there's volume.. Everything else is derived from those two.
And yes even the VIX which gets mentioned in the news as an "indicator" of volatility is derived from them, the vix moves because of them, so it indicates shit.
I would suggest you talk with people before you are set on investing other people's money for a living. The reason I say this is disregarding anything about the algo if it fails or works miracles: People are assholes. You will get endless praise as long as you are making money for people. The moment you have negative returns people jump ship. Then they will try to call you at all hours send you death threats and all sorts of crazy.
Separate names with a comma.